U.S. Department of Labor

Office of Administrative Law Judges

800 K Street, NW

Washington, DC 20001-8002

(202) 693-7300

(202) 693-7365 (FAX)

Issue Date:

17 July 2019

Case No.: 2016-OFC-00006

In the Matter of:

Recommended Decision and Order

This case arises under Executive Order 11246 (30 Fed. Reg. 12,319 ), as amended, and the rules issued pursuant thereto at 41 C.F.R. Chapter 60 . Jurisdiction over this action exists under §§ 208 and 209 of Executive Order 11246, and 41 C.F.R. § 60-1.26 , 41 C.F.R. § 4.8 and 41 C.F.R. Part 60-30.

BACKGROUND

On June 8, 2016, the Office of Administrative Law Judges received an Administrative Complaint from the Regional Solicitor, Philadelphia Office, U.S. Department of Labor, on behalf of the Office of Federal Contract Compliance Programs ("Plaintiff" or "OFCCP"), for alleged violations of the above Executive Order by Enterprise RAC Company of Baltimore, LLC ("Defendant" or "Enterprise"). On June 30, 2016, Defendant filed Defendant's Motion to Dismiss Plaintiff's Administrative Complaint. Plaintiff filed an Opposition to Defendant's Motion to Dismiss Plaintiff's Administrative Complaint on July 14, 2016. Defendant filed Defendant's Reply in Support of Defendant's Motion to Dismiss on August 4, 2016. On August 8, 2016, Chief Administrative Law Judge Stephen R. Henley issued an Order denying Defendant's motion. (AX 2).1 Thereafter, on August 23, 2016, this case was assigned to me for adjudication.

[*2] On October 25, 2016, I issued an Order setting this case for hearing on August 14, 2017 and establishing deadlines for various prehearing events. (AX 1). On June 14, 2017, I issued an Order rescheduling the hearing to begin on June 19, 2018 in Baltimore, Maryland, and I adjusted the other prehearing deadlines accordingly. (AX 5). On September 6, 2017, I issued an Order resolving a dispute over pretrial deadlines. (AX 6). At the parties' request, I subsequently modified the schedule and the deadlines, and the hearing ultimately began on June 19, 2018 and ended on June 27, 2018. (AX 7).

Defendant is a car rental company. In addition to renting cars to the general public, it rents cars to the United States Government through a government contract that began on October 1, 2002. On May 1, 2008, Plaintiff issued a letter scheduling a compliance review of Defendant's Linthicum, Maryland, car leasing facility. Plaintiff audited Defendant's hiring practices for the period August 1, 2006 through July 31, 2008. On March 13, 2013, Plaintiff issued a Notice of Violation alleging that Defendant: (1) discriminated against African-American applicants on the basis of their race in hiring for Management Trainee positions; (2) failed to maintain all the data used in its recruitment and selection process; (3) failed to conduct an adverse impact analysis of its total selection process for all positions; and (4) failed to develop and implement an auditing system that periodically measured the effectiveness of its total affirmative action program.

Between April 2013 and May 2014, Plaintiff and Defendant held six conciliation meetings regarding the alleged violations.3 The conciliation meetings failed to resolve the dispute. Plaintiff issued a Notice to Show Cause why it should not initiate enforcement proceedings regarding the violations alleged to have occurred from August 1, 2006 through July 31, 2008. Thereafter, Plaintiff filed its complaint alleging that the violations occurred from August 1, 2006 through July 31, 2008, and that discrimination was continuing to occur.

On October 19, 2016, Plaintiff served Requests for Admissions, Interrogatories, and Requests for Production of Documents, several of which related to dates that extended beyond July 31, 2008. On December 27, 2016, I received Defendant's Motion for a Protective Order. Plaintiff filed its Opposition to Defendant's Motion for Protective Order on January 17, 2017. Defendant's motion was denied in an Order issued on March 27, 2017. (AX 4).

On June 27, 2018, the final day of the hearing, it was agreed that the evidentiary record would remain open until July 13, 2018 for Defendant to submit job applications and interviewer notes compiled by Defendant's expert witness, Dr. Paul F. White. (TR 1487). On July 13, 2018, Defendant submitted DX 7366 by email, which contained a link to an electronic ShareFile folder that purported to contain the information compiled by Dr. White. I was unable to access the [*3] ShareFile folder and Defendant subsequently provided the same information to me on a thumb drive.4 Plaintiff filed objections to DX 7366 the same day and Defendant filed a response on July 21, 2018. In an Order dated August 27, 2018, I admitted DX 7366 and extended the deadlines for the submission of final briefs.

Plaintiff and Defendant submitted final briefs on September 28, 2018 and reply briefs on October 19, 2018. The evidentiary record is now closed.

FINDINGS OF FACT AND CONCLUSIONS OF LAW

The parties stipulated to the following.

• (1) OFCCP brought this action to enforce Executive Order 11246 and has jurisdiction over this matter under Sections 208 and 209.
  
• (2) Enterprise is a car rental company with a business office in Baltimore, Maryland.
  
• (3) Enterprise is a wholly-owned subsidiary of Enterprise Holdings, Inc.
  
• (4) Enterprise is a federal government contractor within the meaning of OFCCP rules with an ongoing federal contract that began on October 1, 2002, to provide rental cars to the Department of Defense Military Traffic Management Command.
  
• (5) At all relevant times, Enterprise provided services as a government subcontractor within the meaning of Executive Order 11246.
  
• (6) Pursuant to Section 202 of Executive Order 11246 and Title 41, Section 60-1.4(a)(1), of the Code of Federal Regulations , Enterprise agreed not to discriminate against any applicant for employment because of race.
  
• (7) The management trainee position at issue in this case is an entry-level position that can lead to promotion to positions of greater responsibility and higher compensation. The management trainee position is compensated on an hourly basis that at the time of the hearing paid approximately $45,000.00 per year.
  

(TR 9-11).

The parties agreed that the issues that were in dispute were the following.

• (1) Whether Enterprise discriminated against African-American5 applicants seeking employment as management trainees between August 1, 2007 and July 31, 2017.
  • (a) Whether Enterprise engaged in a pattern or practice of disparate treatment of African-American applicants — intentional discrimination; or
    
  • (b) Whether Enterprise is liable for the disparate treatment impact its hiring process had on African-American applicants.
    
  
  
• (2) [*4] Whether Enterprise failed to preserve and maintain all personnel and employment records for a period of two years from the date of the making of the record or personnel action involved pursuant to Section 202 of Executive Order 11246 and Title 41, Section 60-1.4(a)(1), of the Code of Federal Regulations .
  
• (3) Whether Enterprise failed to conduct an adverse impact analysis of its total selection process for all positions in accordance with Title 41, Section 6-3.4 and 3.15(a)(2)(A) , of the Code of Federal Regulations.
  
• (4) Whether Enterprise failed to develop an auditing system to periodically measure the success of its affirmative action program in accordance with Title 41, Section 60-2.17(d), of the Code of Federal Regulations .
  

(TR 11-12).

TESTIMONIAL EVIDENCE

JEFFREY W. ARNEY

Mr. Jeffrey W. Arney was called as OFCCP's first witness. (TR 99). Mr. Arney is the human resources manager for Enterprise Group 18. He explained that the company is organized into various numbered groups and the Baltimore group is Group 18, signifying that it was the 18th group the company created. Group 18-99 indicates the administrative headquarters for Group 18, which is located in Linthicum, Maryland. (TR 100). Mr. Arney has been in his current position since April 2003. His responsibilities include employee retention, diversity and inclusion efforts, employee relations and benefits administration. He began managing the talent acquisition team about two years ago. (TR 101).

Ms. Terri Morris serves as the talent acquisition manager and reports directly to Mr. Arney. She oversees the entire talent acquisition effort, such as recruiting, maintaining relationships with business partners and universities, and handling employee referrals. Part of her effort includes outreach to historically black colleges and universities ("HBCU"), including three HBCUs that are located in the Group 18 area: Morgan State University, the University of Maryland Eastern Shore, and Coppin State University. (TR 102-103). Ms. Elizabeth Hardesty also has the title talent acquisition manager, but she reports to Ms. Morris. There is a new talent acquisition coordinator whose first name is Sandy and her last name starts with the letter Z, but Mr. Arney could not spell it or pronounce it. Mr. Wes Wucher was the talent acquisition coordinator for two years before that. The talent acquisition coordinator's job requires a lot of data processing, running background checks and scheduling new hires for new employee orientation. (TR 103-104). At some point, Mr. Wucher began reviewing management trainee applications that are submitted on the Internet. (TR 104). Ms. Lauren Hughes was the talent acquisition coordinator for about two years prior to Mr. Wucher and she was involved in screening Internet applications as well. Ms. Tiffany Gomez also performed some talent acquisition coordinator functions at some point around 2000. (TR 105).

Enterprise's fiscal year begins on August 1 and ends on July 31 of the following calendar year. In fiscal years 2007 and 2008, Ms. Sara Lichter was the talent acquisition coordinator and was involved in reviewing applications for management trainee positions that were submitted [*5] over the Internet. She had some involvement in conducting telephone interviews to screen applicants for the management trainee program. She was trained on her duties by Ms. Morris and Ms. Hardesty. Ms. Lichter left her position in the fall of 2008 when she was reassigned to the loss control division when Enterprise reallocated some of its personnel during the economic downturn. The position remained vacant for a period of time until Ms. Hughes assumed those duties. (TR 106-107).

Enterprise continued recruiting for its management trainee program during the 2008 recession. It is considered an "evergreen position," which means that Enterprise is constantly recruiting for new management trainees. Applicants are considered individually rather than against each other. Mr. Arney estimated that on average about 10 people are hired as management trainees each month and there are about 125 management trainees in Group 18 at any given point in time. There are about 1,500 employees in Group 18 spread across 55 to 60 branch offices. Mr. Arney described a branch as a "brick-and-mortar, roof-and-four-walls operation in which a team of employees are providing service to the community for their rental needs." (TR 108). A typical branch has a manager, an assistant manager, some number of management trainees and some number of other individuals who service and drive the rental vehicles. Someone who rents a car from Enterprise has the option of picking it up at the branch location or Enterprise will send someone to pick up the customer. Anyone in the branch, from the manager to a management trainee to one of the service personnel, could go out on a pickup run. (TR 108-109).

Mr. Arney began his employment with Enterprise as a management trainee in February 1999. The management trainee program is an entry point into the organization. Enterprise also has management interns who are college students hired to work over the summer. (TR 110). Management trainees begin their employment at Enterprise with a one week orientation course. (TR 111). The orientation course is run by the talent development manager, Ms. Noell Jeanie Landolt. There is a new course starting about every two weeks. Mr. Arney participates in the course and talks about general company policies and procedures, EEO policies and procedures, personal appearance standards and employee benefits. His part of the course lasts for about three hours. (TR 112). Others speak to the trainees about company culture and core values, how to use the computer system, how to write tickets and how to take a reservation, among other things. At midweek, management trainees go to their branch offices and then they come back to the headquarters where they are introduced to some other departments, like risk management, corporate marketing, and vehicle repair. A group rental manager provides a talk on an introduction to sales and then the week is capped off with a test of the management trainees' comprehensive knowledge. (TR 113).

Management trainees return to the headquarters for more sales training at about the 30-day, 90-day, five-month, and eight to nine-month marks. They also get on-the-job training. Branch managers are responsible for SOS — Show, Observe, Shape — where they show trainees how to do something, observe them doing it and then help to shape their behavior. The SOS approach is used for both sales and customer service training. (TR 114-115). Mr. Arney said that his sales and customer service experience before joining Enterprise was working at Walt Disney World and he spent a lot of time in his initial interview talking about his Disney experience. (TR 116). Management trainees work at their branch locations about 48 to 49 hours [*6] per week. They are paid at an hourly rate and get overtime pay for the hours that exceed 40 hours per week. They help customers rent cars, fill out rental contracts — also called tickets — and take in rental car returns. (TR 117). They are expected to do anything that needs to be done at the branch location, including picking up customers and washing cars when it is necessary. (TR 118).

Mr. Arney reviewed PX 1 and PX 2, which are management trainee job descriptions in effect during the charged period. (TR 119-121). He also reviewed PX 48, which are responses to interrogatories that he assisted in writing on behalf of Enterprise, and PX 47, which are supplemental responses. (TR 121-122). On page 50 of the supplemental response, in the answer to interrogatory 17, Enterprise stated that it provided OFCCP with all of the management trainee job descriptions for the period from August 1, 2008 to July 31, 2017. (TR 123). Mr. Arney agreed that PX 2 is what is referenced in the response. (TR 124). Mr. Arney reviewed PX 3, which is recruiting materials that Ms. Morris and Ms. Hardesty utilized. He said the materials may have come from the corporate office. (TR 124-125). He agreed that Enterprise advertises itself as a good place to start a career and that those who join the company and do a good job are afforded opportunities for advancement. (TR 126).

New hires remain in management trainee positions for about 10 months and that leads up to a position as an assistant manager. (TR 128). An assistant manager is second in command at a branch location and is responsible for fleet movement and employee development when the manager is not available. A person serves in an assistant manager position for about six months. Mr. Arney said that he worked on the rental side as a management trainee and assistant manager for about a year and a half before moving into HR where he started out as a human resources coordinator with mainly administrative responsibilities. (TR 129-130). He stayed in that job for about a year and then became a human resources generalist where he had added responsibilities in employee relations and providing support to managers on employee matters. He was in that job for about two years before he was promoted to his current position as HR manager in April 2003. (TR 130-131).

Mr. Arney reviewed PX 5, which he described as the compensation and career path guideline that is provided to applicants during the interview process to illustrate what their compensation will look like during the first year of employment. (TR 131-132). It states that management trainees are guaranteed $33,500 for the first year based on a 49 hour work week, including performance bonuses. (TR 133). That figure includes an hourly wage plus overtime, a $500 skills test bonus, a $1,200 management qualification interview bonus, and a $500 additional performance bonus. (TR 134). The management qualification interview — also called an MQI — takes place at the nine to 12 month mark and tests whether the management trainee is ready to move up to an assistant manager position. The oral portion of the MQI is conducted by an area manager. An area manager is essentially a district manager and is responsible for overseeing three to six of Enterprise's branch locations. (TR 135).

Mr. Arney reviewed PX 6, which are additional compensation and career path handouts. (TR 136-137). The compensation plan dated August 2010 indicates that a new management trainee will earn between $33,000 and $35,000 in the first year. The compensation plan dated August 2012 indicates a total compensation of $36,082 in the first year. The plan dated August [*7] 2015 estimates that a first-year management trainee will earn $42,246.72 for the year. The August 2015 plan includes additional performance bonuses for management trainees who are on the fast-track and could earn a higher income. (TR 138-140). The fiscal year 2017 plan says that a management trainee who is on the fast-track can earn $44,194 in the first year of employment. (TR 141).

Mr. Arney reviewed PX 7, which he said is a benefits practice guide, personnel policies and benefits summary plan description. (TR 142). The guide provides management trainees an overview of the benefits that are available to them, such as health insurance, retirement savings and profit-sharing. (TR 143-144). Enterprise offers a retirement savings plan that includes employer matching contributions of up to three percent. (TR 144).

Mr. Arney testified that he had "[h]eavy involvement" in preparing and reviewing impact ratio analyses, or IRAs. The IRAs for hiring are prepared each year by Ms. Morris and Ms. Hardesty. The statistics evaluate applicant-to-hire ratios and whether there are any potential adverse impacts. (TR 145). A third party vendor, Outsoft, assists Enterprise in compiling its annual affirmative action plans. Anything over two standard deviations on the hiring IRAs would be cause for a review of the hiring process. (TR 146). Mr. Arney was provided a copy of his deposition transcript, which is PX 49, and he was asked to look at page 170. (TR 147). At his deposition, he said that Ms. Morris prepared hiring IRAs twice a year and that Enterprise retained IRAs back to 2006. (TR 147-148). Mr. Arney said at the hearing that the IRAs for fiscal years 2007 and 2008 showed adverse racial impacts for African-American management trainee applicants. (TR 149). He said that he, Ms. Morris and former general manager Mr. Kevin Molitor met and discussed the adverse impact following the 2007 IRA and they decided not to make any changes in their recruitment and hiring processes. (TR 149-150).

Mr. Arney reviewed PX 23, which is an affirmative action plan — also called an AAP — for fiscal year 2008. (TR 151-152). He said an AAP and an IRA are not the same thing. (TR 153-154). Mr. Arney reviewed PX 24, which he said are the IRAs. (TR 154-155). On page 2 of PX 24, beneath where it says "Impact Ratio Analysis, Hires," Mr. Arney said "Group 2" represents "professionals" and that would include primarily management trainees and assistant managers, as well as some other professional positions, like accountants. (TR 156). The IRA shows 47 total hires, including 14 minorities out of 71 minority applicants. (TR 157-158). There are 33 white hires out of 99 white applicants. (TR 158). The standard deviation is 1.96, which is rounded up to two. (TR 158). Any time the standard deviation is two or more, Mr. Arney said it is "cause for conversation about why that number is what it is. And we look at our processes and take into consideration if we want to make adjustments." (TR 160). This IRA showed results for fiscal year 2009 and it was evaluated going into the beginning of fiscal year 2010. Mr. Arney did not recall there being any adjustments made as a result of the IRA. (TR 160). The information for fiscal year 2010 shows a standard deviation of 1.99 and Mr. Arney did not recall if that led to any changes in recruitment and hiring practices. (TR 160-161). The information for fiscal year 2011 did not trigger the computer program to calculate a standard deviation. (TR 161). The information for fiscal year 2012 shows a standard deviation of 2.85 and on the printout under the heading "adverse impact" the software program inserted the word "yes." (TR 162). Mr. Arney did not recall if the IRA for fiscal year 2012 led to any changes in recruiting and hiring practices. The IRA for fiscal year 2014 did not cause the computer program to calculate a standard deviation. (TR 163).

[*8] Mr. Arney reviewed PX 25, which he said is the first letter he received from OFCCP District Director Tom Wells. The letter is dated May 1, 2008. (TR165). The letter was sent to the Group 18 headquarters and it was addressed to the Group 18 former general manager, Ken Molitor. Mr. Molitor passed the letter along to Mr. Arney and tasked him with responding on behalf of Enterprise. In the letter, OFCCP requested Enterprise to submit its AAPs and data for review. (TR 166). Enterprise provided a copy of its government contract as part of its submission to OFCCP and the contract is at PX 26. (TR 167). OFCCP sent a predetermination notice dated January 26, 2012 to Mr. Arney. It advised that a predetermination notice is issued "when the findings of an evaluation indicate the existence of a potential pattern and practice of discrimination." (TR 168).

On March 13, 2013, OFCCP issued a notice of violations. The notice was addressed to Mr. Gavin Watson, Vice President and General Manager of Enterprise Group 18, who by that time had replaced Mr. Molitor. (TR 169). Violation 1 stated that "[d]uring the period August 1st, 2006 through July 31st, 2008, Enterprise discriminated against black applicants on the basis of their race in hiring for management trainee positions." (TR 170). Violation 2 stated that "Enterprise failed to maintain all the data used in its recruitment and selection process, including applications and review notes." (TR 171). The applications and review notes were compiled by Ms. Morris and Ms. Hardesty and there was some data that could not be produced and provided because of a corrupt database. (TR 171-172). Violation 3 stated that Enterprise failed to conduct an adverse impact analysis of its total selection process for all of its positions in violation of applicable regulations. Mr. Arney said Enterprise ran IRAs for every job group. (TR 172). That included Job Group 2, which included the management trainee positions that are in dispute. Enterprise did not run an IRA on all the job groups combined. Violation 4 states that "Enterprise did not develop and implement an auditing system that periodically measures the effectiveness of its total affirmative action program." (TR 173). Mr. Arney disagreed that Enterprise did not review its total affirmative action program. He acknowledged that they reviewed IRAs by job groups. (TR 173). The notice of violations was signed by Mr. George Rouse, the District Director for OFCCP's Richmond District. (TR 174). Mr. Arney also acknowledged that in its response to interrogatories, Enterprise said it implemented no changes to ensure compliance with Executive Order 11246 and the rules and regulations issued pursuant to 41 C.F.R. Chapter 60 between August 1, 2006 and July 31, 2017. (TR 174-175; PX 47 at 74-75).

On cross-examination by Enterprise, Mr. Arney said that the historically black colleges and universities where Enterprise recruits also produce white graduates. (TR 178). With respect to management trainee positions being "evergreen," a particular branch would at times have filled all of its management trainee requirements. Enterprise tries to place management trainees at branch locations that are close to where they live. If the nearest branch did not have any management trainee openings then a new hire would be redirected to another branch location. (TR 178-179). Mr. Arney said that when he was discussing PX 24 earlier, when the term "adverse impact" was used, he was referring to a statistical adverse impact as reflected on the document and he was not drawing a legal conclusion. (TR 183). The designation "Group 2" in the document includes more than just management trainees and the term "minority applicants" includes more than just African-Americans. (TR 184). For fiscal years 2011, 2014 and 2016, no adverse impacts on minority applicants were reported. (TR 184-187).

[*9] On re-direct examination by OFCCP, Mr. Arney said that Ms. Morris would prepare an IRA that looked at African-American applicants and hires. (TR 188). It compared African-American applicants to the total pool of applicants. The analysis was initially run by job group, but if anything prompted review, Ms. Morris would do a "step IRA" that would break it down in more detail. Mr. Arney did not recall whether there were any step-IRAs run for management trainees during the charged period. (TR 189).

Enterprise called Mr. Arney as its first witness. Mr. Arney said that the management trainee position is the entry point into the company and the starting point for advancement into higher managerial positions. Management trainees provide direct customer service and they make the initial first impression on customers. (TR 190). "Core competencies" are the different skills and experiences Enterprise determined necessary for success in a given position. Sales ability and interest in a career with Enterprise are "extremely important." (TR 191). Enterprise is looking for people who are interested in sales and who are likely to stay with the company for three or four years and advance into higher level management positions. (TR 193).

Compliance with OFCCP regulations is part of the Mr. Arney's job responsibilities. He is involved in preparing Enterprise's affirmative action plan. (TR 193). The plan — referred to as an "AAP" — is prepared by an outside vendor, Outsoft. Every year Outsoft submits material for Enterprise to consider putting into its AAP. Outsoft relied upon information provided to it by Enterprise. (TR 194). The information was provided by the Enterprise corporate office and it included an employment database and records of existing workforce job group analysis. Enterprise provided an AAP in 2007 and in subsequent years. (TR 195).

The last page of the AAP is a "utilization analysis" that shows the number of people in each job group by race and gender and whether there is underutilization compared to the "Metropolitan Statistical Analysis," which are statistics for the relevant geographic region. (TR 195). The Metropolitan Statistical Analysis is based upon government data for the Baltimore metropolitan region. (TR 196). Mr. Arney reviewed DX 85, which is the Group 18 AAP for fiscal year 2009 — which is August 1, 2008 through July 31, 2009 — and DX 86, which is a utilization analysis for the same period. (TR 197-199). The document is signed by Mr. Arney and Mr. Molitor. It shows that in 2009 the professional group at Enterprise that included management trainees was made up of 35.8 percent minorities while availability per the metropolitan statistical analysis was 30.8 percent. (TR 199). Since minority representation at Enterprise was five percentage points above the regional availability rate, the underutilization column on the AAP said "no." (TR 200). Mr. Arney said that in the professional category, Enterprise always exceeded the utilization rate. (TR 202). DX 88 shows the utilization rate for fiscal year 2010 and DX 91 shows the rate for fiscal year 2011. Enterprise exceeded the minority utilization rate in both of those fiscal years. (TR 202-204).

Mr. Arney said the term "diversity initiative" means efforts within the organization to make the workforce reflect the community it serves through training focused on hiring, promotions, advancement and representation. (TR 206-207). Mr. Arney, Ms. Morris and Ms. Hardesty "shadowed" area and group managers when they conducted interviews to ensure compliance and to evaluate their skills. (TR 208). Employees on the talent acquisition team also [*10] received diversity training. Enterprise identifies itself as an equal opportunity employer and mangers are trained not to discriminate against applicants. Mr. Arney testified that diversity is extremely important to him. (TR 208-209). Ms. Morris and Ms. Hardesty spend a lot of time at the area's historically black colleges and universities, and they have a great relationship with the staff members on those campuses. They are invited to give talks and at Morgan State University they created a "round robin speed networking event." (TR 209-210). Mr. Gavin Watson, the current general manager for Group 18, is a member of the board of the Greater Baltimore Urban League, as was his predecessor, Mr. Molitor. The Urban League is "geared towards African-Americans" and the outgoing Urban League CEO Howard Henderson came to Enterprise and provided career advice to minority employees. (TR 211).

Mr. Arney said Enterprise does IRAs every six months and has been doing so since he joined HR in 2000. (TR 211-212). They have also done AAPs for as long as he has been in HR. It is standard practice for Enterprise to retain each application file. (TR 212). The application files are kept in bankers' boxes and they were turned over to OFCCP when they were requested. Each file contains an application and whatever other documents that are generated during the hiring process. (TR 213). Enterprise reviewed the IRAs every six months to see if the numbers reflected anything that might cause concern. (TR 214).

Mr. Arney said that when he started out at Enterprise in 1999 it had a great reputation for inclusion and providing opportunities for success. Mr. Arney testified that he did not know anyone when he moved to Maryland to start work. He noted at the hearing that he was wearing a pride pin on his lapel because he was proud to be a member of the LGBT community. He said his focus since joining Enterprise has been to have a positive impact on the culture of the group and the company. He was able to serve on the North American Diversity Team, a group of senior leaders who make key decisions on the culture of the company and how the company will proceed in the future. Inclusion is one of the components of training for new employees. One of Enterprise's core values is an open door policy where discussions about diversity are always welcome. Mr. Arney said training includes the importance of multicultural customers who as a group are extremely loyal Enterprise customers. Enterprises strives to make its workforce reflect the community it serves and vice versa. Many of the Group 18 branches are located in African-American communities, and many management trainees, assistant managers and branch managers are African-Americans. (TR 215-217).

On cross-examination by OFCCP, Mr. Arney said he is not a recruiter and he did not interview any management trainee applicants during the charged period. (TR 218). With respect to DX 86, the utilization analysis, he agreed that the availability statistic says that it is a "weighted factor" and he agreed that he did not know how the weighting is done. (TR 219). Mr. Arney reviewed PX 45, Defendant's Response to Plaintiff's First Requests for Admissions, and agreed that he helped prepare the responses. He agreed that Request for Admission Number 45 asked Enterprise to admit that it was unable to provide the applications for 14 named individuals and Enterprise responded that four of the applications had previously been provided and that there were 23 electronic applications for management trainee positions submitted by 23 named individuals between August 1, 2006 and July 31, 2008 that Enterprise was unable to produce and provide because the electronic files were corrupted. (TR 222-224).

[*11] On redirect examination by Enterprise, Mr. Arney said that the AAPs that Enterprise submitted to OFCCP were never sent back to them to be redone. (TR 225).

In response to questions by me, Mr. Arney agreed that when Enterprise looks for management trainees it looks for individuals interested in making Enterprise a career, not just individuals who want a job for a few years. He also agreed that there were 200 to 225 people in the professional job category under Group 2 in the IRAs and that management trainees made up about 75 percent of the Group 2 total. He said about 100 to 120 management trainees are hired each year and he agreed that the annual turnover rate is about half of those who enter the program. (TR 225-226).

On re-cross-examination by OFCCP, Mr. Arney said the turnover rate did not include management trainees who were promoted into assistant manager positions. The turnover numbers mainly reflected management trainees who left Enterprise to pursue other opportunities and some that Enterprise chose to terminate. (TR 227).

TERESA L. MORRIS

(Initial Testimony)

The second witness OFCCP called was Ms. Teresa L. Morris. (TR 229). Ms. Morris said that she joined Enterprise as a management trainee in October 1989. It was her first — and only — job after graduating from college. She later became a branch manager, then a human resources supervisor, and then human resources manager. Her current job title is talent acquisition manager. She has been involved in recruiting at Enterprise for 25 years. (TR 230-231).

Ms. Morris agreed that management trainees help customers rent cars, take reservations, fill out rental contracts, accept returns, do pick-up and drop-off runs, and at times wash cars. Part of the job involves marketing, which includes going out and meeting with clients — body shops, car dealerships and insurance agencies — to make sure their needs are being met. (TR 233-234). Management trainees work 48 hours per week, which has been a consistent practice from fiscal year 2007 through fiscal year 2017. (TR 234). Normal branch location hours are Monday through Friday from 8:00 a.m. to 6:00 p.m. That usually means coming in around 7:30 a.m. to get ready for the morning period, which tends to be the busiest part of the day, and staying until about 6:30 p.m. to wrap-up all of the loose ends at the end of the day. Branches are open on Saturdays from 9:00 a.m. to 1:00 p.m. and flagship stores have Sunday hours from 10:00 a.m. to 2:00 p.m. Management trainees work every other Saturday and they get a flex day that week to help maintain a good work/life balance. The Baltimore airport location is "pretty much a 24/7 hours — absolutely, 24/7." (TR 235-236).

Ms. Morris said one of the things she likes to share with potential applicants is the opportunity to do "career window shopping" at Enterprise. By that she means that Enterprise promotes from within, so someone starting out as a management trainee who does well can later go into fleet services, human resources or any of the other product lines and they will not have to hop from company to company in order to have an opportunity to move up in the organizational structure. (TR 237). To move up from a management trainee position to an assistant branch [*12] manager position requires a minimum of six months of work experience and the norm is about 10 to 12 months. (TR 237-238). Assistant branch managers get base pay plus a percentage of the profits from their stores, which gives them a sense of buy-in and an incentive to grow their operations. The next step is branch manager, which requires leading a team, doing marketing, handling administrative duties and anything else that needs to get done to make the branch a success. (TR 238).

Management trainees start with a one-week orientation course. It teaches them about the history and culture of the company, how to use the computer system and how to process transactions. The goal is to make them ready to start working the next week when they report to their branches. (TR 239). They get on-the-job-training from their branch managers and assistant branch managers and they come back to headquarters for more training at four stages over the course of the first year. (TR 240). They come back to headquarters at the 30-day point for stage one training, which is a half-day of training focused on customer service. (TR 241). They come back for stage two training at the 90-day point for a half-day of training focused on internal sales (upselling customers) and external sales (marketing). (TR 242). Stage three training is at the five to six month mark and focuses on career planning. Stage four training is "big picture" and includes business ethics. (TR 243). Ms. Morris talks about the employee referral program during the initial orientation course. Management trainees can earn a bonus if they refer future applicants to the company. (TR 244).

In the fiscal year 2007 and 2008 timeframe, applications were completed online using the SHL system, which was sometimes called IRISS. (TR 245-246). The system allowed Enterprise to track the applicant from the submission of an application, through the interview process, to a hiring decision. The online application gave the applicant the opportunity to self-identify his or her race. (TR 246). The recruiting coordinator, Ms. Lichter, pulled the information off of the system on a daily basis. She reviewed the information on her computer screen to make sure applicants met the minimum qualifications for the management trainee program. Ms. Lichter reports directly to Ms. Morris as does Ms. Hardesty who is a talent acquisition specialist. (TR 247). Ms. Lichter's job was more administrative in nature than Ms. Hardesty's job. (TR 247). If Ms. Lichter was not sure whether an applicant met the minimum qualifications, she called the applicant on the telephone to get additional information. Ms. Morris said she made sure her team understood that if it was unclear whether an applicant was qualified they should give the benefit of the doubt to the applicant. (TR 248). Ms. Lichter did the telephone screenings and then either scheduled the applicant for an interview or assigned an "S" code giving the reason why the applicant was rejected. (TR 249). Ms. Morris and Ms. Hardesty trained Ms. Lichter on how to do the telephone screenings. She listened in as Ms. Morris and Ms. Hardesty did telephone screenings and then they listened in as she began to do them herself. By 2008, Ms. Lichter could do the telephone screenings and apply "S" codes to rejected applicants without having to consult with Ms. Morris or Ms. Hardesty, although she could still ask them for advice if she was in doubt. (TR 250-251).

Ms. Morris said PX 8 is an overview of how Enterprise reviews applications. (TR 251). Ms. Morris said that she believed she prepared the overview and it accurately reflects the hiring process used in fiscal years 2007 and 2008.

[*13] Step 1 was an application review, which required someone from the human resources recruiting team to log into the tracking system daily to review applications. (TR 252-253). Management interns could also apply using the online system. (TR 253). Each application was reviewed to make sure the applicant met the minimum qualifications. That included at least two years of college at the time, but around fiscal year 2009 the requirement was increased to a four year college degree. (TR 254). It also included at least two years of sales, customer service or related work experience. The overview listed sales or customer service experience, "such as retail sales, restaurant sales, commission sales, et cetera." (TR 255). The listing was intended to provide some examples of qualifying work experience and was not meant to be an all-inclusive list. Other requirements included job stability, defined as not more than two non-temporary or seasonal jobs in the previous 12 months, and work eligibility, meaning either U.S. citizenship or a visa that granted the applicant legal work status in the United States; a valid driver's license and not more than two moving violations in the previous three years; no criminal convictions within the previous five years for drunk driving or an alcohol or drug related offense and no felony convictions for offenses involving dishonesty or violence. (TR 256-257). Any individual who failed to meet any of the requirements or declined to provide the information was eliminated from consideration at that step. All of those who met the minimum qualifications and those where the information provided required further clarification were selected for telephone screening interviews. (TR 258).

Step 2 was a telephone screening interview. The screener asks questions about the six criteria listed in Step 1 and clarifies any ambiguities about the applications. (TR 258-259). The screener also asks about the individual's interest in a career involving sales and customer service, and the applicant's willingness to accept the pay, hours and job locations that Enterprise offered. (TR 259-260). The purpose of asking the latter questions was to make sure that if the applicant came in for an interview he or she came in with an understanding of what the job involved. (TR 260-261). Applicants could be eliminated at this step if an answer showed he or she did not meet one of the minimum requirements or if he or she self-selected out to pursue another opportunity or was unwilling to accept the pay, hours, or work locations. (TR 261-262).

Step 3 required the applicant to come in for a face-to-face interview with an Enterprise recruiter. The recruiter could be the same person who did the telephone interview if Ms. Morris or Ms. Hardesty happened to do the telephone screening. (TR 263). Ms. Lichter did not do face-to-face interviews. The first interview was designed as a "behavior-based interview" where applicants are asked to provide examples of things they had done that was related to the core competency the interviewer described. (TR 263). For example, Ms. Morris said that she might ask an applicant to describe a time he or she had to deal with a difficult customer. A behavior-based question is asked for each of the six core competencies. A candidate who is lacking in any of the six core competencies is eliminated at this step. (TR 264-265). The first interview is conducted at the Group 18 headquarters by Ms. Morris or Ms. Hardesty. The second interview is conducted at a branch location by a branch manager. (TR 265). The second interview follows the same behavior-based approach as the first interview and an applicant who fails any area is eliminated from further consideration. (TR 266). The third interview follows the same format and is conducted by a group rental manager. (TR 266). There are normally three group rental managers in Group 18. The group rental manager makes the final decision on whether to extend a job offer to the applicant. (TR 267).

[*14] Ms. Morris reviewed PX 44, which is a record of her interview by OFCCP Compliance Officer, Ms. Gail Steele. (TR 277-278). It is dated August 16, 2012 and it is signed by Ms. Morris. (TR 278). Ms. Morris agreed the statement that she and Ms. Hardesty conducted the first face-to-face interviews at the headquarters in Linthicum before applicants were forwarded for branch, area and group manager interviews is accurate. She said the only times a first interview is not conducted at headquarters is when they are on college campuses and do an in-person, on-campus interview. (TR 279). On-campus interviews are currently arranged through an online program called Handshake, but in the past prospective applicants arranged interviews through the career centers at their schools. (TR 280). In the fiscal year 2007 and 2008 period, some applicants interviewed on-campus were rejected for failing to meet minimum requirements. There were six or seven applicants that Ms. Morris could recall and they were rejected for reasons like not having driver's licenses. (TR 281). Those applications were accepted and "S" codes were applied. Ms. Morris said they recruit on college campuses in the fall and spring, but she could not estimate how many applications came in through the on-campus process. (TR 282). She estimates that she spends 30 to 40 percent of her time doing on-campus work during the peak fall and spring periods. (TR 284).

Ms. Morris agreed that in the interview with Ms. Steele she said that she assisted in developing the written procedures and practices for recruiting and hiring management trainees in the August 1, 2006 through July 31, 2008 period for Group 18 and that the corporate office assisted and approved the procedures and practices. (TR 284-285). The procedures are at PX 8. (TR 285).

PX 10-A contains the screening criteria implemented on January 1, 2006. (TR 287). Those criteria included 48 college credit hours (and a bachelor's degree strongly preferred), at least 21 years of age (or at least 18 years of age if the applicant had a college degree), previous or current experience in sales or customer service, no more than two non-temporary or seasonal jobs in the previous 12 months, legal eligibility to work in the United States, no alcohol or drug convictions within five years, no felony convictions for offenses that are sufficiently job-related, and a valid driver's license. (TR 288-294). As examples of sales experience, Ms. Morris said they look for applicants who have had direct customer contact, like waiters or bartenders in restaurants, bank tellers and military recruiters. As examples of customer service experience, she listed working at a summer camp, serving as a telephone customer service representative at a place like Verizon, or working as a home health aide. (TR 290-291).

PX 10-B contains the screening criteria implemented on August 1, 2008. (TR 294). There were some changes from the 2006 criteria. The education requirement increased to a college degree. If the applicant was not a business or communications major, he or she was required to have at least one year of sales experience within the previous three years. (TR 295). While it was not included in the written criteria, Ms. Morris testified that a business or communications major was expected to have six months to a year of sales experience while other majors were expected to have at least a year of sales experience. She said the same applied to customer service experience. (TR 296).

[*15] PX 10-C contains the screening criteria implemented on August 1, 2009. (TR 297). It required a business or communications major to have at least one year of relevant sales experience within the last five years and other majors must have one year of relevant sales experience within the last three years. (TR 298).

PX 10-C also contains the screening criteria implemented on August 1, 2015. (TR 299). It included "S" codes Enterprise used in it applicant flow logs. Ms. Morris said Enterprise implemented its new applicant tracking system called ICIMS in 2015 and the new system had a drop-down menu for the "S" codes where you picked from a list of screening criteria disposition codes. (TR 300). The criteria implemented on August 1, 2015 required a minimum of one year of experience within the last six years in two or more of the following areas: sales [(S3) commission sales strongly preferred], customer service in a service industry [retail, restaurant or a call center], military leadership, or college athletics. (TR 300-302). Applications were reviewed for spelling, grammar and missing information (S4). (TR 302). Applicants were evaluated on their professionalism [(S5), including whether the applicant was rude, used inappropriate language, gave negative responses or displayed unprofessional etiquette], flexibility [(S6) was the applicant accommodating in scheduling the interview?], geographic compatibility [(S7) does the applicant want to live and work in the area?], compatible career interests [(S8) if the applicant wants to be a lawyer and plans to go to law school within two or three years then his or her interest is not compatible with the management trainee program], a credit check [(S9) Ms. Morris said credit checks ended in 2011], a criminal background check [(S10) Ms. Morris said criminal background checks ended in 2016], a Maryland motor vehicle record [S11], a valid driver's license [S12], reference checks [S13] meeting the age requirement [S14], whether the applicant was reapplying within the required timeframe [S15], whether the applicant had the unrestricted right to work in the United States [S16] and whether the position was filled [(S17) Ms. Morris said, for example, the Eastern Shore Branch is three hours away and may not have an opening when an interested applicant applied]. (TR 302-308). An applicant could be rejected at the screening stage for not meeting any one or more of the criteria. (TR 303). PX 10-C also contains the "S" codes for fiscal year 2017 beginning on August 1, 2016. (TR 308).

Ms. Morris said PX 9 is an overview of the responsibilities of the recruiting coordinator position. She wrote the document and sent it to Mr. Arney, but she was not sure when that happened. (TR 309). One of the responsibilities is to manage the online application system and keep applicant flow logs updated. (TR 310). The applicant flow logs show the applicant's name, the job applied for (management trainees were job code 2 through FY08 and then became job code 4-B), how the applicant became aware of the position (for instance, by employee referral), self-identified race and gender, the course of action through the screening and interview process and then the final decision on whether the applicant was hired or rejected. "S" codes are included in the applicant flow log. (TR 312). Ms. Morris said management trainee positions are evergreen because Enterprise is constantly recruiting and hiring, and she said there was a high turnover rate for management trainees. (TR 313). Applicants are not compared against each other and if an applicant is qualified he or she will get a job offer. (TR 314).

Ms. Lichter screened applications as she pulled them off of the online system in fiscal years 2007 and 2008. It was a daily task and if she was not there another team member would do [*16] it. (TR 314). Ms. Hardesty reports directly to Ms. Morris and if Ms. Morris was out of the office then Ms. Lichter reported to Ms. Hardesty. Ms. Morris said PX 11 is the script for telephone screening. (TR 315). It is the script that Ms. Morris, Ms. Hardesty and Ms. Lichter used to conduct telephone screenings. The form states that it is for "individuals who meet the minimum criteria." Ms. Morris said that ideally the phone screen is done for applicants who meet all of the minimum criteria, but it could include some where there are questions about an applicant's qualifications. (TR 316). The form provides scripted questions that ask the applicant about his or her educational background, work experiences, interest in a sales career, experience in sales or customer service positions, how often the applicant changed jobs in the prior two years, whether he or she will work for the anticipated pay, his or her willingness to work the required hours and at the available area branch locations, whether he or she has a driver's licenses and about the applicant's criminal record. The person who conducts the interview can add notes at the bottom of the form and mark whether Enterprise is going to "pass" on the applicant and enter a disposition code giving the reason for rejection or schedule a first interview and note the date and time it will take place. The term "pass" means that Enterprise is not interested in pursuing the applicant and will pass on hiring him or her. (PX 11; TR 317-324). Ms. Lichter could see Ms. Morris's and Ms. Hardesty's calendars on her computer and she scheduled them to conduct first interviews with applicants who were continuing in the hiring process. (TR 325).

When an applicant comes in for a first interview, he or she begins by reviewing his or her application and making any corrections that are necessary, and by filling out a form prioritizing 13 job factor preferences from most important to least important. (PX 12; TR 325-326). Ordinarily an applicant would not be rejected based upon how he or she filled out the job factor preference form, but Ms. Morris recalled rejecting an applicant after having to give the applicant the same form six times because the applicant could not fill it out properly. (TR 327).

Ms. Morris and Ms. Hardesty used an interview evaluation form to conduct face-to-face interviews. (PX 13; TR 328). The form has headings for each of the six core competencies, definitions for each one, examples of effective and ineffective behaviors, a "No" and "Yes" option on whether the applicant met the competency, and space for comments. At the end, the interviewer can continue the applicant along in the process or reject the applicant. The form provides space to explain why an applicant was rejected. (TR 329-332). The same form is used for second and third interviews as well. (TR 331). A set of behavioral interview questions are included for those conducting second and third interviews in order to provide the interviewers a point of reference in addressing all of the core competencies. Those conducting second and third interviews are specifically trained on how to conduct behavioral interviews. They are not required to use the questions on the form, but they are required to ask similar questions. (PX 14; TR 332-333). A candidate who provided a response that showed a negative outcome on a core competency would be rejected. (TR 334-335).

Ms. Morris identified PX 15 as formal training provided to assistant managers on conducting behavioral interviews. It contains a leader's guide for those conducting the training and a participant's guide for those attending the training. Versions from 2005 and 2014 are included in the exhibit. (PX 15; TR 336-338). The training is presented by Ms. Morris and Ms. Hardesty and it lasts for about half a day. (TR 339).

[*17] PX 16 is a collection of documents. Ms. Morris identified page 1-16 as a PowerPoint presentation on conducting second interviews from an area managers meeting. (TR 339-340). Pages 17-18 are a training schedule for July 2007. Pages 19-29 are slides from a presentation Ms. Morris gave sometime prior to 2006. Pages 32-33 are a recruiting quarterly update showing women and minority hiring efforts that Ms. Morris thought was prepared around 2008. (TR 340). Page 35-36 is a top 10 list Ms. Morris prepared on the employee referral program. (TR 341). Page 37-47 are slides from a presentation on interviewing Ms. Morris gave at a women's forum for female Enterprise employees in March 2008. (TR 341-342). The exhibit also contains advertising materials, a telephone screening interview sheet, an interview evaluation form and a branch observation checklist. (TR 342-343).

Ms. Morris testified that she believed the interviewing presentation at the area managers meeting was in 2006. (TR 344). She said the phrase "second interview audits" on page 5 of PX 16 refers to the practice of she and Ms. Hardesty going out and sitting in on interviews with area managers once every two years. (TR 345). The bullets on the slide were things she wanted to call to the attention of the area managers. The bullet that said area managers spoke 70 to 75 percent of the time and candidates 20 to 25 percent was to encourage the area managers to make the interviews more conversational and probing. The bullet that said "Telling NOT selling the opportunity" was a reminder that it was a tight job market and the area managers should be selling the applicants on the benefits of working for Enterprise. The final bullet said there were inconsistent interviews from one area manager to the next, which Ms. Morris said was something she had observed. (TR 346). This type of training on how to conduct behavioral interviews was given regularly as refresher training for assistant branch managers, branch managers and area managers. (TR 347).

The slides for the recruiting department's mid-year summit begin on page 19 of PX 16. Ms. Morris did not recall how often the meeting was held, but a mid-year summit is what her general manager called it. Ms. Jill Trout, whose name is on the first slide, was Ms. Lichter's predecessor. (TR 348). One of the issues addressed was under 90 day retention, which Ms. Morris said was important to Enterprise. OFCCP's counsel asked what the bullet that said Enterprise was looking for the "right" candidate — with "right" being the only word in quotation marks in the bullet — and that it had been key in addressing retention meant. Ms. Morris said the "right' candidate meant a candidate who met the minimum qualifications. She said the slide that said minorities were at 21 percent in Group 18 and needed to be at 25 percent meant Group 18 needed to increase the percentage of minorities in the management trainee program to meet the average for the Metropolitan Statistical Analysis or MSA. (TR 350). Ms. Jill Kavalsky, who is listed as one of the presenters along with Ms. Morris at the women's forum in March 2008, was an area manager. (TR 351; PX 16 at 37). Page 53 of PX 16 is recruiting information that was placed on Monster.com to market the management trainee program. (TR 351-352).

Ms. Morris said PX 17 is an interview guide for management trainee candidates and that it was not something she used often. (TR 352-353). The document was created as a handout for Enterprise by SHL. (TR 353). One of the topics addressed in the document is cultural sensitivity and it notes, as an example, that in some cultures avoiding eye contact is a sign of respect. (TR 354-355). Ms. Morris said that she never conducted cultural sensitivity training, [*18] but a trainer does provide diversity training, which covers the same points. The document states that research shows that interviewers give higher ratings to candidates like themselves in terms of attitudes, gender and race, and lower ratings to candidates who are different. It says to be fair to candidates and avoid basing decisions on "gut feelings." (TR 355; PX 17 at 4). Ms. Morris said that she and Ms. Hardesty include training on bias research as part of their assistant managers training module. She said the training is intended to raise awareness so that interviewers are attuned to subconscious bias. (TR 355-356). In a part of the document on potential errors, it includes "similar to me error," which it says is judging more favorably individuals who are similar to the interviewer. Ms. Morris said this is part of the assistant managers training and it is a common interview error, although she noted that it is not unique to Enterprise. (TR 356).

PX 18 is a branch observation checklist that is given out for the second interview. (TR 358). Assistant branch managers normally conduct the branch visits and the checklist assists them in giving the management trainee applicant a tour of the branch. There is space at the end of the document for notes and it encourages the assistant branch managers to provide feedback on the candidates, including hiring recommendations. (TR 359-360). Assistant branch managers do not have authority to reject applicants or to extend job offers. PX 19 is an old master evaluation sheet. (TR 360). Ms. Morris was not sure how old the sheet was, but it had a one to five rating scale rather than the yes or no choices used now. (TR 361).

PX 20-A is an applicant flow log. (TR 361). The document has 10 headings: region, applicant name, complete date, position, gender, race, interview status, offer status, reason for decline, and hire date. (PX 20-A at 1). Region 1899 is Group 18 for the Baltimore region and 99 indicates the Group 18 headquarters. The complete date is the date the applicant completed the application. (TR 362). Under position, "trainee" indicates an applicant for the management trainee program. (TR 362-363). Under race, code D is for black or African-American and F is for white. The race codes and EEO codes are shown in PX 21. (TR 363). A zero under race meant the applicant did not provide race information. Also, the number 1 indicates white or other and 2 indicates black or African-American. Ms. Morris did not recall when the indicator switched from a letter to a numeral. (364). If an applicant failed to meet the minimum qualification for more than one core competency the flow log should indicate all of the reasons as an "I" code or an "S" code under the column entitled "reason for decline." (TR 367). PX 20-A is the applicant flow log for fiscal year 2007 and PX 20-B is the applicant flow log for fiscal year 2008. (TR 370). PX 20-C is a log of all expressions of interest in the management trainee program. (TR 372). PX 20-D is a spreadsheet the Enterprise corporate office prepared of all candidates that expressed interest in the management trainee program. (TR 373). PX 22 is an applicant flow log curated in an Excel spreadsheet. (TR 375).

Ms. Morris identified PX 60 as the application file for S.M.6 (TR 376). The first page was a rejection letter sent to S.M. on June 2, 2008 that was signed by Ms. Morris. Page three through eight is the online application that S.M. submitted. (TR 377). The application was printed out and the date 5/20/2008 appears at the bottom of each page. (TR 378). After counsel for both parties conferred, it was agreed that the application file for S.M. contained in DX 4643 [*19] (consisting of the same documents) would be used during Ms. Morris's testimony. (TR 378-379).

S.M. learned about Enterprise from a friend, S.G. (TR 379). S.M. was a liberal arts major at Morgan State University, which is among the historically black colleges and universities in the Group 18 area. Ms. Morris said she imagined that the percentage of applicants from Morgan State University who were African-American was "very high." (TR 380). There were some handwritten notes on the printed application. Ms. Morris said those were corrections S.M. made when she came in for the first interview and it reflected that she had graduated from college. Under employment information, S.M. said she interned at WMAR ABC 2 News. Ms. Morris said that may or may not be considered sales or customer service experience since there are a variety of jobs at a radio station and it may or may not involve doing sales calls to sell radio advertising. (TR 381). That was the type of thing Ms. Lichter might have investigated during a telephone screening. S.M. also worked as a hostess at T.G.I. Friday's in November 2007. Ms. Morris said she would consider that sales or customer service experience since it involved direct customer contact. S.M. interned at the Defense Security Cooperation Agency, which Ms. Morris said she would not consider sale or customer service since it appeared to be more of an administrative role. (TR 382). S.M. interned at WEAA 88.9 FM where she assisted producers, wrote public service announcements, responded to listener requests and assisted with on-air talk programming. (TR 382-383). Ms. Morris said she might consider that sale or customer service experience since she dealt with listeners. S.M. provided contact information for references, but Ms. Morris said references are not contacted until the background check stage, which takes place after Enterprise decides to extend a job offer. (TR 383). S.M. also provided information on how many hours per week she would like to work, her salary expectations, her current employment situation, and her driver's license. (TR 385-386).

S.M.'s resume is included in her application file. Ms. Morris said her work experience as a hostess at T.G.I. Friday's stood out. (TR 387). She would also credit her with customer service experience for working at McDonald's and possibly at the Victims Resource Center if she was working on a hotline or talking with people. (TR 388). Ms. Morris said the handwriting on the telephone screening form was not her handwriting, so Ms. Lichter or Ms. Hardesty did the telephone screening. Based upon the telephone screening, S.M. was scheduled for a first interview. (TR 389-390). The telephone screening form is included in the packet and it is available to the interviewer at the first interview. Ms. Morris interviewed S.M. and circled "no" under the core competency "persuasive" and wrote "very limited experience with lack of confidence." (TR 390; DX 4643 at 11). She also circled "no" under the core competency "leadership" and wrote "didn't give a leadership example more of a goal example." (TR 391; DX 4643 at 12). Because she received a "no" on two of the core competencies, S.M. was rejected. TR 391).

Ms. Morris reviewed the application file of J.Y. (TR 394; PX 70). She said J.Y.'s marketing internships and customer marketing while she was working at Bank of America stood out as potential sales or customer service experience. She said the applicant brings his or her resume to the interview. (TR 395). Ms. Morris said J.Y.'s telemarketing position stood out as sales experience. Ms. Morris interviewed J.Y. on October 23, 2006 and scheduled her for a second interview. (TR 396). The second interview was conducted by Mr. David Curry, who [*20] was an area manager. He marked "no" under the core competency "persuasive" and wrote "did not give examples of sales or results that were positive." He also marked "no" for "communicating orally" and wrote "did not answer the questions that were asked." (TR 397-398; PX 70). Mr. Curry rejected J.Y. and Ms. Morris sent her the rejection letter. The application package included a background check on J.Y. Ms. Morris said Enterprise starts a background check when an applicant is sent on for a second interview. (TR 398).

Ms. Morris reviewed the application package for I.J. (TR 399; PX 71). I.J. attended Morgan State University. His retail sales experience at Walmart qualified as sales or customer service experience. It appeared that Ms. Lichter did the telephone screening. She wrote on the screening form that I.J. worked at the customer service desk at Walmart and did outbound cold calls at MCI. Ms. Morris said that would count as sales and customer service experience. (TR 400). Ms. Morris conducted the first interview and circled "yes" on all six of the core competencies. (TR 401). I.J. did a branch visit with area manager David Curry and assistant manager Mark Richardson on May 30, 2007. At the end of the Branch Observation Checklist, someone wrote, "David and Mark — see the personality!!" (TR 402; PX 71 at 13). Ms. Morris said she was sure that meant he came in and was enthusiastic. The second interview was conducted by Mr. Curry the same day as the branch visit. He circled "no" under the core competency "persuasive." He wrote, "He did not close the deal with Mark or myself that he would be competitive and aggressive in our environment." He also circled "no" under "communicating orally" and wrote "he did not answer the questions directly and did not bring any energy or motivation to the interview. His personality did not shine or (illegible word) out during the interview based on the behavior questions you provided us to ask." (TR 402-404; PX 71 at 14-15). Mr. Curry rejected I.J and wrote at the end of the form, "He did not sell us on his personality, motivation and energy to work at Enterprise." (TR 404; PX 71 at 15).

Ms. Morris reviewed the application package for R.Z. (TR 405; PX 72). His job as an insurance agent would qualify as sales and customer service experience. (TR 405). He attended Washington College, which is not a historically black college or university. The telephone screening form showed that he worked at a high-end grocery store, he is good with people, he started a Frisbee team at Washington College and recruited players, and he was on the college's sailing team. (TR 406). Ms. Morris said that R.Z.'s experience managing his band, which was listed on his resume, where he landed more than 20 gigs in Maryland and Delaware and negotiated the band's contracts, would be qualifying experience. Ms. Morris conducted the first interview and passed I.Z. along to a second interview with Mr. Curry. She noted under "persuasive" that he booked gigs for his band and under "leadership" she noted that he was captain of his high school lacrosse team and was the leader of his band. (TR 407). Mr. Curry rejected R.Z. He circled "no" under the core competencies of "customer service" and "persuasive." Under the reason for rejection, Mr. Curry wrote that R.Z. "did not provide any examples on our persuasive questions that connected him to be successful at Enterprise" and "did not provide any examples of customer service he did or could provide to us." (TR 408-409; PX 72 at 14).

Ms. Morris reviewed the application package for B.M. (TR 413; PX 99). The only employment listed on his application was working as an office assistant at a law firm for nearly four years where he said he maintained and retrieved case files and performed general office [*21] upkeep. (TR 414-415; PX 99 at 3, 5). His resume showed that he hosted and produced radio programs, directed a university television program and managed a gas station where he handled closing at night and overseeing employees. (PX 99 at 13). Ms. Morris specifically recalled B.M. because she said she "met him at Salisbury University quite a bit." (TR 415). There was no telephone screening form in his application package and Ms. Morris believed she may have done an on-campus interview instead. The final interview was by group rental manager Kevin Cutron and B.M. was offered a job as a management trainee. (TR 415). The letter extending the job offer was dated June 23, 2008 and was signed by Ms. Morris. (PX 99 at 31).

Ms. Morris said the impact ratios are automatically calculated by a computer program. She plugs in the numbers from the applicant flow log and the program does the calculations. She did not know the name of the program and she said it came from the Enterprise corporate headquarters. (TR 416-417). Ms. Morris agreed that the impact ratio analysis at the end of fiscal year 2009 showed there were 170 applicants for the management trainee program who disclosed their races, 14 minorities were hired out of 71 minority applicants, 33 whites were hired out of 99 white applicants, and the impact ratio was .59, which meant that African-American applicant were hired at 59 percent of the rate at which white applicants were hired. (TR 416-420; PX 24 at 2). She acknowledged that she was the person who generated the reports for each of the years reflected in PX 24 and that fiscal years 2007 and 2008 are not included in the exhibit. (TR 420-421). Ms. Morris reviewed PX 110 and agreed that it was the affirmative action plan for fiscal year 2008. (TR 422-423). She reviewed PX 111 and agreed that it was the affirmative action plan for fiscal year 2007. (TR 423). She acknowledged that neither of the plans included impact ratio analysis data. (TR 424). The affirmative action plans were prepared by Mr. Arney in conjunction with Outsoft, and approved by Mr. Arney and Mr. Molitor. (TR 425).

The impact ratio analysis for fiscal year 2008 showed 114 management trainee hires out of 672 applicants who identified their races. There were 34 minority hires out of 320 minority applicants and 80 white hires out of 352 applicants. The impact ratio was .47, which meant a white applicant was twice as likely to get hired as a minority applicant, and the standard deviation was 4.17, which the program said was a "yes" for adverse impact. The impact ratio analysis for fiscal year 2007 showed 152 management trainee hires out of 619 applicants who identified their races. There were 73 minority hires out of 361 minority applicants and 79 white hires out of 258 applicants. The impact ratio was .66 and the standard deviation was 2.96, which the program said was a "yes" for adverse impact. (TR 426-429; PX 112). Ms. Morris said they did not want a standard deviation over 2.0 and they looked at their processes to see if there was anything that needed to be corrected. She said there were no corrections made as a result of the 2007 or 2008 data. She said they were well ahead of the MSA every year and they could not pinpoint anything that needed to be changed in their hiring process at that time. (TR 430). Enterprise did not have hiring quotas, but it had a hiring goal of at least 80 percent of the MSA ratios. (TR 431-432).

On cross-examination by Enterprise, Ms. Morris said that sales and customer service experience are important factors in selecting people for the management trainee program because those are the two key performance metrics for advancement. Even though new hires get classroom training and on-the-job training, they need a foundation in sales and customer service coming into the job. (TR 433-435). She said that the management trainee position is an [*22] evergreen position because Enterprise is constantly recruiting and hiring, and they did so through the recession, although the hiring numbers dropped off during that period. (TR 436).

Ms. Morris said PX 20-D is not an applicant flow log, it is a spreadsheet showing information for all of those who expressed interest in the management trainee program. It pulls the information from their formal applications into a spreadsheet. (TR 438).

Ms. Morris remained on the witness stand for direct examination as a witness for Enterprise. She said that while the screening criteria for the plan year that began on January 1, 2006 said "work experience in sales and/or customer service," actually both were required and that criteria was better defined later on. (TR 441-442).

DX 75 is an excerpt of the minimum qualifications each year for the management trainee position. (TR 443). Ms. Morris said Enterprise kept trying to get more refined with its work experience requirement. In fiscal year 2015, for instance, the requirement was for one year of experience in two of the following: sales, customer service, military leadership and college or professional athletics. (TR 445). Between 2006 and 2017, the criteria for hiring for the management trainee position did not change on the basis of race.

Applicants are asked to voluntarily identify their races on their electronic applications. It helps Enterprise make sure it has a diverse workforce and it is required by OFCCP. (TR 446-447). In the 2006 to 2008 period when the SHL system was used, the corporate office prepared the flow logs and sent them out the following month. Ms. Morris and her team did not have the self-identified race information at the time they screened or interviewed applicants. (TR 447).

While there were two application tracking systems that Enterprise used during the relevant period — SHL and ICIMS — the initial application screening process and the telephone interview process remained virtually the same. (TR 448). The SHL system required someone to go online every day and review the applications. The ICIMS system is more robust. It checks to see if the applicant meets the minimum qualifications and, if not, it generates a message to the applicant thanking him or her for applying and advising the applicant that he or she did not qualify — referred to as a "DNQ" — and explains the reason why the applicant was rejected. It also informs the applicant that if the information that produced the DNQ is incorrect to contact Enterprise and the application can be corrected. (TR 449). Ms. Morris said that an applicant can be rejected in the initial screening stage for a lack of sales or customer service experience based solely on his or her occupation if it is apparent that it is a purely administrative or technical job. She said race was never the basis for any decision in the initial screening stage and the race of the applicant was unknown when the telephone screening was conducted. (TR 450). She said race was not a factor in any decision made at the first in-person interview stage. (TR 451-452).

Ms. Morris said she was a branch manager before she moved into human resources. The branch manager runs the operation at his or her branch. It is a profit driven business, so managing and developing the staff to deliver sales and customer service is crucial. There are 50 to 60 branches in Group 18 at any given point in time. (TR 452-453). An area manager is responsible for overseeing three to five branch locations. They are focused on team development [*23] in order to deliver results and achieve goals as well. There are 11 to 15 area managers in Group 18 at any given point in time. (TR 453).

In the first, second and third interview stage, it is possible for someone to get passed along to the next stage and then get rejected. (TR 453-454). Enterprise determined that putting applicants through three interviews was making the process too long and they went to just two interviews starting in November 2014. The third interview was with a group manager. A group manager oversees the airport and entire cities with 30 branches and multiple areas. There are normally three group managers in Group 18. (TR 456-457). Ms. Morris was not aware of any branch, area or group manager making any screening decisions on the basis of race. (TR 457).

In the background check stage, Enterprise looked at the applicant's driving record report, criminal background, credit report, and made sure the social security number matched. Enterprise stopped checking an applicant's credit report in 2011. When credit reports were a screening factor, Enterprise looked at the applicant's history over the prior three year period to see if he or she had three or more delinquent accounts. Sometimes a credit report problem would be due to a major life event, such as a divorce. Ms. Morris said that if she went to Mr. Molitor to get approval to forward on an applicant with a bad credit report he knew that 90 percent of the time it was a female applicant or a minority applicant. (TR 459). She said there were no variances at all at any stage in the screening process based on race. (TR 460).

PX 15 is the training material for new assistant branch managers on how to conduct interviews. (TR 460). It introduces them to doing behavioral interviews and stresses the importance of consistency. (TR 461-462). Ms. Morris said she engrains in the assistant branch mangers to only ask questions that are job related and non-discriminatory, and to document their hiring decisions. (TR 462). Part of the training includes an exercise where they go around the room and give examples of interviewing errors based on age, race, religion, sexual orientation, place of origin, veteran status and disabilities. (TR 463). Assistant branch managers are trained to use the STAR method and ask questions that lead the applicant to talk about the situation, the task, the action and the result. (TR 464). Ms. Morris personally conducts the assistant branch manager training along with Ms. Hardesty. (TR 467).

DR. SALLAMAR WORRELL

Dr. Sallamar Worrell is an operation research analyst at the U.S. Food and Drug Administration ("FDA") in the Center for Prescription Drugs. She works with FDA management to help them understand their data and their mandates from Congress. She makes sure their information systems capture the data required and allows them to do data-driven analytics. She is not a direct supervisor, but she does lead interdisciplinary teams on specific initiatives. (TR 470-471). She has about 10 years of federal service, including 9 months with the Navy and the rest with the FDA. She has a doctorate in engineering and engineering management from the George Washington University. (TR 471). Engineering management is similar to a master's in business administration in that it focuses on running an enterprise efficiently, but it is more technical in nature. (TR 471-472). She received her doctorate in May 2017. Dr. Worrell also has a master's in systems engineering from the George Washington [*24] University and a bachelor's degree in applied mathematics and statistics from Johns Hopkins University. She graduated from Johns Hopkins in 2008. (TR 472-473).

Dr. Worrell applied for the Enterprise management trainee program in 2008 prior to graduating from Johns Hopkins. (TR 473). PX 90 is a copy of the application Dr. Worrell submitted to Enterprise. She learned about the position on the internet. (TR 474-475). She said she had rented a car from Enterprise before, the job seemed like good training for a career, she believed she was qualified, and it was something that was of interest to her. (TR 475). Dr. Worrell listed Omnitest on her application as a previous employer. She said she worked there tutoring students on the SAT and ACT as well as mathematics, chemistry and physics. She started out tutoring at the Omnitest site and then graduated to traveling out to schools to conduct tutoring sessions. This was a part-time position she held while she was a student at Johns Hopkins. (TR 476-477). Dr. Worrell worked at Sprint Nextel as a retail sales consultant at their store in Columbia Mall. She sold telephones and cellular plans and assisted customers with problems with their phones and plans, and she tried to upsell them to a better plan that would better serve their needs. (TR 477-478). Dr. Worrell worked for Spring Nextel from December 1, 2006 to February 1, 2007. She started work for Omnitest on February 1, 2007 and was still working for them when she applied to Enterprise on April 25, 2008. (TR 479-480; PX 90). Dr. Worrell worked for HEBCAC Youth Opportunities Baltimore tutoring adults and students for the GED from October 1, 2006 to December 1, 2006. (TR 480; PX 90). She was a research assistant at the Johns Hopkins Bloomberg School of Public Health from June 1, 2004 to July 1, 2005 where she assisted with a study on the correlation between lung size and a predisposition to contract emphysema. She was a tutor at Rapid Success Tutoring where she provided in-home tutoring in math and science to students. (TR 481; PX 90).

Dr. Worrell recalled her telephone interview with Enterprise, although she said she did not remember all of the details. She recalled that it seemed to get off to a good start, but when she got off of the phone she thought to herself that it had not ended well and she wondered if she had done something wrong. (TR 482). She said she prepared for the telephone interview by following the advice she got from the university's career center to research the company and ask the interviewer probing questions about the company. (TR 482-483). Dr. Worrell said she thought her work as a tutor where she dealt directly with students and their parents and her sales experience at Sprint Nextel were examples of sales and customer service experience that addressed the interview question exploring that core competency. (TR 483-485).

On cross-examination by Enterprise, Dr. Worrell said she started her master's program in 2012 and graduated in 2014. (TR 486). She said she did not recall the specific questions she asked the telephone interviewer, but she had looked at the Enterprise website and tried to ask questions to figure out how she would fit into the Enterprise mission and to get a better idea of the day-to-day activities the job involved. When asked why she remembered the responses but not the questions, she said it was because it was one her first interviews and it was awkward. She said she felt as if she had done something wrong and "[i]t was very unsettling to me, so it stuck in my memory." (TR 487). Dr. Worrell said she felt like the interviewer shutdown in the latter part of the conversation and was trying to rush through. She did not know what caused that to happen and she did not know if the interviewer was aware of her race. (TR 488). She said about 75 to 80 percent of her time was devoted to sales when she worked for Sprint Nextel. The [*25] application asked for the highest annual wage earned for each job, so the $23,000 figure she provided for Sprint Nextel was what she believed she would have earned in a year, not what she actually earned in the three months she worked there. (TR 489). Dr. Worrell did not know why Enterprise chose to reject her application. (TR 494).

TERESA L. MORRIS

(Continuation of Testimony)

Direct examination of Ms. Morris by Enterprise continued. DX 81 is the last applicant flow log generated under the SHL tracking system. (TR 505-506). At that time, data generated by SHL went to the Enterprise corporate office and the following month the flow log arrived at Group 18 headquarters. Ms. Morris and her staff entered some information on the applicant flow log, specifically in Column M they entered interview status, in Column N they entered offer status and in Column O they entered disposition codes. (TR 507). Under disposition codes in Column O, some codes begin with an "S" and some with an "I" indicating the applicant was rejected at either the screening stage or the interview stage. There are also some "SSO" codes for applicants who chose to self-select out of the hiring process. (TR 508-509).

Ms. Morris reviewed DX 191 and identified it as a document she prepared that shows a breakdown of the disposition codes. (TR 509).

The S1 code was applied if an applicant did not meet the minimum education requirement. (TR 510). DX 6298 is the application file for L.S. Ms. Morris conducted the telephone screening and entered an S1 code on L.S.'s application because she did not have a college degree. (TR 511).

The S2 code was applied when an applicant's employment history showed a lack of job stability. (TR 514). DX 727 is the application file for J.B. Ms. Morris entered an S2 code on J.B.'s application because he was terminated from a job at Walmart for a misdemeanor crime committed while he was working there. She said someone getting fired from a job might be a reason to use the S2 code and it would definitely be used if someone committed a misdemeanor while on the job. (TR 514-515).

The S3 code was applied if the applicant lacked the required sales and customer service experience. The minimum requirement changed over time and increased from the initial six months that was required in 2006 to 2008. (TR 516). DX 3475 is the application file for B.K. Ms. Morris conducted the telephone screening and wrote on the application that B.K. said "I do not have any sales experience" and entered an S3 code. (TR 516-517). DX 3384 is the application file for J.J. Ms. Morris did the telephone screening and wrote on the form "Could not note any sales experience" and she entered an S3 code. (TR 517-518).

The S4 code was applied when the quality of the application was poor, notably the spelling and grammar. Ms. Morris said it was used when an application contained a lot of typographical mistakes. (TR 519).

[*26] The S5 code was applied if the applicant lacked professionalism, notably rude communication whether by telephone or email. DX 6795 is the application file for C.W. Ms. Morris conducted her telephone screening and scheduled her for an interview. She noted on the application that the interviewed had to be rescheduled several times and C.W. did not handle the scheduling professionally. She entered an S5 code on the application. (TR 520-521).

The S6 code was used when an applicant was not flexible and was not accommodating in setting up the telephone screening or the first interview. Ms. Morris said the S6 code was not used very often. (TR 521).

The S7 code was used when an applicant was not compatible with the geographic region that Group 18 encompassed. Ms. Morris said if that was the case, they tried to put the applicant in contact with the region that covered the applicant's area of geographic preference. (TR 522). DX 5680 is the application file for B.S. This was an applicant they liked, but during the second interview he said he wanted to move to California. Ms. Morris said they forwarded his application to Group 32 in California and entered an S7 code for the disposition of the Group 18 application. (TR 523).

The S8 code was applied when an applicant's career goals and interests were not compatible with a career at Enterprise. DX 1364 is the application file for B.C. who said she wanted a career working in rehabilitation. (TR 524-525).

The S9 code was used when an applicant had a poor credit history. Ms. Morris said Enterprise stopped running credit reports on applicants in 2011. (TR 525). DX 6307 is the application file for P.S. who was rejected for a poor credit history. (TR 526-527).

The S10 code was applied if the applicant had a felony conviction that was related to the requirements of the job. Ms. Morris said just a week before the hearing she had gotten an applicant approved who had a criminal conviction on his record. (TR 528). DX 461 is the application file for S.A. He made it to a second interview, but his criminal records check showed a felony conviction for possession of marijuana with intent to distribute and his application was rejected. (TR 529-530; PX 461 at 21-22).

The S11 code was applied when an applicant's motor vehicle record showed more than two moving violations within the previous three years or a drunk driving offense within the previous five years. (TR 531). DX 513 is the application file for L.B. whose driving record showed two speeding violations and an improper lane change violation within the previous three years. (TR 532).

The S12 code was applied if an applicant lacked a valid driver's license. DX 4188 is the application file for T.M. She was rejected because she did not have a driver's license. (TR 533-534).

The S13 code was applied if an applicant's background check was unfavorable. Ms. Morris said it was a code that was not used often. (TR 535-536).

[*27] The S14 code was used when an applicant did not meet the age requirement. Ms. Morris could not recall ever using the S14 code. (TR 536).

The S15 code was used when an applicant reapplied for the management trainee program too soon after having been previously rejected. She said they normally want an applicant to wait at least six month after a rejection before applying again. (TR 536-537).

The S16 code was used when an applicant was not legally eligible to work in the United States. It was not a code that was used very often. (TR 537).

The S17 code was used if the applicant was applying for a position that had already been filled. Ms. Morris said that the code that was not used often, but it was used during the recession when there were not many job vacancies to fill and it was used at times when an applicant was interested in the Maryland Eastern Shore area where there were not many Enterprise locations or positions to fill. (TR 537).

Ms. Morris testified that "I" codes were used to identify the reason an applicant was rejected at any stage of the interview process. The I1 code was used if an applicant lacked communication ability: made grammatical errors, was unable to articulate thoughts clearly, displayed poor body language or rambled off topic, for example. (TR 538). DX 4607 is the application file for J.M. Ms. Morris conducted the first interview of J.M. and rejected him and assigned an I1 code for, among other reasons, giving short, abrupt answers and never smiling. (TR 539-540).

The I2 code was used if an applicant lacked leadership ability. Ms. Morris said that when Enterprise hires management trainees it is looking to develop future company leaders, so they look for applicants who have taken lead roles in groups or on teams, or have successfully held supervisory positions or management roles. DX 2015 is the application file for J.E. Ms. Morris conducted the first interview of J.E. and rejected him and assigned an I2 code because he was hesitant in responding to her questions and he did not provide a good example of a personal leadership accomplishment. (TR 542).

The I3 code is used if an applicant lacked customer service ability. DX 4289 is the application file for I.M. Ms. Morris conducted the first interview of I.M. and rejected her and assigned an I3 code because she did not give a good example of how she had dealt with a difficult customer. (TR 543-544).

The I4 code was used if an applicant lacked persuasiveness. Ms. Morris said persuasiveness is important because making sales is a management trainee's primary mission. She would ask an applicant to give her an example of when he or she was able to overcome an objection and turn a no into a yes. (TR 545).

DX 5396 was the application file for A.R. Ms. Morris conducted his first interview and rejected him for failing to meet five of the core competencies. Ms. Morris assigned I codes showing A.R. was rejected for a lack of leadership and a lack of communication skills. (TR 546-549).

[*28] DX 4443 was the application file for C.N. Ms. Morris conducted his first interview and rejected him for not providing a good example of customer service, a lack of persuasiveness, a lack of leadership and poor communication skills. (TR 549-550).

DX 3398 was the application file for M.J. Ms. Morris conducted her first interview and noted that she failed to give specific examples of how she performed several of the core competencies, saying instead that "she does it all the time." Mr. Morris rejected M.J. and assigned an I3 code. (TR 550-552).

The I5 code was used if an applicant was not flexible. Ms. Morris said Enterprise looked for applicants who were able to adapt, accept change and be flexible in setting up interviews. DX 6303 is the application file for T.S. Ms. Morris conducted his first interview and marked "no" for flexibility because he said that if he was hired he would need two or more hours off on Mondays to take a class. She used the I5 code as the reason for rejection. (TR 552-554).

The I6 code was used if an applicant did not display a results-driven work ethic. DX 3273 is the application file for B.J. Ms. Morris conduct his first interview and marked "no" for results-driven and communications. She said he was interested in moving laterally and not upwards and he spoke very methodically and without smiling, which were important traits in a customer service industry. (TR 554-556).

The I7 code was used if an applicant did not have a compatible career direction. Ms. Morris said this code was applied if the applicant indicated he or she was interested in pursuing a career in something other than sales. (TR 556).

The I8 code was used if an applicant did not display professionalism. Ms. Morris said this code was applied if an applicant was not punctual, did not follow up after missing an interview, was poorly dressed or groomed, displayed bad manners or etiquette, or was not well-spoken. (TR 556).

The I9 code was used if the applicant was not prepared for the interview. Ms. Morris said an applicant should have some basic knowledge about the company and have some questions to ask about the position, particularly since so much information is readily available online. (TR 556-557).

The I10 code was used if any applicant did not show a sincere interest in the position. Ms. Morris said this was very similar to some of the other reasons for rejecting an applicant, like compatible career direction. (TR 557).

Ms. Morris testified that she never took race into account when she applied any of the "S" or "I" codes, no supervisor had ever instructed her to do so, and no Enterprise policy instructed her to do so. (TR 557). She said Enterprise attempts to recruit minority applicants and that she and her team take pride in their minority recruitment efforts, particularly their efforts at the area's historically black colleges and universities. (TR 557-558). Ms. Morris said Enterprise is an equal opportunity employer and had provided her with training on anti-discrimination [*29] and bias. (TR 559-560). She said she had "[a]bsolutely not" made any decisions concerning the hiring of management trainees based upon race and that she had never entered a disposition code into an applicant flow log that did not reflect the true reason the applicant was rejected. (TR 560).

On cross-examination by OFCCP, Ms. Morris said that until fiscal year 2015 she did not have access to an applicant's race until she received an applicant flow log the month after the submission of the application. Beginning in fiscal year 2015, the tracking software had a dropdown menu that would display race, but Ms. Morris said she never used it and waited instead for the applicant flow log to come in the following month. (TR 561). No one else had access to the race data until Ms. Hardesty was promoted to manager sometime within the past year. Ms. Morris said she knew what implicit bias meant, but she had not had any formal training on it and she had not taken an implicit association tests. (TR 562). She had provided training to area managers on unconscious bias within the previous 90 days. Ms. Hardesty assisted her in conducting the training. The purpose of the training was to make area managers aware of the issue and to stress sticking to the core competencies. (TR 563-564).

Ms. Morris said that in the training she provides on conducting behavioral interviews she explains that as a federal contractor Enterprise has a duty not to discriminate in hiring. (TR 565-566). One of the common errors discussed as part of the training is judging more favorably people who are similar to the interviewer. Ms. Morris said that she is white and her judging white people more favorably would be an example of a "similar to me" error. (TR 566). The training for managers includes avoiding stereotyping. She advises them to make decisions based upon the facts rather than upon personal biases. The training she did recently for area managers on unconscious bias included being aware of cultural bias; for example, recognizing that direct eye contact is considered inappropriate in some cultures. She did not recall cultural bias being specifically included in other training she conducted between 2007 and 2017. (TR 567).

Ms. Morris acknowledged that even if Enterprise exceeds African-American hiring in comparison to the metropolitan statistical area it would still be inappropriate to reject African-American applicants on the basis of race. Ms. Morris was a branch manager for about eight months in 1992 and several management trainees in her branch were African-Americans. She said at the time she was not involved in the interviewing and hiring process. (TR 570-571). Ms. Morris participated in a conference call with the Enterprise corporate office and went over the training module that she recently presented to the Group 18 area managers. Group 18 keeps track of its diversity hiring goals through a monthly report from the corporate headquarters that shows how they compare to the MSA.

In response to questions from me, Ms. Morris said that any sales experience would qualify under the criteria mandating relevant sales experience. With respect to applicant I.J. (PX 71) — who area manager Curry rejected for lack of personality, motivation and energy despite Ms. Morris sending him through for a second interview and with positive comments about his personality during his branch visit — I asked if there were any cross-checks in cases where there were apparent inconsistencies in the way applicants were assessed. (TR 572). Ms. Morris said that when she saw inconsistencies she picked up the telephone and initiated a conversation about it. (TR 572-573). She agreed that generally there was about an even split between white and [*30] minority applicants and about a 60/40 gender split with more male applicants than female applicants. Given that, I asked her how she squared minorities and women accounting for half or less of the applicant pool with her testimony that 90 percent of the time when she went to Mr. Molitor for approval of an applicant with a bad credit report it was for a minority or female applicant. She said there had only been a few times when she went to Mr. Molitor for such approvals, but when she did 90 percent of the time it was for a minority applicant or a female applicant. (TR 573).

ELIZABETH HARDESTY

Ms. Elizabeth Hardesty started her career at Enterprise as a management trainee in 2000 right after she graduated from college. (TR 579-580). She was a management trainee for about a year and then an assistant branch manager for a few months. (TR 580). To become an assistant branch manager, a management trainee has to reach certain performance metrics and then pass a management qualification interview ("MQI") with an area manager. (TR 580-581). The performance metrics include an Enterprise Service Quality Index ("ESQI"), which is a team score based upon surveys of Enterprise customers who rent vehicles from a branch. (TR 581). It also includes sales metrics, which includes how well a management trainee does upselling rental vehicles and ancillary products like insurance, prepaid fuel and satellite radio service. (TR 581-582). The sales metric is an individual score that measures Total Additional Revenue ("TAR") the management trainee adds to a customer's total. (TR 582). All of the management trainees and management assistants in Group 18 are ranked from first to last on one matrix and all of the branch managers and assistant branch mangers are ranked on another matrix. The rankings come out on a monthly basis and they are distributed to the entire group, so every management trainee knows exactly where he or she stands in relation to his or her peers. (TR 584). Ms. Hardesty said the rankings were not prepared by her department and she was not familiar with the specifics on how they are compiled, but she said the ESQI, a sales metric, the TAR and at one point the percentage of contracts where a protection plan was added were all included. In order to be eligible to take the MQI, a management trainee has to rank in the top half of his or her peers for a certain number of months. (TR 586). Ms. Hardesty thought the minimum was a top half ranking for six months and she said six months was the current standard. Management trainees who do not reach that standard by the 15 month point are placed on probation and may be terminated for performance reasons or transferred other divisions where their talents are better suited. (TR 587).

Ms. Hardesty progressed from being a branch manager to a recruiting supervisor in August 2002. She reports to Ms. Morris. After Ms. Lichter left, Ms. Hardesty performed some of the initial applicant screenings. (TR 589). She said with the SHL tracking system she had to manually enter disposition codes, but ICIMS has a drop-down menu where she just clicks on the appropriate choice. (TR 590). The disposition codes remained the same, more or less. (TR 591). She said one purpose of the telephone screening is to convey information to the applicant, like the job responsibilities, compensation, hours and opportunity for advancement. (TR 591-592). Another purpose is to determine whether the applicant meets the minimum qualifications. Ms. Hardesty said they give applicants the benefit of the doubt if it is unclear whether a qualification is or is not met. For instance, if the application says the applicant has sales and customer service experience and it appears that it might satisfy the minimum requirement, [*31] Enterprise would want to talk with the applicant to see if the minimum requirement is in fact met. (TR 592). Ms. Hardesty said she started doing applicant screening as a primary function sometime in the fall of 2009 after she returned from maternity leave. (TR 593). She did not recall how much sales and customer service experience was required to meet the minimum requirement at that time. (TR 593-594).

Ms. Hardesty said customer service experience could come from a variety of types of jobs and it really depended on whether the applicant was providing "customer-facing" service or service over the telephone. (TR 594). Sales experience would depend on whether the applicant was just taking the customer's order or trying to direct the customer towards a purchase decision. A cashier at McDonald's would most likely be the former, but could be the latter if he or she tried to get customers to add French fries or order a larger drink. (TR 595-596). She said once she entered a disposition code for an applicant she had rejected she could not recall anyone every reversing her decision. (TR 596-597). If she determined that an applicant did not meet the minimum requirements the applicant would receive a rejection letter and sometimes she was the one who sent the letter. (TR 597). Under ICIMS, the computer scans applications and sends a rejection email if one or more of the minimum requirements is not met. Ms. Hardesty said that often an applicant rejected by ICIMS will email her and say he or she made a mistake on the application and he or she actually meets the requirement. In that case, the application moves forward. If the rejected applicant does not contact Enterprise then no human being ever sees the application. (TR 597-598). If there is more than one reason for rejecting an applicant Ms. Hardesty said she selects the main reason since ICIMS only allows her to enter one disposition code. She said it was a discretionary decision and if a person did not meet the education requirement and did not have a driver's license it was irrelevant which disposition code she picked because the two reasons are of equal importance. (TR 599). ICIMS allows the entry of more than one disposition code for applicants rejected during the interview phase. (TR 600).

Ms. Hardesty reviewed PX 79, which is the application file for D.Z. (TR 601). She said he played on the lacrosse team and often that involves recruiting team members, so that might qualify as sales experience. (TR 601-602). She said he worked as a doorman at a couple of bars and as a parts consultant at a car dealership, and those jobs would involve customer service. (TR 602). Ms. Hardesty also noted that D.Z. was a business major and that may involve classes in sales, which might count as sales experience, too. (TR 603). Ms. Hardesty did the first interview for D.Z., found that he met all six of the core competencies and sent him on to a second interview with Mr. Curry and Mr. Richardson. (TR 605-606). Mr. Curry and Mr. Richardson found that D.Z. met all six of the core competencies and sent him on to a third interview with Mr. Andy Wendler, who was a Group 18 rental manager. (TR 606). Mr. Wendler found that D.Z. did not meet three of the core competencies: results-driven, leadership and communicating orally. (TR 607-608). Ms. Hardesty agreed that reasonable people can come to different conclusions on whether an applicant met the core competencies and she said "to a certain extent, all interviewing is subjective." (TR 608).

Ms. Hardesty said that she had conducted first interviews from the fall of 2009 to the present. (TR 611). She said she was trained by Ms. Morris on conducting behavior-based interviewing. (TR 612).

[*32] Ms. Hardesty reviewed PX 65, which is the application file for C.G. (TR 614). She said he worked a phone-a-thon and had tutored, and those might qualify as sales or customer service experience. (TR 615). Asking him to describe his sales and customer service experience would be a screening question to assess whether he met the minimum requirement and then asking him to describe the details of his employment would be an interview question to determine whether he met the requirements of the core competencies. (TR 616). Ms. Hardesty did the first interview for C.G. and she marked "no" on customer service and persuasive. Her notes said that in the customer service example C.G. gave he did not resolve the customer's problem and in persuasiveness she wrote, "Didn't seem confident. No eye contact when discussing sales."7 (TR 618). As a result, C.G. was rejected after the interview with Ms. Hardesty. (RT 618-619).

Ms. Hardesty reviewed PX 66, which is the application file for K.K. (TR 619). She said K.K. had a degree in communications with a concentration in public relations, so she assumed K.K. had classes related to dealing with the public. K.K. worked with some non-profit groups and helped organize community service events. She also had experience selling snowboards and ski equipment at a sporting goods store. All of that experience together would make K.K. worthy of an interview. (TR 619-620). Enterprise's counsel said that it appeared PX 66 was not the complete application file for K.K. and asked to switch to DX 3628 instead, which is the same application. OFCCP's counsel agreed to do so. (TR 621). Ms. Hardesty conducted the first interview of K.K. She marked "no" for customer service and noted that the example K.K. gave was her involvement in protesting a company. (TR 622). She marked "no" for leadership and noted that K.K. did not give an example of leading. She also marked "no" for communicating orally and wrote that K.K. went off on tangents. K.K. was rejected and the disposition codes for communication (I1), leadership (I2) and customer service (I3) were entered. (TR 624).

Ms. Hardesty reviewed PX 67, which is the application file for A.J. (TR 624). Ms. Hardesty conducted the first interview for A.J. and sent him on to a second interview with Mr. Curry and Mr. Josh Kitchens. They sent him on to a third interview with Mr. Wendler and he approved A.J. for a job offer. (TR 626). A.J. had a three-month summer internship where he did door-to-door sales and he was a resident hall advisor and resident hall director for two years when he was a student at Virginia Tech. (TR 627; PX 67 at 2-3).

Ms. Hardesty reviewed PX 68, which is the application file for B.K. (TR 627). On the telephone screening form, in the notes following the question about sales and customer service experience, it states: "Healthcare Parking — started as valet worked his way up to asst manager — heavy. Venture Packaging — cust follow up, solve cust problems. Tennis @ Roanoke College 4. Team Captain Sr year." (TR 628; PX 68 at 21). Ms. Hardesty said being on the tennis team might qualify as sales experience depending on how it was communicated since it may involve recruiting players. Also, valets work for tips, which is not sales, but is customer service. She said this would, in her view, meet the minimum requirement for sales experience. (TR 628). [*33] Ms. Hardesty conducted the first interview for B.K. and found that he met all six of the core competencies. She sent him on to a second interview with Mr. Curry and Mr. Kitchens. They marked "no" for persuasive and wrote that B.K. had no sales experience, preferred customer service to sales, and did not grasp the concept. They also marked "no" for results-driven and wrote that he did not have a strong awareness of personal performance and he set goals for himself that he did not achieve. (TR 629-630; PX 68 at 14-15). B.K. was rejected and Ms. Hardesty said the I4 and I6 disposition codes would apply. (TR 630).

Ms. Hardesty reviewed PX 69, which is the application file for R.K. (TR 630). He had worked as a specialty sales associate at Home Depot from February 2006 until the time he applied to Enterprise and was interviewed. He went through three interviews and was offered a position as a management trainee. (TR 631-634; PX 69 at 6, 17).

Ms. Hardesty reviewed PX 76, which is the application file for E.M. (TR 634-635). Enterprise's counsel said that it appeared PX 76 was not the complete application file and asked to switch to DX 4125 instead, which is the same application. OFCCP's counsel agreed to do so. (TR 636). E.M. was working as a regional loss prevention manager at Abercrombie & Fitch when he applied to Enterprise. (TR 635; DX 4125 at 3-4). His duties included recruiting, hiring and training loss prevention staff. (PX 4125 at 8). He had previously worked in asset protection at Target from 2003 to 2007. (TR 635-636; DX 4125 at 8-9). Ms. Hardesty said E.M.'s work history would satisfy the sales and customer service requirement. (TR 636). Ms. Hardesty conducted the first interview, found that E.M. met all six of the core competencies, and sent him on to a second interview with Mr. Brian McVicker and Mr. Larry Taccone. (TR 638). At the second interview, K.M. received a "no" on persuasiveness with a written comment that he had "no sales background." He also got a "no" on communicating orally and he was rejected. (TR 639). Asked about the disparity between her assessment of K.M. and the assessment of Mr. McVicker and Mr. Taccone, Ms. Hardesty said that as a talent acquisition specialist, her goal is to hire people while the goal of branch, area and group managers is to maximize the performance of their units. (TR 639).

The talent acquisition department has hiring goals, including goals on minority hiring. They get monthly reports on minority hiring and they provide quarterly updates to the general manager on outreach to try and meet their diversity metrics. (TR 640). She said she and Ms. Morris prepare the quarterly report and present it to Mr. Arney who, in turn, presents it to Mr. Gavin Watson. Part of Ms. Hardesty's performance evaluation is based on meeting the minority hiring goal. The goal is to be above 80 percent of the MSA. Another goal is for the minority impact ratio analysis not to show an adverse impact on minority applicants. (TR 641). Ms. Hardesty assists Ms. Morris in taking the information from the applicant flow logs and inputting the data into a spreadsheet that is provided by the corporate office. The spreadsheet does the calculations on its own. The spreadsheets are provided to Mr. Arney and Ms. Hardesty was not sure what he did with them. (TR 642).

Ms. Hardesty identified PX 4 as an internet posting for the management training program that looked like it came from the Enterprise careers website. (TR 643). It is dated July 7, 2014. The posting explained that nearly all senior Enterprise employees began as management trainees where they learned the business from the ground up. (TR 643-644). It said that new hires are [*34] assigned to branch locations in their geographic areas where they get hands-on training. (TR 644). Over the first year, new hires actively participate in all phases of branch operations and will be tested and evaluated for rewards, higher pay and promotions. (TR 645).

Ms. Hardesty testified that she assists Ms. Morris in conducting training on how to conduct behavioral interviews. The training for new assistant managers is done quarterly. (TR 645). They get the formal training once and they get hands-on training from sitting in on interviews with their managers. (TR 646). Ms. Hardesty and Ms. Morris did interview training for the area managers a few months prior to the hearing. (TR 646-647). Mr. Arney was involved in the training as was Group 18 vice-president, Mr. Watson. The training materials were prepared by the corporate office and sent out to the field. The area manager training in Group 18 lasted for half a day. (TR 647). Part of the training included awareness of "similar to me" where an interviewer may view someone similar to him or her more favorably on the basis of their similarities. Ms. Hardesty acknowledged that she is white, but she said "similar to me" was not limited to race and could include attending the same college or having worked for the same employer, among other reasons. (TR 648-649).

Ms. Hardesty reviewed PX 15, which she said is a leadership guide for the behavioral interview training. It is dated 2014. (TR 649). The training discusses stereotyping, which it defines as "making a decision based on a standardized mental picture that represents an oversimplified opinion, prejudiced attitude or uncritical judgment." (TR 650; PX 15 at 24). At the training, they go around the room and have the participants give examples of stereotyping. (TR 650). They also talk about discrimination. Race is not specifically addressed as part of the formal training on stereotyping and discrimination, but Ms. Hardesty said race is often cited in examples of the behavior. (TR 651). She said she sees the numbers when the impact ratio analysis is computed and that there had not been any adverse impact in recent years for the management trainee program, so there had not been any need to go back and do any retraining. (TR 652).

Ms. Hardesty reviewed PX 111, which is the Enterprise Group 18 Affirmative Action Plan for August 1, 2006 through July 31, 2007. (TR 656, 565). OFCCP's counsel directed her attention to a chart entitled "Workforce Analysis" that in the upper left corner says "20201 — Human Resources."8 (TR 656). Ms. Hardesty said she is included in the chart as the "Recruiting Supervisor." The "Recruiting Administrator" position involves mainly administrative support duties, such as screening applications and sending out letters. That would have been Ms. Lichter's position. (TR 657). The "Recruiting Manager" position refers to Ms. Morris and the "Generalist Manager" refers to Mr. Arney. The chart shows there were six people working in human resources at the time and under the column heading "Minority" the total number was zero. (TR 658-659; PX 11 at 31).

On cross-examination by Enterprise, Ms. Hardesty described the efforts made to ensure that interviews for the management trainee program are conducted in a consistent manner. (TR 662). She said formal training is provided as employees move up in the organization to become assistant branch managers, branch managers and area managers. (TR 663-664). Human [*35] resources provides the interviewers with the questions to ask during the interviews. (TR 664). When an applicant goes out to a branch for a second interview it is usually an assistant branch manager who facilitates the branch visit and completes the branch observations checklist. The checklist is supposed to ensure the assistant branch manager covers all of the points that should be covered during the applicant's branch visit. The assistant branch manager can provide his or her assessment of the applicant on the form, but ultimately it is the branch manager or the area manager who makes the final decision. (TR 665-666).

Sales is a big part of the management trainee position and, along with customer service, is a key determinate in promotions. Ms. Hardesty said they do not want to put new hires in uncomfortable positions where they will not be successful, so they look for applicants who have proven in the past that they can succeed in sales and customer service. (TR 667-668). When asked what happens if an applicant does not list any sales experience, Ms. Hardesty responded:

Then, you know, depending on what experience they put on, it's hard to say, you know, with a definitive yes or no, without looking at the experience — but if nothing on their application shows sales experience or, you know, because sometimes they look at customer service as well, then they wouldn't meet that screening criteria.

Ms. Hardesty said the current minimum requirement for sales and customer service experience is at least one year of experience in two or more of the following areas: (1) sales, preferably in a commission-based environment, (2) customer service, preferably in retail or a restaurant, (3) military service and (4) college or professional athletics. The applicant has to check yes or no, and if they say they have the experience then someone reviews the applications, but if they say no the system automatically disqualifies the applicant and sends a rejection email. (TR 670). Sometimes an applicant gets a rejection email and contacts human resources and explains that he or she made a mistake on the application and actually has the experience required. In that case, the application continues on in the hiring process. (TR 671).

Ms. Hardesty remained on the witness stand for direct examination by Enterprise. (TR 671). She said that at no time between 2006 and 2017 did the hiring criteria vary at all based on an applicant's race. Applicants were able to self-identify racial data, which went into a database that Ms. Hardesty did not maintain. She said she was not aware of an applicant's race at any time in the initial screening stage. She started conducting telephone screenings when Ms. Lichter left. (TR 672). The purpose of the telephone screening is to make sure the applicant meets the minimum requirements and to pass along information to the applicant about the management trainee position. There were times when Enterprise needed to expedite the process [*36] and did open houses and skipped the telephone screening step. In those cases, applicants who applied online and met the minimum requirements were invited to an open house where they got a first interview with an area manager. (TR 673). If that went well, the applicant still went out for a branch interview and the process continued from there. (TR 673-674). An applicant who met the minimum requirements did not necessarily get hired because meeting the minimum requirements and demonstrating that the applicant possesses all of the core competencies are not the same. Ms. Hardesty said she was not aware of an applicant's race when she did telephone screenings. (TR 674).

The rejection notification gives the reason the applicant was rejected. In some instances, an applicant follows-up to inquire about the reason for the rejection and most handle it professionally and appreciate the feedback. (TR 675). On occasion, an applicant will feel passionately that they have all of the core competencies and human resources will choose to arrange another second interview with a different area manager. Ms. Hardesty said it has not happened often, but there have been cases where that resulted in an applicant getting hired and going on to be a very successful Enterprise employee. She said if an applicant was rejected for not being persuasive and then is persuasive enough to persuade her that a second chance is warranted the applicant gets another chance since the HR goal is to hire people. (TR 676).

Ms. Hardesty said she never made a decision at a first interview based on an applicant's race. (TR 676). She said the first interview was the first point in the process where she was visually aware of an applicant's race. (TR 676-677). She was not aware of a branch manager, area manager or group manager ever making a screening decision based on race. The screening process did not change based upon race between 2006 and 2017. (TR 677).

Ms. Hardesty reviewed and explained the disposition codes for applicants who choose to self-select out of the process, which are listed in DX 191. (TR 678-679). Those codes are applied when the applicant makes the decision that he or she is not interested in the job and chooses to opt out from further consideration. (TR 679).

Ms. Hardesty reviewed and explained the disposition codes for applications that are rejected during the screening stage. She said the S1 code is entered in the SHL system when an applicant does not meet the education requirement, but in ICIMS you choose from a drop-down menu and it writes out education. The education requirement changed over the years. In the 2006 to 2008 period the requirement was for two years of college, but they preferred someone with a college degree. (TR 680). Corporate made the decision around the time the recession hit in 2008 to make the minimum requirement a bachelor's degree. Ms. Hardesty said she would never use the education disposition code if it was not the true reason for rejecting an applicant. (TR 681). She reviewed DX 829, which is the application file for L.B. (TR 681). She did his telephone screening on December 18, 2013 and he said that he would not have a degree until December 2014. (TR 682; PX 829 at 6-7). This is an example of when she would enter an S1 code indicating the applicant was screened out for not meeting the education requirement. (TR 683).

[*37] The S2 code is for job stability. Enterprise wants to hire employees that do not hop around from one job to another. (TR 683). It is not a code Ms. Hardesty used often and she said she would not use it if it was not the true reason for rejecting an applicant. (TR 684).

The S3 code is for sales and customer service work experience. Ms. Hardesty said she would not use this code if it was not the true reason for rejecting an applicant. She reviewed DX 3459, which is the application file for G.K. (TR 684-685). He was rejected for lack of relevant experience because he worked selling food at a carnival, which was more transactional than selling, and he waited tables at Ruby Tuesday's for three months. This did not give him the minimum relevant work experience required for the position. This is an example where an S3 code would be entered for not meeting the work experience requirement. (TR 685).

The S4 code is for the quality of the IRISS application, spelling and grammar. Since switching to ICIMS, the same reason would be described as professionalism. (TR 686). Ms. Hardesty said this code was not used often and she would never use it if it was not the true reason for rejecting an applicant. (TR 687).

The S5 code is used when there is a lack of professionalism, rude communication, inappropriate language or negative responses. Ms. Hardesty said it was similar to the reasons for an S4 code. (TR 687). Ms. Hardesty reviewed DX 1466, which is the application file for K.C. She said she did the telephone screening for K.C. and rejected her because she said she was unemployed and just wanted a job, could not explain why she was interested in working for Enterprise and did not recall what position she applied for. Ms. Hardesty said this is an example of when an S5 code would be appropriate. (TR 688).

The S6 code is used when an applicant is not flexible in setting up a telephone screening or a face-to-face interview. (TR 689). The S7 code is used when the geographic area is not compatible. As an example, she said that if an applicant lives in the District of Columbia there is no reason to commute 45 minutes to the Baltimore area when the applicant could interview for a job with Enterprise in the District of Columbia group. (TR 689-690).

The S8 code is used when an applicant does not have compatible career interests and is not looking for an entry level position in sales, customer service, management and marketing or the pay or hours are not acceptable. (TR 690). As an example, Ms. Hardesty said if an applicant really wants to be a doctor and plans to apply to medical school then the applicant does not have a career interest that is compatible with being a management trainee at Enterprise. (TR 691). She reviewed DX 1237, which is the application file for P.C. She conducted the telephone screening for P.C. and rejected him because he said he wanted to be a lawyer, which is incompatible with a career as a management trainee. (TR 691-692).

The S9 code was used back during the time when Enterprise ran a credit report on applicants. Ms. Hardesty said they stopped running credit report on management trainee applicants in about 2011, although they still run them on applicants for accounting positions. (TR 692).

[*38] The S10 code is used for an adverse criminal records check. (TR 692-693). The application used to ask if the applicant had ever been convicted of a felony. Now, Enterprise runs a background check before making a job offer. Those that come back with criminal records are assessed on a case-by-case basis. Ms. Hardesty reviewed DX 6528, which is the application file for R.T. He made it through the telephone screening and the first interview, and he appeared to be someone Enterprise was interested in hiring, but a background check conducted by a third-party vendor disclosed that R.T. had a felony conviction for attempted second degree rape and was on the Maryland sex offender registry. Enterprise did not hire R.T. (TR 694-695; DX 6528 at 8-9).

The S11 code is used when an applicant is rejected due to an adverse Maryland motor vehicle records check. (TR 695). Ms. Hardesty reviewed DX 4422, which is the application file for D.M. An applicant cannot have more than two moving violations in the previous three years and in this case D.M. had three violations. Ms. Hardesty said D.M. waited until enough time passed for one of the violations to drop out of the three year period and he was then hired. (TR 696-697).

The S12 code is used when an applicant does not have a valid driver's license. (TR 697). The S13 code is used when there are problems with an applicant's references during the background check. Ms. Hardesty said that code is not used often. (TR 697-698). The S14 code is used if someone applies who is less than 18 years of age. (TR 698). The S15 code is used if an applicant has applied before and been rejected and then reapplies within six months. (TR 698-699). The S16 code is used if an applicant is not legally eligible to work in the United States. (TR 699). The S17 code is used when there are no positions for the applicant to fill. Ms. Hardesty said that code was used a few times during the recession when there were no openings and when an applicant wants to work on the Eastern Shore and there are no positions available. (TR 699-700).

Next, Ms. Hardesty went through each one of the "I" codes, which she said are used when an applicant is rejected during any of the interview stages. (TR 700).

The I1 code is used when an applicant is rejected for a lack of communication ability. (TR 700). For example, Ms. Hardesty said if an applicant does not speak clearly, make eye contact and has poor body language I1 would be the appropriate disposition code to enter. She said she had applicants fall asleep during interviews. She reviewed DX 5661, which is the application file for C.R. Ms. Hardesty did his first interview and rejected him for multiple reasons, including communication ability (I1), leadership (I2), and customer service (I3). She wrote on the interview form under communication that he did not give specific examples, said that when he had managed a site there was always trouble, and he made derogatory comments about people who do not speak English. (TR 701-702). Enterprise is looking for management trainees that will advance to become managers, so someone who says there is always trouble when he manages a site does not show strong leadership potential. The core of Enterprise's business is customer service, so someone who makes derogatory comments about people who do not speak English is not likely to foster good customer relations. C.R. was rejected and disposition codes I1, I2 and I3 were recorded. (TR 702-703).

[*39] The I4 code is used when an applicant is rejected for a lack of persuasive ability and the I5 code is used if an applicant is rejected for failing to demonstrate that he or she could be flexible in a changing environment. (TR 704). Ms. Hardesty reviewed DX 3816, which is the application file for D.L. She conducted the first interview and rejected D.L. for a lack of persuasiveness and flexibility. He only gave examples of how he provided customer service, not how he had persuaded someone to make a decision. He gave an example of when he had rented a car and how unhappy he was when they did not have the exact model or color he wanted, and how customers at the wood business where he currently worked would get upset with him for being too picky and inflexible. (TR 705-706; DX 3816 at 16-17).

The I6 code is used when an applicant is rejected for not demonstrating that he or she is results driven. Ms. Hardesty reviewed DX 4022, which is the application file for M.L. Ms. Hardesty did the first interview with M.L. and rejected her for customer service and results driven reasons. (TR 707). For customer service, M.L. said she only went the extra mile when the boss was around because people complained just to get things for free. For results driven, she could not give examples of achieving organizational goals or of setting and meeting her own personal goals. (TR 708-709).

The I7 code is used when an applicant is rejected for not having career ambitions that are compatible with a career at Enterprise. For example, Ms. Hardesty said if an applicant's goal is to attend law school or medical school then his or her career ambition is not compatible with a job in sales. She reviewed DX 7080, which is the application file for C.W. Ms. Hardesty did the interview and rejected C.W. because he said his goal was to become a lawyer. He met all of the other core competencies, but the fact that he had applied to law schools indicated that he did not have long-term career ambitions at Enterprise. (TR 709-710).

Ms. Hardesty said it costs about $7,000.00 to hire a new management trainee to replace one that departs, so retention is an important issue. One of the performance metrics for human resources is 120-day retention. They meet with senior level managers to talk about the reasons management trainees leave and they examine their training to see if they need to make adjustments to promote retention. They survey new hires to see how things are going and to ask what they can do to improve the process. They do branch visits to meet with employees and they do recognition programs to reward successful performers. Ms. Hardesty said she believes the hiring criteria help to promote retention by bringing in people equipped to perform well in a sales environment. (TR 711-713).

The I8 code is used when an applicant is rejected for a lack of professionalism: arriving late for an interview without calling ahead, poor grooming and appearance, bad manners or not well-spoken. Ms. Hardesty reviewed DX 1649, which is the application file for R.D. (TR 714). Ms. Hardesty conducted the first interview and gave R.D. a "yes" on all of the core competencies. She rejected R.D. because she was 20 minutes late for the interview. Ms. Hardesty said she told R.D. when they spoke on the telephone to allow plenty of time to get to the interview because R.D. was relying on public transportation, which is a challenging way to get to the Group 18 headquarters location. Despite the warning, R.D. was late and was rejected for a lack of professionalism. (TR 715).

[*40] The I9 code is used when an applicant is rejected for not being prepared for the interview. Ms. Hardesty said that in this day and age there is no excuse for going to an interview and not knowing anything about the company where you are seeking employment. (TR 715). She said the I9 code is not used often. She reviewed DX 1395, which is the application file for D.C. D.C. was rejected for customer service, persuasiveness, flexibility, leadership and oral communication reasons. For customer service, she said that when a customer gets angry she just goes and gets the manager. (TR 716). She did not have an example of when she persuaded someone and for flexibility she said she called security when a customer was shoplifting because dealing with a shoplifter was not her job. She said she did not want to be the captain of the team, she wanted to be a player. Under oral communication, Ms. Hardesty noted that D.C. used "we was" quite a bit and did not elaborate on examples. She also noted that D.C was half an hour late for the interview and when she asked her if she had done any research on Enterprise, D.C. said she had not because she was playing basketball. (TR 716-718).

The I10 code is used when an applicant is rejected for not showing a sincere interest in the management trainee program. Ms. Hardesty said this reason is very similar to the I7 incompatible career ambitions disposition code. She reviewed DX 360, which is the application file for W.A. Ms. Hardesty conducted his first interview and rejected him for several reasons, including saying he applied to be a management trainee because he was "leaving no stone unturned" and thought it was "worth a shot." (TR 719-720; DX 360 at 10).

Ms. Hardesty denied that she ever took an applicant's race into consideration in assigning disposition codes and she said that she was never instructed to do so by a supervisor or corporate policy. She said that Enterprise recruits minority applicants for the management trainee program. Group 18, she said, was fortunate to be in a diverse area and it had excellent relations with the historically black colleges and universities in the vicinity. (TR 721). They participate in career fairs at colleges, through veterans' organizations, on-post at Fort Meade, and with Goodwill Industries. (TR 722). They post on hiring websites like Monster and they search online for potential applicants. Enterprise is an equal employment opportunity ("EEO") employer and everyone involved in the recruiting and hiring process is trained on EEO. (TR 723). Ms. Hardesty testified that in her time working for Enterprise she never made a hiring decision based on an applicant's race. (TR 724).

I asked Ms. Hardesty whether as an applicant proceeded through the interview process the subsequent interviewers had the benefit of seeing what prior interviewers had said about the applicant. She said no, an interviewer would only be aware that the prior interviewer found that the applicant met all of the core competencies, otherwise the applicant would not have progressed to the next stage, but the next interviewer does not see the comments from the earlier interviews. (TR 724-725).

SUSAN O'HARA

Ms. Susan O'Hara was called as a witness by OFCCP. Ms. O'Hara works as an income maintenance administrator for the Pennsylvania Department of Human Services. She supervises seven employees and each of them supervises about seven people. She is responsible for overseeing the administration of benefits to eligible recipients. (TR 727). She had worked for [*41] the Department of Human Service for seven years at the time of the hearing and she had progressed from being a caseworker to a supervisor to an administrator. (TR 728). Ms. O'Hara is a 2008 graduate of Morgan State University, which is a historically black college or university. She estimated that about two percent of the student body at Morgan State is white. (TR 728-729).

Ms. O'Hara reviewed PX 89 and identified it as her application file. She said that at the time, she was a recent college graduate and was looking for a job that provided good pay and benefits and that would enable her to progress as a professional. (TR 729). She chose to apply for the management trainee position at Enterprise because she had previous sales and customer service experience, which matched with the management trainee position. (TR 730). She listed a sales associate position at Gymboree — a children's clothing store — on her application. (TR 731). Part of the job required her to try and sell customers on promotional items. For example, she said if a customer spent a certain amount of money the customer got an amount taken off of the purchase price, so she tried to sell them on reaching the promotional dollar limit. (TR 731-732). She was lead tutor at the Morgan State University Office of Residence Life. Her job required her to meet with students, assess their needs, and match them with tutors. She also put on training sessions, workshops and labs. (TR 732-733). She was an institutional advancement intern at Maryland Public Television where she contacted people about pledging and performed other tasks to help raise money to support the network. (TR 733-734). She did not recall the specifics of the telephone screening conversation with the Enterprise recruiter. If Enterprise had offered her a job she said she would have accepted it. (TR 735).

On cross-examination by Enterprise, Ms. O'Hara said she believed she was qualified for the management trainee program because she had a college degree and sales and customer service experience. She believed that she was aware those were requirements for the position from looking at an online advertisement. She did not recall the specifics of the telephone interview or the exact responsibilities of a management trainee. (TR 735-736). She said she understood that the job meant working in a branch office renting cars and she said she would have taken the job. Ms. O'Hara said she also applied for a position with Target that was similar to the management trainee position at Enterprise and she applied for a position at AFLAC selling insurance. (TR 737). She recalled receiving the rejection letter from Enterprise and she remembered calling, but she was not certain if she called before or after she got the letter. (TR 738). Ms. O'Hara testified that she did not know why Enterprise rejected her and she did not know if anyone she spoke with for the telephone screening knew her race. (TR 739).

I asked Ms. O'Hara to describe what she meant when she said in her application that she enjoyed acting and writing and had participated in and directed a stage production. (TR 741; PX 89 at 4). She said she participated in two plays while she was at Morgan State University and the money raised went to a domestic violence shelter. I asked if she could recall any of the substance of the conversation during the telephone screening and she said that she could not. (TR 742).

MICHELLE C. EVANS

[*42] Ms. Michelle Evans was called as a witness by OFCCP. (TR 744-745). Ms. Evans said that previously her last name was Green. She is a human resources administrator at Security Plus Federal Credit Union, a position she has held for two years. She has been employed by Security Plus for 11 years. (TR 745). Ms. Evans graduated in 2006 from Morgan State University. (TR 746). She earned a MBA from Morgan State University in 2016. Morgan States University is a historically black college or university and Ms. Evans said 85 to 90 percent of the students and faculty are African-Americans. (TR 747).

Ms. Evans learned about the management trainee program from a classmate who worked at Enterprise. He said it was a really good program and she was interested because she wanted to work for a good company and gain management experience. She said she wanted management experience because she always strived to do more than average and to be a leader. She said she hoped to grow with the company. (TR 748).

Ms. Evans identified PX 86 as her application for the management trainee program. I noted that the copy of the exhibit provided to me contained blank pages and the page numbers appeared to be in random order. The parties conferred and agreed that was the case. Ms. Evans's application file that was in the defense exhibits was the same as the one in the plaintiff's exhibits. The parties agreed that all of the pages are included in both of their exhibits, but the pages are not in sequential order. (TR 749-751).

On her application, Ms. Evans listed that she was a business development intern/internship coordinator at MECU of Baltimore. (TR 751-752). She said she met with businesses to try and get them to join so their employees could enjoy the benefits of credit union membership and as internship coordinator she worked with schools across the country to try and get them to partner with the credit union so their students could get internships. It was a paid position that allowed her to work part-time during the school year and full-time over the summer. (TR 752). She was working at MECU of Baltimore when she applied for the management trainee program. Her job was both sales and customer service. (TR 753). She also worked at Victoria's Secret as a sales associate and she interned at the Afro-American Newspaper as a marketing intern. The job at Victoria's Secret included selling merchandise as well as trying to sell customers on a Victoria's Secret credit card. The marketing internship involved selling newspaper subscriptions. Both of those positions were part-time while Ms. Evans was in college. (TR 754-756).

Ms. Evans successfully completed the telephone screening and a first interview at the Enterprise headquarters in Linthicum, Maryland. (TR 758). She went for a second interview with Mr. McVicker and afterwards she was rejected. Mr. McVicker marked "no" for results driven, but he did not explain why he came to that conclusion. He wrote at the end of the form under reason for rejection "looking at ERAC for experience." (TR 759; PX 86 at 17). Ms. Evans agreed that she was looking at Enterprise for experience, but she wanted to gain experience so she could advance and move up in the company. She said she thought both of her interviews went well and she was surprised when she found out that she had been rejected. (TR 760).

[*43] On cross-examination by Enterprise, Ms. Evans said she had previous sales experience where she was required to meet sales goals. She said she got a $300 performance bonus at MECU of Baltimore. (TR 761-762). The friend that referred her to Enterprise is African-American and was working at Enterprise at the time she applied. (TR 763). She did not recall if her first interview was with a man or a woman. The second interview was at a branch office. (TR 764). She remembered that it was raining and the person who interviewed her was a white male about 40 years of age. (TR 765). She did not recall if the interviewer's name was Brian McVicker. (TR 766). She did not recall if the interviewer asked whether she was results driven. (TR 768). Ms. Evans testified that she did not know why she was not hired, whether she was treated differently than any other applicant, or whether her race was a factor. (TR 769-770). The job she got at Security Plus Federal Credit Union required a lot of market research and about 20 percent of the job was sales. (TR 772). She acknowledged that she had no evidence that she was treated differently by Enterprise because of her race. (TR 773).

On re-direct examination by OFCCP, Ms. Evans said that if Enterprise had offered her the job as a management trainee she would have accepted it. (TR 773-774). In her current position, Ms. Evans says she has responsibilities related to diversity. (TR 775). She said they seek diversity in their applicants and want the interviewing and hiring process to be fair. She said discrimination can happen and she sees it on a regular basis. (TR 776).

WESLEY WUCHER

Mr. Wesley Wucher was called as a witness by OFCCP. Enterprise hired him as a management trainee in November 2014 and in February 2016 he became a recruiting coordinator. (TR 778-779). He began conducting telephone screening calls in December 2016. He left Enterprise about a month prior to the hearing for a job closer to his residence that paid more money. (TR 779). He currently works for Astor & Sanders Corporation. (TR 780).

Mr. Wucher said his management trainee program began with a five-day orientation course that focused on sales techniques, delivering customer service and presentations by different departments. (TR 780-781). His class came back for day-long training sessions at regular intervals. That training was more intensive training on the same topics covered in the orientation course. (TR 781-782). Mr. Wucher started out in Group 21 in the Hampton, Virginia, area. He was a management trainee from November 2014 until February 2016. He did not take the MQI because he decided to pursue the recruiting route rather than staying in sales. (TR 782-783). He moved to Group 18 in July 2015

When he first started working in human resources, Mr. Wucher was not involved in conducting the orientation course, but as he gained experience he began doing presentations on recruiting. He was also involved in coordinating and arranging the courses. (TR 784-785). The number of applications was greatest around college graduation season, which was May and December. (TR 786-787). He did not have a hiring quota and he was not evaluated on the number of new hires he brought in. He did not have a minority recruitment goal, although he believed Enterprise did. He said his goal was "just get people in." (TR 787).

[*44] Mr. Wucher began conducting telephone screening calls in December 2016. Ms. Morris or Ms. Hardesty would review the application first and then, if it appeared the applicant met all of the minimum requirements, one of them would tell Mr. Wucher to reach out to the applicant and schedule a telephone screening. He never entered any disposition codes until there had been a telephone screening. (TR 788). Prior to conducting telephone screening calls on his own, Mr. Wucher sat in with Ms. Morris or Ms. Hardesty for about 10 months and listened in as they did the screening calls.

He identified the information in DX 75 with a date of August 2015 as the qualifications for the management trainee position. (TR 789; DX 75 at 2-3). He said that during telephone screenings he asked applicants if they had driver's licenses and about moving violations. When asked if he had to make subjective judgments on applicants' level of work experience, he answered, "[g]ive or take, yes." (TR 790). He said he had the application in front of him during the call and knew the applicant's work history, so what he was looking for was for them to talk about it, translate and explain the work experience, and "basically persuade me to have them come in." (TR 791). When asked if the minimum of one year of relevant work experience was a "hard and fast rule," he answered:

No. I wouldn't say so. Again, going back to kind of my first point of we really did give people the benefit of the doubt. So, for instance, if I had a candidate with let's say ten months of some sort of retail or service industry experience, but again they had that personable personality, then I would bring them in. It wasn't a cutoff. Now, with that being said, if I had someone with one month, no. But at the same time, that never really happened because if a candidate only had one month I — they probably wouldn't have gotten to the phone screen portion.

Mr. Wucher said that ICIMS kept track of racial data on applicants and that he could access that data if he chose to do so. He entered disposition codes for the applicants he rejected by clicking on a drop-down menu and then clicking on the reason he rejected the applicant. (TR 793). He said that normally when he rejected an applicant he had Ms. Morris or Ms. Hardesty review his notes to make sure his reasoning was sound, but that was not done 100 percent of the time. (TR 793-794).

Mr. Wucher said the job titles shown on page 25 of DX 100, which was from August 2015, were the titles used when he worked in human resources. His official title was human resources coordinator and Jasmine Brown had the same title. Mr. Wucher acknowledged that he is white and he said that Ms. Brown was the other human resources coordinator who was counted under the minority heading. (TR 795). He said Enterprise never provided him any training on unconscious bias, implicit bias or subconscious bias, and that all of his training on recruiting was provided informally by him sitting in with Ms. Morris or Ms. Hardesty. (TR 797).

On cross-examination by Enterprise, Mr. Wucher said that while ICIMS contained racial data on applicants, he never looked at it. He said he never took an applicant's race into account [*45] when he conducted telephone screenings. (TR 798). He said the other human resources coordinator, Ms. Brown, was African-American. (TR 799).

On direct examination by Enterprise, Mr. Wucher expounded on how he assessed sales and customer service experience during his conversation with an applicant during the telephone screening. He said his objective was to bring people in, so he wanted them to talk and show that they were personable rather than providing one word answers. (TR 799). If an applicant had experience on his or her application and did not bring it up in response to his question, Mr. Wucher said he would follow up and ask him or her to elaborate on how the work experience relates. The applicant did not need to talk about the job he or she had as long as the applicant conveyed confidence, leadership and persuasion. (TR 799-800). With respect to customer service experience, Mr. Wucher said there "wasn't necessarily some sort of that qualifies or that doesn't qualify" for particular past positions, instead he looked for whether the job required dealing with the public. (TR 801). With respect to sales experience, he said that was "even broader" and really focused on whether the applicant was persuasive and persuaded him that he wanted to bring them in for an interview. (TR 801). He said he was not aware of an applicant's race when he conducted the telephone screening. (TR 802).

Mr. Wucher identified DX 191 as the ICIMS disposition codes used at the self-select out, screening and interview stages. (TR 803). He said he input codes into ICIMS and did so for all three stages. (TR 803-804). As an example, an applicant might tell Ms. Morris or Ms. Hardesty that he could not work the hours or she thought the pay was not acceptable, and they would tell him to enter that the applicant self-selected out for whatever reason and he would enter it into the system. He said he never conducted an in-person applicant interview. (TR 804). If an applicant was rejected at the first, second or third interview, he checked the notes on the interview evaluation form and entered the code into ICIMS for the reason that the applicant was rejected. (TR 805).

Mr. Wucher said there were 17 codes for the reasons an applicant was rejected during the screening stage and that he "barely used hardly any, for the record." Mr. Wucher added:

Work experience is easily the one that I used probably the most, S3. And that would be, again, going back to it if they weren't personable, if they didn't have that persuasiveness in their voice and they didn't have the experience on their application and they weren't able to implement it into our phone conversation, it would be an S3.

Mr. Wucher identified DX 3044 as the application file for M.H. (TR 809). Mr. Wucher conducted the telephone screening and rejected M.H. for lack of communication skill and entered an S1 disposition code. His notes indicate that M.H. did not answer the telephone when he called at the scheduled time and he was vague and inconsistent on when he expected to graduate from college. (TR 810-812).

Mr. Wucher identified DX 4883 as the application file for Y.O. (TR 812-813). He was unable to pronounce the last name and he spelled it for the record. He conducted the telephone screening and rejected the applicant based on work experience. According to the information provided in the application, Y.O. graduated from high school and law school in Nigeria and earned a master's degree at the University of Baltimore. Y.O. was working as a geriatric nursing assistant at a nursing home and as a driver for Enterprise at the time his application for the management trainee program was submitted. Mr. Curry, a Group 18 area manager who was mentioned often in the testimony of Ms. Morris and Ms. Hardesty as a second interviewer, was listed as Y.O.'s current supervisor. (DX 4883 at 7). Mr. Wucher wrote on the telephone screening form that Y.O. did not give any sales examples and was rambling rather than focused during the conversation. Mr. Wucher rejected Y.O. and entered an S3 disposition code. (TR 813-814; DX 4883 at 11).

Mr. Wucher identified DX 6417 as the application file for C.T. (TR 814). He conducted the telephone screening and rejected the applicant for not having a valid driver's license for at least one year and he assigned an S12 disposition code. He said the conversation was going well and C.T. would likely have been invited in for an interview until she said she had not had a driver's license for at least one year, which was a requirement to satisfy the minimum qualification. (TR 815). The application just asks the applicant whether he or she has a valid driver's license, not whether he or she had the license for at least one year, so checking "yes" would have gotten C.T. through without being automatically rejected by the computer screening. (TR 816).

Mr. Wucher identified DX 2047 as the application file for E.E. He conducted the telephone screening and rejected her for not having compatible career interests. His notes indicate that she said she was more interested in an office or administrative job rather than a sales and customer service job. He entered an S8 disposition code. (TR 816-817).

Mr. Wucher identified DX 3580 as the application file for J.K.-L. (TR 817-818). Mr. Wucher conducted the telephone screening and rejected the applicant based on a lack of relevant work experience. His notes indicate that J.K.-L. was interested in communications and marketing rather than sales and there was no sales involved in his prior job. Mr. Wucher entered an S3 disposition code. (TR 818-819).

Mr. Wucher identified DX 5990 as the application file for J.S. and he said he conducted the telephone screening. J.S. was rejected for lack of relevant work experience. Mr. Wucher noted that while J.S. had sales experience in the jobs he listed on his application, he failed to [*47] elaborate on his experience during their conversation. He said J.S. just wanted a full-time job, gave short answers in response to his questions, and he was not personable. Mr. Wucher said ICIMS only allowed him to pick one reason for rejecting an applicant in the screening stage, so deciding which one to choose was "the tricky part." He said he tried to select the most obvious reason if there were multiple reasons for rejection and in this case the inability to articulate sales experience was the most obvious choice. (TR 819-822).

Mr. Wucher identified DX 3910 as the application file for W.L. He conducted the telephone screening and rejected W.L. for incompatible career interest, S8. Mr. Wucher said W.L. stated that he was not interested in sales or customer service and he wanted to be an area manager. He also said the $44,000 starting salary was not acceptable. W.L. worked for a nonprofit organization helping customers, but the job did not involve doing any sales work. Mr. Wucher said he could have rejected W.L. for lack of relevant work experience, but incompatible career interest was the more obvious choice. (TR 822-824).

Mr. Wucher testified that he never took an applicant's race into account when he applied disposition codes and he was never instructed to do so by a supervisor or company policy. (TR 824-825). He said he never made any decisions related to the screening of management trainee applicants based on race. (TR 825).

On cross-examination by OFCCP, Mr. Wucher said he never looked in ICIMS to find an applicant's race. (TR 825). He denied that he would have a pretty good idea about an applicant's race by looking at the information provided in the application. (TR 825-826). When asked if he would have an idea whether an applicant named Da'Shawn was African-American, he answered "Sure, I could make an educated guess." (TR 826). When asked if an applicant was more likely to be African-American if he or she attended a historically black college or university, Mr. Wucher said he did not know the area colleges very well and could not say which ones were historically black colleges and universities. (TR 826). He said he did not look at the activities applicants listed in their applications, so if an applicant noted involvement in the NAACP or similar organizations he probably would not see it and would not draw any inferences about race. He said he never did a telephone screening where an applicant's accent or dialect caused him to think the applicant was African-American. (TR 827).

Mr. Wucher was referred back to DX 4883, the application file for Y.O. Under education it said Y.O completed high school in Nigeria and Mr. Wucher agreed that would suggest Y.O. was likely African-American. (TR 828-829). He said Y.O.'s name would not cause him to suspect one way or the other whether Y.O. was African-American. [Note — On direct examination by Enterprise, when Mr. Wucher was asked if he could identify DX 4883, he answered, "I can. Probably mess up the name, but Yetunde something," and he spelled the last name for the record. (TR 813)].

Mr. Wucher was referred back to DX 3580, the application file for J.K.-L. (TR 829). He agreed that he had never met a white person with a first name spelled "Juleon." (TR 830). He was referred back to DX 2047, the application file for E.E. who graduated from Morgan State University. Mr. Wucher said he did not know whether Morgan State University was a historically black college or university. (TR 830-831).

[*48] On re-direct by Enterprise, Mr. Wucher said that when he worked in human resources his desk was right outside Ms. Morris's and Ms. Hardesty's offices. He saw some of the applicants that he had telephone screened when they came in for in-person interviews and some of them were African-Americans. (TR 832).

In response to questions from me, Mr. Wucher said that an applicant who did not have one year of sales and customer service experience, but who sold him on their personality and salesmanship, would move on to an in-person interview. He said the work experience requirement "was a little bit more subjective than the rest of my phone screen." That was a practice he said he learned from observing Ms. Morris and Ms. Hardesty. (TR 833). I asked Mr. Wucher if he could elaborate on the difference in an S3 disposition code for work experience and an S8 disposition code for career compatibility. He said that an applicant who really made a good impression but just did not have enough work experience could gain more experience, reapply to Enterprise, and perhaps get hired, while an applicant who really wanted a career in some other field was unlikely to change his or her mind and decide he or she wanted to do sales. He also noted the applicant who had a driver's license for less than one year (applicant C.T. — DX 6417) made a good impression during the telephone screening and could reapply in a few months when the one year driver's license time requirement was met. (TR 834).

MIA L. LOWERY

Ms. Mia Lowery was called as a witness by OFCCP. She is currently employed as the supervisor of the soil evaluation section, water management, at the Baltimore County Environmental and Sustainability Department. She also works as a cashier at Office Depot. She has worked for Baltimore County for three years and for Office Depot for one year. (TR 836-837). She supervises two people in her job with Baltimore County and at Office Depot she spends half her time performing cashier duties and half her time assisting customers on the sales floor. In her job with Baltimore County, she interacts with residents, engineers and consultants who are developing properties that are on wells and septic tank systems. (TR 838). She graduated from Florida A&M University, which is a historically black college or university. She estimated that the student body was 90 to 95 percent African-American. She graduated in 2007. (TR 838-839).

Ms. Lowery applied for a management trainee position. She said she was a biology major and wanted to go to medical school or dental school, but she wanted to take a break from school before she did and she wanted to get a job. (TR 839-840). Ms. Lowery identified PX 88 as the application file for the application she submitted. (TR 841). She listed work as a cashier at Office Depot in Tallahassee, Florida, where she covered the technology section. She assisted customers in choosing electronics and tried to upsell them. She worked there from July 2006 to April 2007. (TR 842-843). She worked as canvasser for the Florida Democratic Party where she went out into communities and offered information on candidates and registering for absentee ballots. (TR 844). She held that job for about a month while she was a student in college. (TR 845). She worked as a hostess at Duclaw Brewing Company from June 2002 to August 2002. (TR 845-846). She worked as a cashier at Stride-Rite Shoes when she was a high school student from November 2000 to August 2001. (TR 846-847). She had a vague recollection of the [*49] telephone screening call with Enterprise where she was asked questions about her resume and her work experience. (TR 847). She said that had she been offered a management trainee job she would have accepted. (TR 848).

On cross-examination by Enterprise, Ms. Lowery said she did not know why Enterprise did not hire her. (TR 851). She said she did not recall the details of the conversation she had during the telephone screening. (TR 851-852). She recalled that she also applied for a job at Sam's Club, Center Stage Theater, and with a temporary placement agency. (TR 855).

SARA LICHTER

Ms. Sara Lichter was called as a witness by OFCCP. (TR 869). She joined Enterprise as a management trainee in July 2002 immediately after graduating from college. She was a management trainee for about two years and then became a customer service representative at the airport location. She moved to the human resources department in November 2006 as a recruiting administrator, a position she held until November 2008 when she was laid off during the Great Recession of 2008. (TR 870). She has not worked for Enterprise since she was laid off. (TR 871).

Part of her job as a recruiting administrator was to review applications for the management trainee program to determine whether applicants met the minimum requirements. (TR 871). Applicants who met the minimum requirements were schedule for an in-person interview with Ms. Morris or Ms. Hardesty. Ms. Lichter reported directly to Ms. Morris. When Ms. Lichter arrived at work in the morning, she logged into the IRISS system and printed out the applications that came in since the last time she had checked the system. (TR 872).

Ms. Lichter identified PX 67 as the application file for A.J. The first six pages of the exhibit is the application as it appeared when she printed it from IRISS. She placed the stamp in the upper right corner of page one where it says "date received" and she wrote in "6/22." (TR 673; PX 67 at 1). The applicant's signature on page six was not there when Ms. Lichter printed it from the IRISS system. The applicant reviews the application when he or she comes in for an interview, makes any corrections that are necessary, and signs the application. (TR 874).

Ms. Lichter said she was trained on the procedure for screening applications. She reviewed PX 10A and said it showed the minimum requirements for the management trainee position. (TR 874-875). She reviewed PX 10B and said it reflected the minimum requirements as modified in 2008 to require a college degree. It states that applicants who meet the minimum requirements will be scheduled for telephone screenings. (TR 875-876). She agreed that when it was unclear whether an applicant met the minimum requirements the applicant got the benefit of the doubt and was scheduled for telephone screening. (876).

Ms. Lichter said PX 11 is a telephone screening form that she was trained to use when she screened an applicant for the management trainee program. (TR 876). She asked every applicant if he or she was interested in a career involves sales or customer service. (TR 876-877). Enterprise was looking for applicants who answered "yes" to that question. She asked every applicant to tell her about their previous sales and customer service experience in order to [*50] confirm the minimum work experience requirement was met. (TR 877-878). There was no minimum amount of sales or customer service experience required. Instead, Ms. Lichter assessed whether examples of actions taken in previous jobs aligned with the management trainee position. She was looking for customer-facing responsibilities, like meeting sales goals and handling customer complaints, the types of things that are typical in a retail sales environment. (TR 878). Ms. Lichter said working as a cashier in a grocery store might count as sales experience depending on how the duties were described. She agreed there was some overlap between sales experience and customer service experience. (TR 879).

If Ms. Lichter determined during a telephone screening that an applicant did not meet the minimum requirement for sales or customer service experience she entered the S3 disposition code. (TR 879-880). If she was "on the fence" on whether an applicant met the minimum requirement she would discuss it with Ms. Morris or Ms. Hardesty before a rejection letter was sent out. She did not recall if it was common for her to be "on the fence" about whether to reject applicants using the S3 code. (TR 880).

Going back to PX 67, the application file for A.J., she said it was her job to get a background check on an applicant from ADP if the applicant was sent on to a second interview. She ordered a criminal records check if an applicant was sent on to a third interview. Page 18 of PX 67 shows a background check and page 21 shows a criminal records check. (TR 880-881). She also ran a credit check at the same time as the background check and page 20 shows the credit check. (TR 881-882).

Ms. Lichter identified DX 191 as the disposition codes that were used in the screening and interview phases of the applicant assessment process. (TR 882-883). She said Ms. Morris or Ms. Hardesty would conduct an interview and, if the applicant was rejected, would write an "I" code on the application and then she would enter the code into the system. She did the same with the self-select out codes. (TR 883-884). She recalled that those involved in recruiting did something with applicant flow logs every month, but she did not recall exactly what it was. With respect to the disposition codes, Ms. Lichter agreed that which code to assign was not always cut and dried, particularly when it came to applying the sales and customer service experience code. (TR 885). She agreed that potentially there was some discretion involved in using the career compatibility disposition code. (TR 886).

Ms. Lichter identified PX 73 as the application file for T.A. and she said that she conducted the telephone screening. (TR 886). T.A. was rejected after the telephone screening and Ms. Lichter assigned an S8 code for incompatible career interest. (TR 887). His work experience was as a legal clerk, a family violence division intern, work in the public defender's office and at a domestic violence shelter, and he said he wanted to go to law school in a few years. (TR 888). Ms. Lichter said all of that indicated to her that he was not really interested in a career in sales and customer service, so she rejected him using the S8 code. (TR 888-889).

Ms. Lichter identified PX 74 as the application file for A.A. (TR 889). The rejection letter to A.A. is signed by Ms. Lichter. (TR 890). The first job listed in the application was retail sales associate at Williams-Sonoma from October 2000 to February 2004, which Ms. Lichter considered to be sales and customer service experience. (TR 890-891). A.A. also [*51] worked as a realtor for RE/MAX beginning in August 2006. The application to Enterprise was rejected in August 2007. Ms. Lichter agreed that work as a realtor was sales and customer service experience. (TR 891-892). She said A.A. told her that she enjoyed sales, but she did not want to be a manager, a "key holder," and that she could not pretend that she believed in a product. (TR 892; PX 74 at 9). Ms. Lichter said this was an application for a management training program, so not wanting to be a manager was inconsistent with the position. Also, the job involves selling intangible products like insurance, so if you do not believe in it you are not going to have a desire to sell it. Ms. Lichter acknowledged that not all successful management trainees move up into management positions and some, like herself, move into other areas like human resources. She said she did not serve as a branch manager before she joined human resources. (TR 892-893).

Ms. Lichter identified PX 75 as the application file for K.A. (TR 893). K.A. had worked as a shift manager at Taco Bell, which Ms. Lichter said she considered to be customer service experience. (TR 894). According to her application, she was working as a substitute teacher in the Baltimore City School System. (PX 75 at 3). During the telephone screenings, Ms. Lichter learned that K.A. had retail experience working at Target and she worked as a receptionist at her college. Ms. Lichter rejected K.A. and assigned disposition codes S3, sales and customer service experience, and S8, career compatibility. (TR 895; PX 75 at 8).

Ms. Lichter identified PX 76 as the application file for E.M. (TR 895). Ms. Lichter conducted his telephone screening. E.M. listed retail experience working at Target. He also worked at Abercrombie & Fitch. When asked to explain why K.A., who had worked at Target, got an S3 code, while E.M., who worked at Target, did not, Ms. Lichter said E.M. provided more details about his customer service experience. (TR 896-897). She agreed that according to E.M.'s resume, the work he did at Target and at Abercrombie & Fitch was in loss prevention, not in sales or customer service. (TR 897).

Ms. Lichter identified PX 77 as the application file for A.K. (TR 897-898). He worked at Best Buy as a salesman and as an assistant supervisor in the television department from September 2003 to April 2005. (TR 898). He described his duties at Best Buy in his application as greeting and assisting customers, executing sales and maintaining the appearance of the section. (PX 77 at 3). Ms. Lichter rejected A.K. and assigned an S3 disposition code for lack of sales and customer service experience. She said that while he "gave me some things at Best Buy he didn't articulate that he was responsible for sales." (TR 899).

Ms. Lichter identified PX 78 as the application file for W.L. (TR 899). He worked as a store cashier and bartender at a tavern and was a supervisory lieutenant at a security service. He described his experience to Ms. Lichter in the telephone screening by saying that as a bartender his job was to keep customers happy and in his security role he was the first point of contact for people entering the Walter Reed National Military Medical Center. Ms. Lichter considered his experience as a bartender to count as sales and customer service. (TR 900). Ms. Lichter sent W.L. on to a first interview. (TR 900-901).

Ms. Lichter identified PX 79 as the application file for D.Z. He listed on his application that he worked as a parts consultant at a car dealership and as a doorman at two taverns. Ms. [*52] Lichter did not conduct his telephone screening. (TR 901). The application file shows that D.Z. went on to a first interview and later received a rejection letter from Ms. Hardesty. (TR 901-902).

Ms. Lichter identified PX 81 as the application file for R.B. (TR 902). R.B. attended Morgan State University, which Ms. Lichter knew was a historically black college or university. She said she may have attended one or two recruiting events at Morgan State University and she attended one at Coppin State University, which is also a historically black college or university. She believed the only other historically black college or university in the area was the University of Maryland Eastern Shore or UMES. (TR 903). R.B. was an assistant manager at Kinko's, which included upselling to local businesses. Ms. Lichter scheduled R.B. to come in for an in-person first interview. (TR 904).

Ms. Lichter identified PX 90 as the application file for Dr. Sallamar Worrell, who testified previously at the hearing. (TR 905). The first job listed on Dr. Worrell's application was three months as a retail sales consultant at Sprint/Nextel where she assisted customers with their telephones, data and wireless internet selections. On the telephone screening form, Ms. Lichter wrote under sales and customer service experience, "Sprint — Tutoring — cust service, have to be personal & sell self or else parents will stop using you." (TR 905-906; PX 90 at 9). Ms. Lichter rejected Dr. Worrell for a lack of experience and entered the S3 disposition code on the form. Ms. Lichter said the primary duty of a tutor is not selling. (TR 906).

Ms. Lichter identified PX 82 as the application file for T.J. (TR 906). T.J. attended Temple University and had been working as a realtor since December 2003. On the application she said she had done 170 real estate transactions, including 89 sold residential properties. (TR 907-908; PX 82 at 4). On the telephone screening form, under sales and customer service experience, Ms. Lichter wrote that T.J. had been in real estate since 2003 and solicited sales by marketing homes to buyers and providing customers with information. Ms. Lichter rejected T.J. for a lack of sales or customer service experience and entered the S3 disposition code. (TR 908).

Ms. Lichter identified PX 83 as the application file for P.K. (TR 908-909). P.K. listed broker agent/customer service representative at Globallink and Sirshmoozalot Communications as her current job. She also reported working as assistant manager at an assisted living facility and she was the owner and operator of a merchandise marketing business. (PX 83 at 3-4). During the telephone screening, P.K. told Ms. Lichter that she worked for Hertz for nine years in reservations and executive sales. Ms. Lichter rejected P.K. for a lack of relevant work experience and entered disposition code S3. (PX 83 at 8). In the notes at the bottom of the form, Ms. Lichter wrote: "Worked at Hertz for 9 years is why she would fit in — only sales & c/s she talked about was in call center environment — Did not sell her experience, no enthusiasm." (PX 83 at 8). Ms. Lichter said screening applications sometimes required her to make judgment calls. (TR 911).

On cross-examination by Enterprise, Ms. Lichter said that an applicant putting sales and customer service experience on the application is not enough to move on to an interview; the applicant has to be able to articulate that experience and show its relevance to the management trainee position. (TR 912-913). She said her progression at Enterprise was management trainee, [*53] management assistant, customer service representative at the airport location, and recruiting administrator. She was working on a master's degree in human resources while she was the recruiting administrator. (TR 913).

With respect to the application of A.K (PX 77), she noted that he had a degree in marine biology, had worked as aquarist for two and a half years, and had applied for a job at the Baltimore Aquarium. Ms. Lichter agreed that indicated A.K.'s career interests were not compatible with a sales and customer service career at Enterprise. (TR 915-916).

With respect to the application of K.A. (PX 75), Ms. Lichter said K.A. did not convey to her any previous sales experience. She said just because she worked at Target, which is a retail store, did not mean she automatically got relevant work experience credit; she had to articulate what the experience was and how it related to sales and customer service. (TR 916-917). Ms. Lichter said two applicants could have held the same job with the same employer and one would get work experience credit and one would not depending on how it was articulated during the telephone screening. (TR 918). With respect to applicant P.K. (PX 83), Ms. Lichter said enthusiasm is important because often management trainees have to deal with people who are in bad situations where the employee being positive is key to good customer service. She said someone who did not convey enthusiasm but was able to articulate a great customer service experience would probably get sent on to an interview and, if not, it would probably be a rejection. (TR 918-919).

On redirect by OCFFP, Ms. Lichter said a telephone screening call lasted for about five to ten minutes and the purpose was to make sure the applicant met the minimum requirements. She said the only applicants that were not telephone screened were ones that Ms. Morris or Ms. Hardesty had already met during an on-campus visit, otherwise all applicants who appeared to meet the minimum requirements were screened to verify their sales and customer service experience. Ms. Lichter acknowledged that enthusiasm was not one of the core competencies Enterprise identified for the management trainee program. (TR 921-922).

Ms. Lichter said that she did not receive any training while working at Enterprise on implicit bias, unconscious bias or subconscious bias. She said she has been educated on it in the course of pursuing her master's degree in human resources. She defined implicit bias as "an inherent bias you may apply without realizing it." (TR 924). She started her master's program at Towson University in September 2006 and she graduated in December 2008, one month after she left Enterprise. She said she could not recall all of the courses she took because it was 12 years ago. (TR 928-929). She currently works for the University of Maryland Medical System. When asked if implicit bias could color ones perception of whether another person is enthusiastic, Ms. Lichter answered "[m]aybe." (TR 930). She agreed that when she conducted telephone screenings the applicants had different accents and dialects. (TR 930-931). She agreed that some candidates sounded like they were British and some sounded like their first language may have been Spanish. When asked if some candidates sounded like they might be African-American, she answered, "I could guess, but —". (TR 931). When asked if in reviewing an application there were sometimes indications that an applicant might be African-America, she answered, "[y]ou could make a guess." She agreed that a person's name or having attended a historically black college or university could possibly be an indication that the applicant was [*54] African-American. (TR 931). Ms. Lichter agreed that in a situation where the screener is aware of the applicant's race the screener should be careful to avoid implicit bias. (TR 932).

On re-cross examination by Enterprise, Ms. Lichter said that P.K. (PX 83) was not rejected because he lacked enthusiasm, he was rejected for a lack of relevant sales experience. (TR 932).

In response to a question from me, Ms. Lichter said she made a conscious effort to find candidates that met Enterprise's requirements and neither race nor anything else was part of her thought process. She said Enterprise had a very diverse workforce and she wanted to find candidates that were going to be successful working for Enterprise. (TR 933).

Ms. Lichter returned to the stand as a witness called by Enterprise. (TR 935). She identified DX 1026 as the application file for T.B. (TR 935-936). Ms. Lichter conducted the telephone screening of T.B. and sent her on to a first interview. T.B. attended Morgan State University. (TR 938). Enterprise offered T.B. a position as a management trainee. (TR 938-939). Ms. Lichter was asked to look at PX 20A again. She identified it as an applicant flow log. One of the columns in the applicant flow log is entitled "race." T.B. appears on the flow log and it shows her hire date and has a "D" under race. (TR 939-940). Ms. Lichter reviewed PX 21, which shows that race code "D" indicates the individual is African-American. (TR 940). She agreed that there were many other applicants hired from Morgan State University. (TR 941).

Ms. Lichter identified DX 6904 as the application file for S.W. who was a graduate of Howard University.9 (TR 941). Ms. Lichter conducted the telephone screening of S.W. and she put him through to an in-person interview. S.W. successfully completed all of the interview stages and was offered a job as a management trainee. (TR 942). S.W. is shown on page 16 of PX 20A, which indicates that he was hired on April 30, 2007 and his race was African-American. (TR 942-943). Ms. Lichter said Howard University is a historically black college or university. (TR 943).

Ms. Lichter identified PX 4171 as the application file for P.M. who was a graduate of Morgan State University. (TR 943). Ms. Lichter conducted the telephone screening of P.M. and put her through to an in-person interview. P.M. successfully completed all of the interview stages and was offered a job as a management trainee. P.M. is shown on page 9 of PX 20A, which indicates that she was hired on April 16, 2007 and her race was African-American. (TR 944-945).

Ms. Lichter acknowledged that applicants from historically black colleges or universities were hired; that she mainly applied "S" disposition codes; she applied "I" disposition codes after an interview was conducted; and she conducted telephone screenings and made decisions on which disposition codes should be applied to applicants she rejected. (TR 946). She denied ever making a decision about an applicant based on race and denied that there was any policy or instruction from her supervisors concerning the race of applicants. Ms. Lichter said the [*55] disposition codes she chose and entered reflected the true reason she rejected the applicants. (TR 947).

On cross-examination by OFCCP, Ms. Lichter agreed that the minimum requirements for management trainees changed on August 1, 2008 and as of that date a college degree was a minimum requirement. She agreed that the criteria shown in PX 10A reflected the minimum requirements she used to screen applicants prior to August 1, 2008 and the criteria shown in PX 10B were the minimum requirements she used after August 1, 2008. (TR 949). One of the minimum requirements throughout her tenure at Enterprise was work experience in sale and/or customer service. (TR 949). There was no minimum amount of work experience required. After Ms. Lichter was trained, Ms. Morris did not review her decisions on whether applicants met the work experience requirement. (TR 950).

Ms. Lichter was deposed on August 18, 2017. (TR 950-951; PX 52). Ms. Lichter acknowledged that on direct examination she had agreed that Enterprise hired plenty of graduates from historically black college or universities. (TR 952). She said she did not have a specific number in mind to define "plenty" and she agreed that every applicant has the right to be considered on the basis of his or her qualifications without regard to race. (TR 953). Ms. Lichter testified that she was never trained that a certain level of minorities was required nor was she given monthly data on minority hiring. She said she was encouraged to hire qualified candidates and there was no discussion about race amongst the recruiting department staff. (TR 954-955).

Ms. Lichter reviewed PX 112, which shows the adverse impact analysis for August 1, 2006 through July 31, 2007. (TR 956-957). Ms. Lichter agreed that everyone at Enterprise in a professional position started out as a management trainee since they promote from within and she agreed that the standard deviation for the adverse impact for that period was 2.96 and the impact ratio was .66. (TR 957-958). The next period was August 1, 2007 through April 30, 2008. Ms. Lichter agreed that it showed a standard deviation of 4.17 and an impact ratio of .47. (TR 958).

Ms. Lichter testified that she did not take any steps to counter implicit bias because she did not believe she applied implicit bias. (TR 959). She said she was confident that she did not intentionally discriminate against African-American applicants and she applied the same standards to African-American applicants and white applicants. (TR 960). Ms. Lichter agreed that perhaps an interviewee's enthusiasm during the telephone screening might be influenced by how comfortable he or she felt talking with her on the telephone, but she said if a person is applying for a job he or she should be able to convey interest and enthusiasm. (TR 962-963). She agreed that enthusiasm was not listed as one of the minimum requirements under the criteria in PX 10A or PX 10B. (R 963).

On redirect examination by Enterprise, Ms. Lichter agreed that from 2006 to 2008 the changes in the minimum requirements as shown in PX 10A and PX 10B included the requirement for a four-year degree rather than 48 college credit hours and for one year of relevant sales and/or customer service experience within the previous three years if the applicant [*56] was not a business or communication major rather than just sales and/or customer service experience. (TR 963-965).

ROBERT MCGRAIN

Robert McGrain was called as a witness by OFCCP. He started working at Enterprise in 1997 as a management trainee right out of college. He moved up to management assistant and then assistant manager before becoming a local corporate account manager in 2000. (TR 967-968). From there he became an area manager and, in 2007, he assumed his current position as the group truck manager. He said his sales and customer service experience prior to joining Enterprise was from working part-time in the fast food industry while he was in college. (TR 969). He was deposed on August 18, 2017. (TR 970; PX 56). He acknowledged that at his deposition he testified that he did not have any customer service experience prior to joining Enterprise and that his only sales experience was delivering newspapers. (TR 972).

Mr. McGrain said he conducted second interviews when he was an area manager. (TR 972). He agreed that applicants he interviewed had already passed a telephone screening and a first interview. He said he could reject an applicant at a second interview if he found the applicant lacked a core competency. If there was a secondary review of his decision Mr. McGrain was not aware of it. (TR 973). If he sent an applicant on to a third interview it went to group manager Andy Wendler and if he rejected an applicant the file went back to human resources. (TR 974). Mr. McGrain said that by the time an applicant gets to a second interview he or she already been screened to determine whether he or she had customer service experience, so the second interview was more of an assessment of whether the applicant was someone who was likely to be successful at Enterprise. (TR 975).

Mr. McGrain reviewed PX 15, which is the participant's guide for behavioral interview training. (TR 975-976). He said he attended the training and the focus was to look at the core competencies and assess whether the individual was likely to be successful working at Enterprise. As an example of the type of question he might ask to determine if an applicant is results driven, he said he might ask an applicant to "tell me about a time where you've had a goal that you accomplished?" He said he would assess the response based on whether the applicant was able to articulate his or her goal and to tie it to some benefit. (TR 976). He agreed that one of the things he looked at in second interviews was the applicant's ability to sell himself or herself as a perspective employee. (TR 977).

Mr. McGrain said that when he was an area manager he was aware of Enterprise's affirmative action plan and the metropolitan statistical area ("MSA"). (TR 977). When he did second interviews, he was aware of Enterprise's minority hiring goals, which he learned about in department head meetings where they discussed where Group 18 stood with respect to the MSA. Branch managers did not attend those meetings and Mr. McGrain said he did not recall the specifics of the discussions on minority hiring because it was 12 years ago or more. (TR 978).

Mr. McGrain reviewed PX 112, the adverse impact analysis for fiscal years 2007 and part of 2008. He did not recall anyone telling him that in those years there was an adverse impact on African-American management trainee applicants. (TR 979). He did not recall ever receiving [*57] instructions on specific steps to take to try and reduce the adverse impact on African-American applicants for the management trainee program. (TR 979-980).

On cross-examination by Enterprise, Mr. McGrain said he was hired as a management trainee in Group 16 in the Washington, D.C., area under that group's criteria. He said he did not recall seeing the adverse impact analysis, PX 112, before the hearing. (TR 980-981).

On redirect examination by OFCCP, Mr. McGrain said it certainly helps for a management trainee to come into the job with customer service experience, although he acknowledged that he had been successful in his career at Enterprise despite coming in without it. He said Enterprise provided him training on sales, customer service, marketing, use of its computer systems, and how to conduct behavioral-based interviews. He assumed, but was not sure, that the core competencies and minimum qualification screening requirements for management trainees came from the Enterprise corporate office. (TR 982-983).

On re-cross examination by Enterprise, Mr. McGrain said that when he was in college he worked as an assistant manager at a taco maker in a mall, a shift manager at a cookie shop, for a painting company, and delivering newspapers. (TR 983-984). He agreed that the two food service jobs were customer-facing and that it was 21 years ago, back in 1997, when he joined Enterprise. (TR 984).

In response to additional questions by OFCCP, Mr. McGrain agreed that in his deposition testimony he said he "did not have a tenured background in customer service and sales" and that he would forward someone on in the hiring process who did not have customer service experience if the applicant was able to communicate things that are important to customer service. (TR 986-987; PX 56 at 39-41).

DR. JANICE FANNING MADDEN

OFCCP called Dr. Janice Fanning Madden as an expert witness. Dr. Fanning is a professor of sociology, real estate and regional science at the University of Pennsylvania and a consultant for Econsult Corporation in Philadelphia, Pennsylvania. Her area of expertise is labor and urban economics and the econometrics and statistics relevant to those topics. (TR 1018-1019). Dr. Madden said that urban economics is the study of cities and the urban geography and how labor markets and conditions in urban labor markets affect outcomes by workers demographically and how changes in the demographic composition of workers affect how cities function. She said labor economics is the study of the workforce; who enters the workforce and how they are paid, the study of unions, and the study of human capital and how workers acquire skills and credentials. (TR 1019).

Dr. Madden said that she conducted a study to investigate whether Enterprise Group 18 discriminated against African-American applicants for the management trainee program from 2007 to 2017. She applied controls for minimum qualifications, including the requirement for a four-year college degree, sales experience, customer service experience and a driver's license. (TR 1019-1020). After controlling for these variables, Dr. Madden said African-American applicants were less than half as likely as white applicants to receive job offers. Her findings [*58] were summarized in a written report. (TR 1020). Her initial report, prepared in December 2017, is PX 105. (TR 1021). She identified PX 106 as a report prepared by the expert for Enterprise, Dr. Paul White. (TR 1021-1022). Her rebuttal to Dr. White's report, which she wrote in April 2018, is PX 107. (TR 1022).

Dr. Madden's curriculum vitae is at Attachment A to PX 105 and it is dated October 2017. She said that since she submitted her curriculum vitae she has published a paper in the Journal of Housing Economics on the paradox of growing black ghettos while metropolitan areas are integrating. The paper was published in June 2018 shortly before the hearing. (TR 1023). Dr. Madden earned a Ph.D. in econometrics, labor economics and urban economics from Duke University in 1972. She said econometrics are statistics commonly used in the profession of economics. She has a bachelor's degree in economics and mathematics from the University of Denver. She has been a faculty member in various schools and departments at the University of Pennsylvania since earning her Ph.D. She was recently appointed an assistant professor of regional science in the Wharton School, the business school at the University of Pennsylvania, which Dr. Madden said is the "[b]est in the world." (TR 1024-1025). The University of Pennsylvania is a private school and part of the Ivy League.

Dr. Madden's dissertation was on the economics of sex discrimination and it was published as a book. She has written numerous articles on race, ethnic and gender discrimination in labor and housing markets. She has been retained as an expert witness regularly since 1973 or 1974. (TR 1025). Some of the reports she prepared as an expert witness were accepted by some of the premiere journals in economics after going through a referee process. Dr. Madden said a referee process involves submitting a paper to a journal who then sends it out to a panel of experts who assess whether it is accurate and advances the field of study. The process is also known as being "peer-reviewed." (TR 1026). The Federal Reserve Bank of Philadelphia and the Federal Judicial Center in Washington have hired Dr. Madden to train judges on the use of statistics in discrimination cases. (TR 1026-1027). Dr. Madden's research has been funded by the National Science Foundation, the National Institutes of Health, the Department of Labor, and the Brookings Institution, among others. (TR 1027). She was one of the founders of Econsult in 1980 and she has testified as an expert 50 to 60 times over the past 45 years. (TR 1028). She has testified for both plaintiffs and defendants.

Dr. Madden had assistance from her colleagues at Econsult who helped "with the actual data crunching." She was assisted in this case by Rachel Carse, a Ph.D. candidate at Temple University, Tabatha LuPinetti, who has a degree in computer science and has been with Econsult for about 25 years, and colleague Dr. David Crawford. (TR 1030-1031).

Dr. Madden chaired the board of the American Academy of Political and Social Sciences from 2011 until earlier in 2018. (TR 1034, 1043; PX 105 at 50). She worked with attorney Johnny Cochran on a study of the success of African-American coaches in the National Football League ("NFL") despite few having been hired at that time. That work led the NFL to implement the "Rooney Rule," which requires teams to interview a minority candidate any time they are seeking to fill a head coaching vacancy. (TR 1041-1042).

[*59] Dr. Madden was tendered by OFCCP as an expert in the field of labor economics, urban economics and statistics. (TR 1046).

Enterprise conducted voir dire of Dr. Madden. She said she has testified as an expert witness four times in the past four years, and two of the four times she testified for OFCCP. (TR 1047-1048). In the other case where Dr. Madden was the expert for OFCCP, Dr. White was the expert for the opposing party. (TR 1048). When she was a full-time faculty member, the University of Pennsylvania limited her to not more than one day of work as an expert witness per every seven days while school was in session. Now that she is a part-time faculty member, Dr. Madden said half to three-quarters of her time is devoted to her work as an expert. She said she has only testified four times since November 2013 because most of the cases she is involved in end up settling. (TR 1049).

Enterprise had no objection to Dr. Madden as an expert in the fields of labor economics and statistics, but noted that it did not appear she relied upon expertise in urban economics in her work on this case. OFCCP concurred and Dr. Madden was accepted as an expert in the fields of labor economics and statistics. (TR 1049).

Dr. Madden briefly recounted some of the cases in which she testified as an expert in discrimination cases. She said her testimony had not been limited by the Daubert standard.10 She estimated that over the past five years there were about ten cases where she analyzed racial data and did not end up testifying in a hearing or trial. She said that at present she is involved in about 25 cases. (TR 1049-1051). Econsult is being paid for Dr. Madden's work in this case and bills $690 per hour. (TR 1051-1052). In this case, she reviewed discovery responses, deposition testimony, and files provided by Enterprise. This included a set of electronic data files and PDF job applications. (TR 1052-1053). She also reviewed information under the heading "Other Sources," which she said was financial data used to calculate damages.

Dr. Madden prepared a PowerPoint slide presentation on the results of her study as a demonstrative aid for her testimony. (TR 1016, 1054; PX 113). She said the central question she studied was whether there were racial differences in the likelihood of getting hired for a management trainee position at the Enterprise Baltimore group within the Baltimore metropolitan statistical area. If she found racial disparities she would also study: (1) which stage in the hiring process accounts for the racial differential; (2) whether African-Americans have inferior credentials that account for the differences in hiring; and (3) the monetary damages experienced by equivalently qualified African-American applicants who were denied job offers. (TR 1054-1055). Dr. Madden said she found that Enterprise made 220 fewer job offers to African-American applicants than it would have if offers had been extended to African-Americans and whites at the same rate. She called this a shortfall of 220 African-American offers. (TR 1055; PX 113 at 1-4).

Dr. Madden said she looked at whether the shortfall of African-American job offers could be consistent with random variation. She found that the statistical probability of the shortfall occurring by chance in the absence of consideration of race was one million times less likely [*60] than a ticket winning the Powerball Lottery. She said the African-American job offer shortfall not being due to chance was a virtual certainty. (TR 1055; PX 113 at 5). She found that the application screening and telephone screening stage (stage one) accounted for the greatest share of the racial differential, although the interview stage (stage two) and background check stage (stage three) also contributed. There were 449 fewer African-American applicants than expected who made it through the application screening and telephone screening stage. Dr. Madden explained that the 220 shortfall in offers and the 449 shortfall in getting through the screening stage are consistent because more people get past the screening stage than ultimately get job offers. She said the odds of the 449 shortfall in African-Americans getting through the screening stage (stage one) absent consideration of race was one million times less likely than winning the Powerball Lottery. (TR 1056-1057; PX 113 at 6-8).

Dr. Madden found that Enterprise selected 91 fewer African-American applicants in the interview stage (stage two) than they would have if African-Americans were selected at the same rate as African-American and white applicants combined. The odds of a shortfall of 91 African-Americans by chance without consideration of race was as likely as the odds of winning the Powerball Lottery. (TR 1057-1058; PX 113 at 9-10). Dr. Madden used the combined success rate of all applicants as the basis for comparison to the success rate for African-American applicants. Had she compared the African-American applicant success rate to the rate for white applicants alone the disparity would have been greater. (TR 1058-1059).

Dr. Madden found that Enterprise selected 16 fewer African-American applicants at the background check stage (stage three) than it would have if African-Americans were selected at the same rate as African-American and white applicants combined. The odds of this happening by chance without consideration of race was two in 1,000 or 2.8 standard deviations. (TR 1060; PX 113 at 11-12).

Dr. Madden found a statistically significant disparity at each of the stages she studied and in most instances the disparity was so great as to constitute a virtual certainty. She assessed the credentials of the applicants, using regression analysis that alternately controlled and did not control for credentials reported on the applications, and she found no differences that would account for the difference in African-Americans proceeding through the application stage. Her conclusion was that racial differences in the credentials of the applicants do not account for racial differences in the hiring outcomes. (TR 1061-1062; PX 113 at 13-14).

Dr. Madden explained that it was impossible to do a regression analysis for the interview stage of the process because there was no independent objective information to assess. In her analysis, she assumed that African-American applicants and white applicants were equal absent evidence to the contrary and she said she looked at the data and there was no objective evidence to indicate that African-American applicants were systemically inferior to white applicants. (TR 1063).

The last part of Dr. Madden's study looked at damages. She said she looked at the shortfall of 220 job offers to African-American applicants, the likelihood of offers being accepted, the compensation those Enterprise hired received, the prospect that applicants were hired elsewhere, the racial disparity, and the IRS interest rate to bring the amount to present [*61] value. Her calculation was that as of December 2017, the total amount was $7,693,184. That amount represented back pay only, no front pay, and assumed the situation was remedied and there were no damages going forward. (TR 1064-1065; PX 113 at 15-16).

Dr. Madden said her study relied upon data provided by Enterprise. She received electronic applicant flow log data that included race information and the outcomes of the application process. She received it in Excel spreadsheet format. There were some applications where applicants did not provide information about race. Those were excluded from her study. (TR 1065-1066). The database included the date each application was received and that is the date Dr. Madden used rather than the date Enterprise made a decision on the application. The fiscal year at Enterprise runs from August 1 of one year through July 31 of the next year, and that is the fiscal year basis Dr. Madden used in her study. She noted that she did not have data for the last half of calendar year 2017 when she calculated damages, so her total reflects interest for that period, but no earnings damages for that period. The database indicated whether an applicant received an offer and it indicated which applicants self-selected out of the process. The self-select outs were not included in her study at the stage the applicant opted out of further hiring consideration. For instance, an applicant who self-selected out in the interview stage would be included in the screening stage analysis but not the interview stage or background check stage analyses. (TR 1067-1068). Dr. Madden said the self-selection out rate was higher among African-Americans, but she did not pursue it because it was not statistically significant. She noted that there were a large number of applications that indicated self-selected out, duplicate, where there were no duplicate applications in the database. In the end, she and Dr. White were both doing their analyses using the same group of applicants. (TR 1069-1070).

Once the database was refined, Dr. Madden conducted an analysis called Fisher's exact test. She explained that if the number of African-Americans and whites in the pool competing are known and you know the number that succeeded from that pool, you can put together all of the potential combinations of African-Americans and whites that could possibly occur and then compare those to what actually occurred. (TR 1070-1071). Dr. Madden said Fisher's exact test to measure racial discrimination is well established and she had used in litigation consistently for 35 years. (TR 1071).

Dr. Madden said that she and Dr. White had some disagreements on what should or should not be included in the database, but she accepted his position because it did not make a difference in the conclusions. (TR 1072; PX 113 at 17). The central question examined was whether the actual number of job offers to African-American applicants for management trainee positions was less than expected. That was the central question because you would expect a number that was racially neutral. Dr. Madden said she looked at the actual number of African-American successes and actual number of white successes, which was easy to determine. Then she had to calculate what the expected number of successes would be with racially neutral practices and then compare that to the actual number. If the actual number is less, then the hiring process is not racially neutral. (TR 1072; PX 113 at 18-19).

Dr. Madden said she compared the success rate of African-American applicants to the success rate of African-American and white applicants combined. A racially neutral process should produce equal rates. (TR 1073; PX 13 at 20). If there is a disparity, she said there are [*62] only two possible explanations: (1) African-Americans have systemically inferior credentials or qualifications to white applicants, or (2) African-Americans with equivalent qualifications or credentials to white applicants are less likely to receive a job offer. She said it could also be some combination of the two. (TR 1074; PX 13 at 21).

Dr. Madden explained that she got the number expected by taking the total number of African-American applicants and comparing it to the total number of African-American applicants and white applicants combined. For 2008, she said the data she and Dr. White analyzed showed there were 969 African-American applicants and 1,027 white applicants, for a total of 1,996 applicants. The African-American share of the total was 49 percent, which is the expected share of the job offers. Offers were extended to 31 African-American applicants and 104 white applicants, for a total of 135 total job offers in 2008. While African-Americans were expected to get 49 percent of the job offers, they received 23 percent. (TR 1074-1075; PX 113 at 22-25). The numbers cited in the slides come from Table 1 of Dr. Madden's rebuttal report dated April 6, 2018, which is PX 107. (TR 1075-1076). The pie chart on slide 27 of PX 113 shows that African-American applicants were expected to get nearly half of the job offers (49 percent) and they actually got less than one-fourth of the offers (23 percent). In 2008, there was a shortfall of 35 offers (66 expected offers minus 31 actual offers). (TR 1077).

In order to make sure the results were not due to chance or just happened to capture a bad year, Dr. Madden said it was necessary to use statistics and probability theory. (TR 1078; PX 113 at 30). Statistical significance reflects the confidence in the relationship that was measured. Social science considers a probability that something is false of five percent or less as significant. A five percent probability that something happened by chance is equal to 1.96 standard deviations. (TR 1079; PX 113 at 31). Dr. Madden used flipping a coin to illustrate the probability that something was not due to chance. She said that if you flip a fair coin you would expect it to land on heads half the time and on tails half the time. If you flipped the coin three times, there are eight combinations of results that could occur. Of those eight combinations, there are four where there are at least two heads and four where there are at least two tails. There is one of the eight possible combinations where the coin flip ends in three tails, so one in eight is a 12.5 percent chance of getting three tails by chance. If the number of flips in a set is increased to five and the coin lands on tails five times in a row, the odds of that happening by chance is 3.125 percent, which means a conclusion that the coin is loaded has less than a five percent chance of being wrong. (TR 1080-1082; PX 113 at 32-36). This type of calculation to determine the probability of an event is common. For example, Dr. Madden said the probability of any one ticket matching the five white balls and the one Powerball in the Powerball Lotter is 1 in 292,201,338. (TR. 1082-1084; PX 113 at 37-40).

Dr. Madden said the same concept used to examine the probability of particular results from coin flips applies to looking at all of the ways African-American applicants and white applicants can be combined to get to 131 applicants that received job offers. (TR 1084). The probability of 31 African-Americans job offers in 2008, according to Dr. Madden, was 0.000000000231, which is less than the 0.0000000342 probability of winning the Powerball Lottery. (TR 1084-1085; PX 113 at 39, 41; PX 107 at 19). She explained that standard deviation is something statisticians use to measure the distribution of outcomes. A set of observation falls within one standard deviation 66 percent of the time and within two standard [*63] deviations 95 percent of the time. The less probable an outcome is the higher the number of standard deviations. (TR 1085-1086). Dr. Madden said that the results reflected in Table 1, Column 5, of PX 107 showing the probability that the differences in the expected and the actual number of job offers for African-American applicants were due to chance show statistical significance because they all have less than a five percent probability of being the result of chance. (TR 1087).

Dr. Madden said that the more data that is analyzed the more accurate the results, so when you assess data for several years it yields a more accurate result than examining data for any single year. She said it is like her coin flip analogy: the probability of flipping a coin three times and getting three tails is 12.5 percent while the probability of flipping a coin five times and getting five tails is 3.125 percent. A small number of observations can have big difference while a large number of observations is sensitive to detecting relatively smaller differences. (TR 1088; PX 113 at 36).

In the 2007 through 2017 period, there were 5,620 African-American applicants for the management trainee program and 4,817 white applicants. Offers were extended to 458 African-American applicants and to 772 white applicants. (TR 1089; PX 107 at 19). Dr. Madden calculated a total shortfall of 220 African-Americans by subtracting the actual number of African-Americans who receive jobs offers from the expected number based upon their percentage of the total pool of applicants. The probability of the number of African-American job offers being the result of chance was, in Dr. Madden's calculation, one million times less likely than a single ticket winning the Powerball Lottery. It was "3.87, and you've got to put 41 zeros in front of that 3, so it's just virtually as close as you can get in statistics to being certain." (TR 1090). The odds are less likely than the odds of flipping a fair coin 45 times and having it land on tails 45 times. (TR 1091; PX 113 at 42). In her analysis, Dr. Madden assumed that the groups of African-American applicants and white applicants were similarly situated, but she did not assume each individual applicant within each group was similarly situated. (TR 1093). Dr. Madden's conclusion was that the shortfall of 220 job offers to African-American applicants was "certainly not due to chance." (TR 1094).

Next, Dr. Madden looked at the stages in the hiring process to see where the differences occurred. (TR 1094-1095). The first stage she evaluated was the application review and telephone screening stage (stage one). In 2007, there were 627 applications reviewed for whites and 755 for African-Americans. Of that number, 305 white applicants and 239 African-Americans moved on to the interview stage (stage two). Had African-American applicants moved on to the interview stage at the same rate as all applicants combined there would have been 297 African-Americans that moved on to the second stage. (TR 1097; PX 107 at 20). That results in a shortfall of 56 African-Americans passing the screening stage. (TR 1098). The probability of this racial distribution in 2007 was 8.59 with ten zeros in front of the eight and the standard deviation was 6.4. In the interview stage, there were 183 white applicants and 175 African-American applicants. (TR 1099). The reason there is a difference between the number that passed the screening stage and the number assessed in the interview stage is that some applicants chose to self-select out of the hiring process prior to an interview. (TR 1099-1100). Of the 183 white applicants that were interviewed, 83 passed. Of the 175 African-American applicant that were interviewed, 60 passed, for a shortfall of ten African-Americans moving on [*64] from stage two. (TR 1100). The probability of that result being by change was .02, which Dr. Madden said was statistically significant. (TR 1100-1101). In the background check stage, one white applicant and one African-American applicant were eliminated, so there was no racial effect in that stage. Dr. Madden said the application screening stage (stage one) had the biggest impact on the results and the interview stage (stage two) had some impact. (TR 1101).

Dr. Madden said the same pattern was true in each year she analyzed. (TR 1101). She said "[y]ou see again big huge probabilities of the racial disparity — of a racial disparity at the first stage of the application review and phone screen" in 2010, 2011 and 2012. There was statistical significance in individual years 2010 and 2012, but "not quite" in 2011. (TR 1102). In 2013, 2014 and 2015 there was "huge statistical significance in the screening," but not in the 2013 or 2014 interview stages. (TR 1113; PX 107 at 21-22). There was no racial difference in the background checks in any of those years. The pattern persisted in 2016 and 2017. (TR 1104; PX 107 at 23).

Looking at the 2007 through 2017 period as a whole, Dr. Madden said 2,336 African-American applicants should have made it through the screening stage, but only 1,887 did, leaving a shortfall of 449 African-Americans.11 Dr. Madden said the probability of that happening by chance was 1.63 with 68 zeros in front of the numeral one. In the interview stage, the expected number of African-Americans to succeed was 675 and the actual number was 584 for a shortfall of 91. The probability of that occurring by chance was 2.07 with ten zeros in front of the two. (TR 1105; PX 107 at 24). In the background check stage, the expected number of African-American applicants to pass was 474 and the actual number was 458, which Dr. Madden said was statistically significant with a standard deviation of 2.8. (TR 1106; PX 107 at 24).

Dr. Madden said that not having access to information about a particular applicant's race would not preclude racial bias because an applicant's name, address, social and professional activities, and the schools he or she attended can convey the same information.12 (TR 1113-1114). Dr. Madden said she did not assume that all applicants were similarly situated, she assumed that African-American applicants as a whole and white applicants as a whole were, on average, similarly situated. She analogized it to a medical study where one half of a group receives treatment and the other half is untreated: It is not assumed that each member of each group is identical, it is assumed that each group as a whole is similar to the other group as a whole. (TR 1119-1120). Here, she said she ran a regression analysis to test to see if the assumption that the groups did not differ in relevant characteristics was accurate. (TR 1120-1121). Dr. Madden explained that a regression analysis measures the impact of explanatory variables on the dependent variable and it results in probabilities that certain characteristics predict who succeeds. (TR 1121-1122; PX 113 at 51-52).

[*65] Her first regression had no controls and only looked at the probability of African-Americans failing at the application screening stage. It showed that African-Americans had a greater probability of failing to move to the next stage than whites. The standard deviation was 14.43. (TR 1123; PX 113 at 53). Then, Dr. Madden ran a regression with controls for college degree, meeting the age requirement, holding a driver's license, seeking full-time employment, possessing a visa or other legal authority to work in the United States, plus the applicable fiscal year, and she found that the probability of failure for African-American applicants increased. (TR 1124). Finally, Dr. Madden ran a regression that added controls for sales and customer service experience. She said she and her team researched the job titles that Enterprise provided to them and they compiled a list of those that appeared to be associated with sales and customer service work. She said some job titles could be added to or taken out of the list, but the same job titles counted as a yes or a no for all of the applicants regardless of race. (TR 1125). When sales and customer service controls were added, the probability of failure for African-American applicants declined slightly, but was still well above the non-controlled probability. (TR 1127). As she added controls to the regression the standard deviation declined. She said that was because taking out some of the data to estimate the effects of other characteristics results in "fewer coin flips." (TR 1128). Dr. Madden identified PX 109 as some of the application files that she reviewed and referred to in her rebuttal report. (TR 1129-1130; PX 109).

Dr. Madden said it was impossible to do a regression analysis on the interview stage because of the absence of objective data. For instance, in the screening stage, whether applicants have driver's licenses is a yes or no criteria. In the interview stage, however, individuals are making decisions based upon subjective judgments and whether race influenced those judgments is the question and the heart of the study. (TR 1130-1131). Dr. Madden did not include Enterprise's disposition codes in her regression analysis because the codes are explanations for judgment calls people made rather than objective data. (TR 1131). She only used information from the applications. (TR 1132; PX 107 at 25).

To calculate damages, Dr. Madden assessed the shortfall in offers extended to African-American applicants, the percentage of applicants that accepted offers, compensation paid to those hired as management trainees, attrition and a measure of compensation from other employers unsuccessful applicants earned. (TR 1134). She also used IRS-approved interest rates to bring the totals to present values as of the date of the report in 2017. (TR 1135; PX 107 at 26-27). Dr. Madden said she used the figures provided by Dr. White to calculate damages. He used assumptions drawn from government reports, which Dr. Madden said was "quite reasonable." (TR 1137). The hiring shortfall was calculated as 191.7, which was multiplied by the average loss minus mitigation, and then interest was added to reach a total of $7,693,184 as of December 31, 2017. (TR 1137-1138; PX 107 at 27).

On cross-examination by Enterprise, Dr. Madden said that she had been retained by OFCCP in five to eight cases over the past five years. She often starts out in a case as a consulting expert. (TR 1139). She said that on occasion she finds no statistical disparity in parts of the allegations raised by OFCCP, but she did not recall finding a total absence of disparity in any OFCCP case in the past five years. (TR 1140). When asked if she believed OFCCP could rely upon her to substantiate claims of discrimination, Dr. Madden said "No. I don't think I've done that. Many times I've taken a lot of their cases off the table, so no." (TR 1140-1141). She [*66] could not provide an estimate of how much she had charged the Department of Labor over the past eight years for serving as an expert because her firm does the billing. (TR 1141). Dr. Madden reviewed page 55 of PX 105, which is entitled "Expert Testimony of Dr. Janice F. Madden Since November 2013." It lists three cases. Dr. Madden testified that she remembered one other case, for a total of four cases. (TR 1142). She said she was retained by the plaintiff in all four of the cases in which she testified since November 2013. She said she has done work for defendants, too, but all of those cases settled as have many cases where she worked for the plaintiffs. (TR 1142).

Dr. Madden reviewed PX 107, pages 57-59, which was a list of the information she reviewed and considered in preparing her analysis and report. (TR 1143). She said that she was not aware that the ICIMS computer system automatically knocked out individual applications that did not meet the minimum requirements. (TR 1143-1144). Dr. Madden believed that OFCCP retained her in this case in early 2016. (TR 1144-1145). She did not recall OFCCP giving her legal or factual assumptions to apply to her analysis and she said she would have disregarded the assumptions if they had because she was doing an economic and statistical analysis. (TR 1145). Dr. Madden identified DX 188 as a transcript of her deposition testimony. (TR 1146-1147). She agreed that at her deposition she answered "[n]ot that I recall" when asked if OFCCP gave her any factual assumptions to apply to her statistical analysis contained in her original report. (TR 1147; DX 188 at 16). She agreed that she did not consider the notes that appeared in application files. She said she did not consider them because they reflected subjective decisions Enterprise employees made, which was the subject of her analysis. (TR 1148). She acknowledged that she relied upon the job titles applicants used in their applications to assess their sales and customer service experience and she did not consider any notes they may have written describing their job duties. (TR 1149).

Dr. Madden said the work product OFCCP requested was a written report. OFCCP did not ask her to perform any analysis of implicit bias. She acknowledged that she had no evidence that any of Enterprise's recruiters harbored any animus towards African-Americans other than the hiring outcomes, which she concluded showed a racial disparity that was based on racial bias. (TR 1150-1151). Dr. Madden defined "racial bias" as when "you make decisions differently for one race than you make for another." (TR 1151). She said she had published extensively on discrimination from an economist's perspective and not from a legal perspective. (TR 1151). She said she had no legal training on discrimination, but she assumed from the fact that she was invited to train people in the legal profession that her perspective on discrimination had legal relevance. (TR 1152).

Dr. Madden reviewed PX 10A. (TR 1153). She agreed that the shortfall analysis in Table 1 showing a shortfall of 220 African-Americans in hiring was the basis for her damages calculation. (TR 1154). She agreed that she did not account for the eight factors listed on PX 10A in her analysis and instead assumed the characteristics were the same by race. (TR 1154-1155). She said the same applied to the eight factors shown in PX 10B. (TR 1155-1156). Dr. Madden testified that in Tables 1 and 2 she assumed that as a group African-American applicants and white applicants were equally likely to have sales experience and then in Table 3 she tested that assumption. (TR 1157). She said she begins with the assumption that the races are equal, absent evidence to the contrary, and then she looks for evidence to the contrary. (TR 1157-1158). [*67] Since there was a "massive racial difference," she thought there was going to be strong evidence to the contrary and she looked for that evidence. (TR 1159). She agreed that she only looked at the information provided in the applications and did not consider notes added by screeners or interviewers during the process because she only considered objective evidence and did not consider subjective judgments made by Enterprise. (TR 1159). Dr. Madden said people conducting interviews do the best they can when they take notes, but it is not a verbatim transcript and they make selective judgments about what to write down. (TR 1159-1160).

Dr. Madden was asked if there were any studies that supported her initial assumption that on average African-American applicants and white applicants had the same sales experience. She said her Table 3 showed the assumption was accurate. When asked again if there were studies that supported her assumption she said that in her 45 years of experience, whether working for plaintiffs or defendants, experts start with the assumption that groups are equal and then look at what the differences are and see if they can explain the differences. (TR 1161-1162). She assumed that for the eight criteria, including education and sales experience, African-American applicants and white applicants were equal in Tables 1 and 2 and then she tested the assumption in Table 3. (TR 1162).

Dr. Madden reviewed her initial report, which is dated December 14, 2017. (TR 1162; PX 105). The four bullet-points on page 4 are the principal results that she wanted to emphasize. The report also has results by year and by the different credentials. (TR 1163; PX 105 at 4). She said she did a report by year because changes in the economy over time may influence the shortfalls. (TR 1164). She agreed that with respect to shortfalls she could not conclude that the results were based on racial animus of the interviewers. (TR 1165). Dr. Madden wrote in her report that her analysis assumed "a racially neutral system or step in the hiring progression is one in which African-American applicants experienced the same probability of proceeding through the hiring process as comparable white applicants do." (TR 1166; PX 105 at 5). She said she did not assume that individuals were the same, she assumed that on average African-American applicants and white applicants as groups have the same distribution of qualifications across the board. (TR 1166). She assumed the two groups, as a whole, were equally qualified. (TR 1167).

Dr. Madden said her presumption that an equal percentage of African-Americans should proceed through the various steps was not the same as a quota. She explained that statistically there can be differences in success rates as long as they remain within the realm of chance while a quota mandates a specific number without regard to chance. (TR 1168). Dr. Madden did not believe that straight quota hiring was permitted by law. (TR 1169). She used the same database as Dr. White, which removed applicants who self-selected out of the hiring process. She said that relied on the entry of self-select out codes by Enterprise, which she believed was an objective decision and not a judgment call as was the choice of disposition codes during screening or interviews. (TR 1170). Dr. Madden did not control for criteria changing over the 2007 to 2017 period, such as the evolution of the work experience requirement, because she assumed there would be no difference between African-American applicants and white applicants. The only thing she analyzed under experience was job title and that analysis was done by computer. (TR 1173). She said that she and her team looked at some of the notes applicants entered to explain their work experience in order to facilitate development of a list of relevant job titles, but then the computer only analyzed job titles based upon the list they [*68] compiled without analyzing any notes. (TR 1174). She could not quantify how many of the notes were reviewed, but she said she was confident that her research assistants continued reviewing them until they were satisfied that they were no longer getting useful information that would aid in developing their list of job titles. (TR 1177-1178). Dr. Madden's analysis for the years from 2008 through 2017 did not control for one year of sales experience; instead, she assumed there was no difference between the racial groups as a whole. (TR 1179).

Dr. Madden said that everyone who expressed interest in the management trainee position, except for those who self-selected out, were included in her shortfall analysis, but who is pulled in or out of the pool does not matter because the African-American share of the applicant pool remains the same and leads to the same result. (TR 1182). She agreed that an employer can set minimum qualifications for positions in its organization provided the qualifications are applied consistently, which was not the case here. (TR 1183). Table 3 of her initial report looks at all of the minimum requirements and finds that if consistently applied more African-Americans would have cleared the screening and interview stages. (TR 1184; PX 105 at 28). The analysis controls for African-American applicants and white applicants having comparable traits and then assesses whether race matters. (TR 1187). Dr. Madden assumed the two groups as a whole were comparable for all of the minimum requirements. (TR 1188). She took the data on qualifications directly from the applications and she did not attempt to verify whether the information the applicants provided was accurate. (TR 1188-1189).

Dr. Madden examined her rebuttal report. (TR 1191; PX 107). The shortfall analysis is in Table 1 at page 19 of the report. She agreed that she eliminated from the pool of applicants those who were not qualified because of a DUI conviction. (TR 1192). She did not eliminate from the pool applicants that did not have a driver's license. She did not account for the fact that information on some applications might be incorrect because she assumed that would be similar across the board and that one group did not inherently lie more than the other group. (TR 1193-1194).

Dr. Madden was asked to review DX 5417, the application file for C.T.13 C.T. marked on the application that she had a valid driver's license. (TR 1194; DX 6417 at 3). The telephone screening form indicates that Mr. Wucher conducted the telephone screening and he marked "yes" on the screening form in the block for valid driver's license. He wrote in the comments that C.T. had not had a license for more than one year as required for the management trainee position and he rejected her at the screening stage with an S12 disposition code. (TR 1194-1195; DX 6417 at 5). Dr. Madden said she never received any information that a valid driver's license for more than one year was a minimum requirement. (TR 1197).

Dr. Madden agreed that she did not exclude from her analysis applicants who were not qualified due to felony convictions, lack of eligibility to work in the U.S., or lack of a degree, and that she did not conduct any analysis of whether evaluation of those factors was racially motivated. (TR 1198-1200). She used the information on the application to determine whether the applicant did or did not have a college degree. (TR 1200). She said whether an applicant had a degree at the time the application was submitted or instead would have a degree by the [*69] time he or she started work at Enterprise was immaterial to her analysis because she assumed it would apply equally to both groups. (TR 1201).

Dr. Madden said Enterprise treated sales and customer service experience as a preferred qualification rather than a minimum qualification because it selected some applicants who did not have the required sales and customer service experience. (TR 1202-1203). She listed seven applicants in her rebuttal report that she concluded were hired despite no sales or customer service experience. (TR 1203-1204). In addition, she discussed two white applicants — C.T. and D.Z. — Enterprise determined met the minimum requirement for sales and customer service experience where she assessed that they lacked such experience. One of them was credited with relevant experience for work as a bartender and as a supervisory lieutenant and the other for work as doorman and as a parts manager at an auto store. (TR 1204; PX 107 at 6). Dr. Madden reviewed PX 78, the application file for W.L. (TR 1204). W.L. described his work experience as a cashier and bartender at a tavern to include "work with a variety of customers … customer service." (TR 1205; PX 78 at 4). Dr. Madden said she believed this would be customer service experience, but not sales experience. She believed the minimum requirement was for experience in both areas. (TR 1206). Dr. Madden reviewed PX 79, the application file for D.Z. (TR 1207). In D.Z.'s description of his duties as a parts consultant he wrote that he had to "[l]ook up parts based on description given by technician, front counter customer, or wholesale customer on phone. . ." (TR 1208; PX 79 at 2-3). Dr. Madden said that was not sales experience. (TR 1208).

Dr. Madden reviewed PX 108, the application files for four of the applicants mentioned in her supplemental report. M.H. submitted a resume with her application and Dr. Madden acknowledge that she only used information from the application and did not review any of the resumes. One of the items M.H. listed was work as a private softball instructor. (TR 1209). Dr. Madden disagreed that being a private softball instructor was a job requiring sales. She said choosing occupational codes is an art, not a science, and jobs could be included in the list or not as long as they are used consistently for everyone. She said she treated every application the same and if someone could show her that the job descriptions would cause a different result she "would say uncle." (TR 1210-1211).

G.H listed work as a runner, bar-back and busser. Dr. Madden said that was not customer service experience because those are behind the scenes jobs that do not routinely require interaction with customers. (TR 1212-1213; PX 108 at 65-66). K.M. only listed two jobs — one, laboratory technician operator and trainer, and two, insulation assistant and laborer — in his application and in his resume. (TR 1213-1214; PX 108 at 116-117, 122-123). J.R. listed two jobs in his application — maintenance manager and painter — and he did not provide a resume. (TR 1215; PX 108 at 141-142). Dr. Madden reiterated that she only looked at information provided in the applications. (TR 1215).

Dr. Madden acknowledged that she accepted what an applicant reported on his or her application about moving violations and did not check anything else to see if the information was true. She said she did not do so because Enterprise did not use that criteria consistently over time. (TR 1217-1218). She cited to her initial report that stated:

[*70] I include all the qualifications used in the application review for the entire period, August 1, 2006 through July 31, 2017. The self-disclosed information on DWI/DUI convictions or drug or alcohol-related offenses, felony conviction records, and loss of driver's licenses are not used in the regression analyses because Enterprise stopped asking the questions about these matters on the applications after 2012 and/or 2013. I presume they stopped asking because the information yielded was inaccurate or not important to the process.

Enterprise asked why Dr. Madden assumed African-Americans and whites are comparable in leadership ability. She said that in social science the basis of discrimination analysis is that the races are presumed equal absent evidence to the contrary and no objective evidence was presented to her that African-Americans who were in other relevant respects comparable to whites were systemically inferior in leadership ability. (TR 1226-1227). She said she did not consider in her analysis the explanatory notes written by Enterprise's screeners and interviewers on applicants' leadership abilities because those were the subjective decisions being analyzed to determine whether there was racial disparity. (TR 1227). She explained that there are subjective differences across individuals and there is nothing wrong with a company considering those differences, but when there are disparate outcomes by race between people who are objectively the same there is a problem with how subjective criteria were applied. (TR 1229). She said: "I mean, I want to see some evidence about why is it, who else is finding Baltimore that African-Americans are so inferior to whites when they've got the same characteristics? I mean, this is just an astounding difference in this case. . . . This is no small difference." (TR 1230).

Dr. Madden agreed that she only considered the eight factors Enterprise evaluated to determine whether applicants met its minimum requirements and said she did so because those were the only factors where she had consistent evidence to evaluate. (TR 1231). She said she did not presume there was or was not racial animus on the part of those who conducted interviews for Enterprise. (TR 1232). She agreed that her calculation of a shortfall of 220 job offers to African-Americans was not calculated from her regression analysis. (TR 1232; PX 107 at 25). She said using Fisher's exact test gives an exact number while regression analysis gives a probability, which requires establishing a cutoff, but the results would probably be similar. (TR 1232-1233). Dr. Madden agreed that her analysis as reflected in Tables 2 and 3 of her rebuttal report (PX 107) combined application review and telephone screening as a single stage. (TR 1233). She said she combined them because she was looking at the outcome of moving on to the interview stage. She agreed that she only looked at the eight minimum requirements criteria in conducting her analysis. She said she did so because those were the criteria used over the entire time period and any of the criteria that had been eliminated she assumed Enterprise had determined were not useful in selecting management trainees. (TR 1234-1235).

[*71] Dr. Madden agreed that Dr. White said that one of her models had a pseudo R-square of approximately 11 percent, but she added that she had published a lot of articles with lower R-squared values. (TR 1235). She described pseudo R-squared as the percent of variance explained, but not as a measure of reliability. (TR 1236). A higher R-squared value means that what you have in the model is more correlated with the outcome. (TR 1237).

Dr. Madden did not consider whether an applicant lied on his or her application and assumed that lying would be equal by race and she would "find it incredible that one would assume that blacks lie more than whites as a justification for employment decisions." (TR 1271).

Dr. Madden reviewed DX 513, the application file for L.B. She agreed that on the application L.B. marked "no" in response to a question that asked whether his driver's license had been revoked, denied, suspended or canceled within the previous five years. (TR 1273; DX 513 at 5). She agreed that the document from the background check showed that L.B.'s driver's license was suspended within five years of the date of his application. She noted, however, that the record showed the letter informing L.B. of the license suspension was "returned by postal authority," so L.B. may have been unaware of the suspension and answered the question truthfully on the application. (TR 1274-1275; DX 513 at 18-19).

Dr. Madden reviewed DX 829, the application file for L.B. She testified that she would not characterize the names Louis or Lester as African-American names. (TR 1275-1276). She agreed that in section two of page two of L.B.'s application he marked he had graduated from the University of Maryland and in the notes on page seven from his December 2013 interview it states that he will not get his degree until December 2014. (TR 1276-1277). She did not agree that her failure to review the notes led her to mistakenly assume L.B. had a college degree. She said she considered the information everyone reported on their applications and analyzed whether people with the same applications were treated differently by race. When asked if she had observed a white applicant who lied about having graduated from college, Dr. Madden said "I did not look at the question."14 (TR 1277).

Dr. Madden testified that she believed the application no longer asked applicants about their driving records after 2013. She did not include driving records in her shortfall analysis and she said that Dr. White's analysis showed there was no racial difference in the data on driving records. (TR 1278-1279). She reviewed DX 727, the application file for J.B. dated October 26, 2016. She agreed that the telephone screening form asked if the applicant had a valid driver's license and in a parenthetical told the screener to "verify minimum requirements for driving record." (TR 1279; DX 727 at 5). She added that she only used information from the [*72] applications and did not consider information Enterprise employees entered on the screening forms. (TR 1279-1280). Dr. Madden stated that she understood telephone screenings were done when there was a question about information provided in an application and when no one from recruiting had met the applicant in-person. (TR 1280). She did not use telephone screening data because not everyone had a telephone screening; she only relied upon data that applied to everyone and was independently provided. She said some applicants that did not have a telephone screening ended up getting job offers. (TR 1281).

Returning to her rebuttal report, PX 107, Dr. Madden agreed that the pseudo R-squared showed that race explained 1.3 percent of those who were selected to progress beyond the first stage and into the interview stage. (TR 1282-1283). She said that was "highly statistically significant" and suggests that there was more than a thumb on the scales, "[t]here's a brick on the scales." (TR 1283). She agreed that Dr. White found a pseudo R-squared of 11 percent for one row of Table 3 in her rebuttal report. She disagreed that a pseudo R-squared of 11 percent indicated that her model did not account for many other selection factors that would have raised the value. Instead, she said it indicated there was a tremendous amount of discretion and subjectivity taking place in the decision-making process. (TR 1284). Dr. Madden did not have Dr. White's report at hand, but she said it showed that the two things that really mattered was the decision made on sales and customer service experience and then the energy of the applicant. (TR 1285).

Dr. Madden agreed that in her Table 3 regression analysis African-Americans were better distributed with respect to the eight characteristics than white applicants and that when controls were added for education, age, driver's license, ability to work full-time, legal authority to work in the U.S. and the year of the application that the standard deviation decreased from no controls at all. (TR 1286-1287). She explained that the decrease in the standard deviation was due to "less coin flips" when the controls were added and that it was still less than the probability of winning the lottery. She agreed that there was also a decrease in the standard deviation when sales and customer service experience was added, but said it was due to "fewer coin flips" and was still "highly, highly, highly significant, probably as much as anything I've ever seen in my entire career." (TR 1287). Dr. Madden acknowledged that to the extent she controlled for racial differences, as control variables were added the standard deviation decreased for the African-American coefficient, but she disagreed that adding all of the factors Enterprise considered would reduce the standard deviation even further because she believe her analysis looked at all of the things that Enterprise considered at the application review and telephone screening stage consistently over the relevant time period. (TR 1289).

Dr. Madden testified that if OFCCP prepared a regression analysis that included additional factors it would not change her opinion. (TR 1290). She testified that if she was told that OFCCP prepared a regression analysis that looked at whether applicants had college majors that Enterprise preferred she would be surprised because she was never told any particular college majors were preferred. (TR 1290-1291). She said the same about any regression analysis OFCCP may have done that took into account whether an applicant had been a college athlete. (TR 1292). Dr. Madden said there were applicants that did not have a bachelor's degree when they applied that ultimately received job offers. (TR 1292-1293). She agreed that her [*73] analysis of what constituted sales and customer service experience was done by word search for job titles.

Table 2 in PX 107 is an analysis of the hiring process by stage in the progression towards a job offer and measures whether there is a difference by race and the likelihood of passing through the various stages. (TR 1294-1295). Dr. Madden said "disparate impact" is a legal term, not one used by economists, and what she looked at was whether race was a factor at the various stages. (TR 1295). It was Dr. Madden's understanding that there was a subjective evaluation component to the telephone screening. (TR 1296-1297). She agreed that an applicant could be rejected in the initial stage for different independent reasons and that she had not analyzed the telephone screening step as a separate component. (TR 1297). She denied that in her study of the first stage that she assumed the screeners and interviewers knew the applicants races, but she suspected that many times that was the case, although not all the time. (TR 1297-1298). She said she did not review or rely upon any particular studies regarding the impact of African-American names in hiring because she has been doing this work for 45 years and such knowledge is part of her expertise. She agreed that she had not reviewed the list of applicant names to determine whether or not there were African-American sounding names. (TR 1298). She agreed that at the interview stage the interviewer would be aware of the applicant's race by visual observation. (TR 1299-1300). She agreed that the standard deviation she calculated at the second stage (interview) for the 2007 through 2017 period was 6.6 and at the first stage (review and screen) was 17.6, and that the standard deviation was smaller in the second stage. She explained that the smaller standard deviation was due to fewer observations or fewer coin flips. (TR 1300-1301).

Dr. Madden testified that she did not calculate damages in her initial report and that in her rebuttal report she accepted the damages as calculated by Dr. White. (TR 1302). In order to calculate economic losses the shortfall has to be multiplied by the percentage of job offers to African-American applicants that eventually became hires, and that is illustrated in Table 5 of her rebuttal report. (TR 1302-1303; PX 107 at 27). Dr. Madden disagreed that her use of 11 specific applicants in her discussion of the impact of the evaluation of sales and customer service experience was so insignificant as to render her examples worthless given that there were 10,000 total applicants. She said the examples were illustrative of the statistics and the statistics are what really matter. (TR 1304-1305). She said the same thing about discussing five applicants as examples of the evaluation of the college degree requirement. (TR 1305-1306).

Dr. Madden reviewed PX 109, the application file for S.J. dated August 2, 2010. S.J. marked "yes" on her application indicating that she had graduated from college, but added under "current status" that she had three credit hours remaining. (TR 1308; PX 109 at 10). Her resume states that she graduated in May 2010. (PX 109 at 15). The telephone screening form dated August 12, 2010 reports that she said she was one class short of graduating from college. (TR 1309; PX 109 at 17). Dr. Madden said that as a professor she would not recommend a student complete an application the way S.J. completed her application, but she read it to mean that S.J. was in the process of completing her degree requirements. (TR 1309-1310).

Dr. Madden reviewed DX 82, an applicant flow log from 2006-2007, and specifically the entry for A.C., which was on line 305. (TR 1311). There were three reasons listed for rejecting [*74] A.C.: S1 (education), S3 (work experience) and S8 (compatible career interest). (TR 1311-1312; PX 82 at line 305, column L). Dr. Madden agreed that she identified A.C. and two other applicants in her rebuttal report and she believed all three had an S1 disposition code and they may have additional codes. (TR 1315). Going back to S.J., Dr. Madden agreed that S.J. received an interview. (TR 1316). She also agreed that A.O. went through to an interview. (TR 1317).

Dr. Madden disagreed that any conclusion that discrimination existed was outside her expertise as a labor economist and she agreed that whether discrimination existed could also be addressed as a legal opinion. She agreed that she was not offering her professional opinion as a legal opinion. (TR 1318).

Dr. Madden testified that until Enterprise began its cross-examination she was not aware that the ICIMS system automatically knocked out applicants who did not report meeting the minimum requirements established for the management trainee program. (TR 1319-1320). The applications that were rejected by ICIMS remained in Dr. Madden's shortfall analysis and she said they should be removed from the numbers reported in Table 3 of her rebuttal report. (TR 1320-1321). When she took them out, she said the results expanded tremendously and showed a "serious increase in the racial disadvantage." (TR 1321). Fifty percent of the applications knocked out by ICIMS were submitted by African-American applicants. Dr. Madden said she advocated for computer review of applications "because we don't worry about racial bias of robots" and her Table 3 analysis had assumed judgments were being made by people. Those who were rejected by ICIMS should not have been in her Table 3 analysis, so she recalculated her figures after the computer rejections were removed. (TR 1321).

There were nearly 3,000 applications rejected by ICIMS. (TR 1327-1328). Dr. Madden had not seen the email rejection notifications ICIMS automatically generated and sent to applicants telling them they did not meet the minimum requirements, explaining which requirements were not met, and telling them to contact Enterprise if there were errors in the information they had provided in their applications. She agreed that this process may account for 14 applicants that got interviews despite having been rejected initially for failing to meet the minimum requirements. She said she would not have included the 14 had she known that they were screened out by ICIMS, but 14 out of 3,000 is trivial. (TR 1330-1331).

On re-direct examination by OFCCP, Dr. Madden explained that Table 1 of her rebuttal report shows the number of applicants and the number of job offers broken down by race and by year. It assumes that African-American applicants and white applicants are similarly situated. The percentages are relevant to her shortfall analysis because she would expect the percentage of African-Americans offered jobs would equal their share of the total pool of job applicants. (TR 1332). She said removing 25 percent of the African-American applicants and 25 percent of the white applicants from her calculation would have no effect on the overall results. (TR 1333).

Dr. Madden said the shortfall shown in Table 1 was very dramatic and certainly was not due to chance, which meant there were only two explanations for why it occurred: (1) Enterprise treated African-American applicants and white applicants differently or (2) as a whole, African-American applicants were inferior to white applicants. (TR 1334-1335). Dr. Madden said "there [*75] is no other possible explanation." (TR 1335). Dr. Madden stated that she was not advocating for a quota system, but the ultimate results should be within the range of probability. A company can establish hiring criteria based on qualifications relevant to the productivity and profitability of the company, but when the criteria are subjective it requires being especially careful to make sure the criteria are applied fairly. (TR 1336-1337).

Dr. Madden reviewed footnote 8 in her initial report, PX 105. The note said that she included all of the requirements Enterprise used to conduct application reviews from August 1, 2006 through July 31, 2017, but she did not use information on DUIs, drug and alcohol offenses, felony convictions and loss of driver's licenses in her regression analysis because Enterprise stopped asking for that information on applications beginning in the 2012-2013 period. (TR 1337). She assumed Enterprise stopped asking for the information because it was inaccurate or had proven not to be important, so she did not include it in her regression analysis. As a labor economist, Dr. Madden said having a full regression analysis that includes the same information on everybody is the better practice. (TR 1337-1338).

Dr. Madden said that Enterprise performed background checks on applicants at different stages in the hiring process during the charged period. They did not do background checks at the application review and telephone screening stage. She did not include background check information in her Table 3 analysis because it assessed applicants successfully getting through the application review and telephone screening stage and into the interview stage. (TR 1340). She agreed that there was nothing wrong with Enterprise doing background checks as part of the hiring process, but it was not done as part of the initial stage that she was examining. (TR 1340-1341).

Dr. Madden said the education and work experience requirements changed over time. The August 1, 2015 criteria required a bachelor's degree with a business or communications degree preferred and one year of experience in two or more of: sales, customer service, military leadership or professional or college athletics. (TR 1341). The August 1, 2014 criteria was generally the same, but only mandated one year of sales and customer service experience. The same was true in 2012 and 2013. In 2010, the work experience requirement was one year in sales and customer service within the previous three years or within five years for business and communications majors. (TR 1342). The same criteria applied in 2009. Dr. Madden explained that in her regression analysis it was important to include criteria that were consistent over time because consistent application indicated the company considered the criteria to be important and, from a statistical analysis perspective, you want to index criteria. (TR 1343). Looking at PX 75, Dr. Madden agreed that Enterprise added athletic and military experience to the criteria in 2015. She said she did not include that in her analysis because she did not see either one reported as sales or customer service work experience on any of the applications. (TR 1344).

Dr. Madden said that she did not do a regression analysis for the interview stage and the background check stage of the hiring process because she did not have objective information to evaluate. In the review and screening stage, she had application information as reported by the applicants that was not affected by the decision-making of anyone at Enterprise. In the latter stages, Enterprise applied subjective judgment and there was no way to test it for discrimination. As an example, she said an Enterprise interviewer could say white applicants have good [*76] communication skills and African-American applicants do not, and code them accordingly, and there would be no way statistically to evaluate those judgments. She said the majority of applicants were rejected in the first stage and the numbers were much smaller in the latter two stage — particularly the background check stage — and the outcomes were more subjective. She assumed that having successfully cleared the first stage that African-American applicants and white applicants selected to move on as a whole had comparable credentials. (TR 1345-1346).

Dr. Madden reviewed Dr. White's chart entitled "Effect of Sequential Removal by Applicant Flow Code on Offer Shortfall Calculations." (TR 1346; PX 106 at 20). His first row reflects no applicants removed from the process, so it includes the entire database that both he and Dr. Madden used. The second row reflects self-select outs and Dr. Madden explained that the shortfall increased because more white applicants than African-American applicants self-selected out of the process. (TR 1347). Dr. White's analysis showed no racial difference in the application of the criteria for job stability (S2) and motor vehicle record (S11). (TR 1347-1348).

Dr. Madden said the individual applicants she cited in her report were illustrative of what was reported on the applications versus how Enterprise coded what was on the applications. (TR 1348). She said she agreed with Enterprise that looking at individual applicants is not probative evidence of anything and she said they were included simply as examples. She said the evidence was in the statistics as reflected in her Table 3. (TR 1349).

Dr. Madden said she and Dr. White used the same database, the difference is in what they did with it. (TR 1349). She agreed with Enterprise that both she and Dr. White found some substantial racial disparities. (TR 1349-1350). Dr. Madden calculated a shortfall of 220 and Dr. White calculated a shortfall of 181.9 before he applied the disposition codes. Both of their analyses showed that stage one — application review and telephone screening — had the greatest racial impact. The big drop in the shortfall tally in Dr. White's chart started when the disposition codes moved from stage one (S codes) to stage two (I codes) and eventually ends with zero shortfalls. (TR 1350). Within the interview criteria, the one that had the greatest racial affect was work ethic (I6) where Enterprise's interviewers were much more likely to say African-American applicants lacked work ethic than white applicants. (TR 1351).

Dr. Madden said pseudo R-squared looks at individual variations rather than group variations. She said she presumed there was pseudo R-squared data in her computer backup material, but she was interested in looking at group differences and R-squared data does not tell her anything about groups. She said she looked at the pseudo R-squared value and it was reasonable. If the pseudo R-squared is a negative number it indicates there is a problem with the study. In this case, it was a positive number, the sample analyzed contained a large number of people, and the results were relatively rare. Dr. Madden said by academic journal standards, the pseudo R-squared for her regression analysis is "really good" and most of the academic paper she published had comparable R-squared values. (TR 1353-1354).

On re-cross examination by Enterprise, Dr. Madden agreed that in the most recent criteria from 2015 it states that applicants will not be asked about their criminal records and no criminal background check will be conducted until a conditional employment offer has been tendered. (TR 1354-1355; DX 75 at 3). Dr. Madden disagreed that Enterprise made many changes to its [*77] requirements over time. She said the details changed, but the broad set of characteristics were consistent. (TR 1357). She agreed that the regression analysis in Table 3 for the 2007 to 2017 period was based on the eight factors. (TR 1358). She did not do an analysis by year because of the "coin flip problems" and she said the smaller group size if done by year would produce less powerful statistical analysis. (TR 1358-1359). She said she did the multiple pools analysis with a control for the year. The multiple pools analysis is the bottom two rows of Table 1. (TR 1359; PX 107 at 19). Individuals who were rejected for failure of any requirement remained in the pool for purposes of Dr. Madden's shortfall disparity analysis. (TR 1360).

Dr. Madden testified that her analysis clearly showed that race was a factor, so the question was why was it a factor? Either Enterprise treated applicants differently by race or African-Americans applicants were systemically inferior in their qualifications. She agreed that the only information she considered was the information the applicants themselves provided in their applications and that she did not consider information reported by Enterprise's screeners or interviewers because it was filtered through the judgments of the group whose racial bias was being evaluated. (TR 1361).

DR. PAUL F. WHITE

Enterprise called Dr. Paul F. White as an expert witness.15 (TR 1362). Dr. White has been a labor economist for about 25 years. Most of his work involves the statistical analysis of alleged discrimination. It also involves analysis of data on compensation, FLSA wage and hour matters, and economic loss calculations. (TR 1363). Dr. White earned a bachelor's degree in economics with a specialty in labor economics from James Madison University, a master's degree in labor economics with a focus on healthcare economics and statistics from North Carolina State University, and a Ph.D. in labor economics, healthcare economics and statistics from North Carolina State University. He did a fellowship while he was in graduate school on the economics of aging that was funded by the National Institutes of Health. (TR 1364). He has been employed for almost three years at Resolution Economics where he is a partner and the head of the Washington, D.C., office. Dr. White says he spends about half of his time doing statistical analysis of compensation to determine whether there are pay gaps by gender or race and to help companies correct them where gaps are found. This is proactive work and does not involve litigation. The other half of his time is devoted to litigation work in cases like this one. His litigation work is about half discrimination and half FLSA wage and hour matters. (TR 1365-1366). He previously worked for 22 years at ERS Group where he was a managing director. His duties were similar to his current duties, although less time was devoted to proactive compensation issues because it was not a hot topic at the time. (TR 1366).

Dr. White was an adjunct professor at Florida State University for six years where he taught two graduate courses, including one on general research methods and one on applied economics. (TR 1366-1367). He was also a faculty member at the Institute for Workplace Equality where he did training on compensation and hiring issues. He has provided training on [*78] statistical analysis of employment issues addressing the types of employment issues government contractors would face under the jurisdiction of OFCCP. (TR 1367). He has done training at seminars and conferences for attorneys, human resources employees and others, including OFCCP personnel. (TR 1367-1368). Dr. White has been a professional journal referee for the Journal of Forensic Economics and two other professional journals. That involves reviewing manuscripts and providing opinions on whether they warrant publication. (TR 1368). He is on the board of directors of the Journal of Business Valuation and Economic Loss Analysis at the National Association of Certified Valuation Analysts. (TR 1369).

Dr. White has been retained to testify as an expert witness over one hundred times and he has testified about ten times. (TR 1370). About 50 of those cases involved allegations of discrimination. (TR 1372). He testified as an expert witness in another OFCCP case where Dr. Madden was the OFCCP expert witness. Of the approximately ten cases were Dr. White has testified as an expert witness, he believed at least five involved allegations of discrimination. (TR 1371). He has been retain 50 to 100 times as a consulting expert. There have been times when he has testified in discrimination cases where he found the data were consistent with the allegations. (TR 1372). Dr. White has published numerous articles for economic and legal publications. (TR 1373-1374; PX 106 at 42-43).

On voir dire by OFCCP, Dr. White acknowledged that he is not an academic and that his academic experience was limited to the work he had done as an adjunct professor. (TR 1375). He said that he did not believe his expert witness testimony in prior cases had ever been limited by the Daubert standard. He agreed that he had advised clients in other cases that his analysis showed the data were consistent with racial discrimination and that whether the data are consistent with racial discrimination in this case depends on which stage of the hiring process is determined to be the appropriate stage to analyze. (TR 1376).

Dr. White was accepted as an expert witness in labor economics and statistics. (TR 1377).

Dr. White used a PowerPoint presentation as a demonstrative aid for his testimony. (TR 1377; DX 7362). He was retained to provide services to Enterprise as a labor economist in early 2017. He was asked to conduct a statistical analysis of job offers for employment as a management trainee and to review and respond to any work done by OFCCP's expert, Dr. Madden. (TR 1378-1379). Dr. White reviewed application files, applicant flow logs, information on race, and payroll data used to calculate economic loss. He said there were a little less than 19,000 application files and about 10,000 were included in the database. (TR 1379). Some of the application files included notes. (TR 1379-1380). He believed the named applicants in the list of applications that were reviewed were also named in the applicant flow logs. The information in the applicant flow logs was the basis for the analysis done by both Dr. White and Dr. Madden. (TR 1380).

Dr. White reviewed DX 32, the screening criteria as of August 1, 2009; DX 189, the screening criteria as of August 1, 2015; and DX 190, the screening criteria as of August 1, 2016. (TR 1381-1382). He said that the handwritten notes Enterprise prepared during the screening, interviewing and selection process were reviewed and compiled in a database by his team. Dr. [*79] White also reviewed deposition transcripts for background and to understand the decision-making process. He believed that he reviewed Ms. Hardesty's deposition transcript. (TR 1383).

The statistical analysis Dr. White conducted was a pools analysis, which he explained means he looked at the African-American representation in the pool and compared it to African-American representation among those selected. Dr. White agreed with Dr. Madden's description of regression analysis. (TR 1384). He said a pools analysis looks at the pool, however the pool is defined, and assumes everyone is equally likely to be selected, and then compares the initial pool to the pool of those selected; it compares the percentages for the two groups. Regression analysis looks at the effect of individual characteristics. As an example, in an analysis of compensation, you can take base salary and then use regression to look at the effects of years of seniority, education level and focus the spotlight on the race affect. (TR 1384-1385). He said a "dependent variable" is the outcome you are looking to measure; in this case, for instance, whether or not someone was selected for the management trainee program. The "independent variable" is the list of factors you use to explain why people were or were not selected. (TR 1385). Dr. White said the "law of large numbers" means that when you have a lot of observations in your analysis the difference between the actual and expected outcomes do not have to be as large to be statistically significant. Using Dr. Madden's coin tossing example, he said if you toss a coin a million times you would expect it come out pretty evenly at 500,000 heads and 500,000 tails, and even a small difference could be statistically significant. Toss a coin four times and it may or may not results in two heads and two tails, and if it ends up three and one it would not be surprising. (TR 1386).

Dr. White said that his general concern with Dr. Madden's analysis is that the goal was to create a statistical model that reflects the decision-making process used to choose management trainees and the model Dr. Madden built does not reflect that process. He agreed with Dr. Madden that most people got kicked out of the process in the application review and telephone screening stage, but he said race is unknown during that stage. (TR 1388).

Dr. White did his own shortfall analysis that took into account all of the reasons Enterprise said it considered and he found that there was no shortfall of job offers to African-American applicants. He said Dr. Madden did not consider any of the requirements and qualifications Enterprise reported in conducting her shortfall analysis. (TR 1419).

Dr. White said he and Dr. Madden agreed on the database and they both used the same initial database. He said her pools analysis was at the beginning of the process and looked at the entire pool and removed the self-selected outs. Dr. White said he began with the entire pool and removed the self-selected outs, and then each line below in his table shows what happens when you remove applicants for the reasons Enterprise gave for rejecting the applicants. (TR 1420-1421; DX 7362 at 21).16 He said that enables the fact-finder to understand what the shortfall is at any stage depending upon where the fact-finder determines the appropriate stage is for evaluation. (TR 1421).

[*80] The first line of his table shows the roughly 10,000 applicants with no removals.17 The next column shows the number of fiscal years that were significant at two standard deviations. He said he ran all 11 years and if no one is removed from the pool all 11 years are statistically significant — adverse to African-American applicants — at two standard deviations and he calculated a shortfall of 209 African-Americans. If the bar is set at three standard deviations, then nine fiscal years are statistically significant and the African-American shortfall is 160.2. (TR 1421). Removing the self-select outs results in 11 years that are statistically significant at two standard deviations and an African-American shortfall of 223 and at three standard deviations there are nine fiscal years that are statistically significant and a shortfall of 201.9 African-Americans. (TR 1422).

The next line shows the entire pool with self-select outs and initial DNQs (does not qualify) and at two standard deviations there are ten years that are statistically significant and a [*81] shortfall of 185.1 African-Americans and at three standard deviations there are seven fiscal years that are statistically significant and a shortfall of 150.9. (TR 1422).

Dr. White said he ran numbers for both two and three standard deviations because of the Hazelwood case.18 He said that he used the disposition codes the interviewers entered as the reasons Enterprise rejected applicants. (TR 1423). Dr. White was aware that ICIMS rejected applicants for not meeting the minimum requirements. He said he became aware of that during a telephone conversation with Mr. Arney, Ms. Morris and Ms. Hardesty on January 12, 2018. (TR 1423-1424).

Dr. White said his team gathered and reviewed all of the notes and entered the notes information into a database. (TR 1425). Team members who compiled notes data did not enter all of the information, just the information they believed was relevant. They were instructed to enter the data just as it appeared on the forms. (TR 1425-1426). The information on why applicants did not move on in the process was provided by Enterprise. (TR 1426). He said "[w]e relied on the code that was provided by ERAC as to why they were excluded." (TR 1427). The shortfall drop from I5 (flexibility) and I6 (results driven-work ethic) at two standard deviations went from 40.3 to 35.9. (TR 1428).

Dr. White said he determined whether an applicant was rejected for poor communication ability by looking at the disposition code Enterprise provided. He agreed that he had the notes information that would have supported the reason for the rejection. Dr. White clarified that his table was cumulative and that every line excludes those rejected for the reason given for that line as well as all of those already excluded on the lines above, so it slowly converges to zero on the last line. (TR 1429). He said the numbers could be recalculated if the factors were arranged in a different sequence and that his chart was intended as an illustration of the evolution of the process. (TR 1429-1430). When asked if his analysis showed any evidence of racial bias in the application of independent variables by Enterprise's recruiters, he responded, "I've not seen evidence one way or the other on that." (TR 1430).

For clarification, I asked Dr. White whether I understood correctly that each line in his chart subtracted all of the applicants eliminated in the prior lines plus those eliminated for the reason shown on the current line and he said that was correct. I asked why some of the shortfall numbers increased moving down the chart when intuitively it seemed they should get progressively smaller. Dr. White explained that it depended upon the representation of African-Americans with the characteristic being excluded, which could result in an increase in the shortfall. I asked if that indicated more African-Americans were being excluded by the particular code and he responded, "[t]hat could very well be the case." (TR 1430-1431).

Counsel for Enterprise asked Dr. White to look at two particularly lines in his chart to illustrate how the shortfall could increase as the progression moved down the chart. The line for disposition code S6 (flexibility) showed a shortfall in column 3 (two standard deviations) of 102.8 while the shortfall in the line beneath it for code S7 (geography) was higher at 103.1. Dr. White said that in a pools model you compare African-American representation to those who were selected versus those that would be expected to be selected based upon their representation [*82] in the pool. (TR 1431). Dr. White said that the goal is to build a model that reflects as closely as possible the decision-making process and try to assure that any differences are due to race and not to modeling differences. (TR 1432). He said that pseudo R-squared is a percentage that is thought of as a predictive power of an expression; it tells you what percentage of the story is told by the factor that you are considering. He said his table reflected a pools model that calculated standard deviations, so it did not call for pseudo R-squared measures that might apply to a regression analysis. (TR 1433). The inclusion of all known variables in a regression increases the explanatory power; it gives you more pieces of the puzzle to look at. Dr. White said he and Dr. Madden calculated damages using the same methodology and that if there is a shortfall of zero there are no damages. (TR 1434).

On cross-examination by OFCCP, Dr. White agreed that he earns a substantial portion of his income from providing expert witness services. (TR 1434-1435). He said that he did not have an expert opinion on whether Enterprise employees involved in the application review and telephone screening stage were aware of applicants' races. (TR 1436). He agreed that Enterprise did not give him any law or factual assumptions to apply in reaching his expert opinions. He agreed that he relied upon the applicant flow logs and that the logs showed the races of the applicants, the disposition codes Enterprise entered, and the stage of the hiring process when each applicant was rejected or hired. (TR 1438).

Dr. White agreed that in his analysis he tested the outcome by race for each of the disposition codes and he did not do any systematic checks to see if the codes were applied in a reasonable manner. He took the codes Enterprise entered as a given and he said that to check whether the codes had been applied consistently would require reviewing all of the applications and making judgment calls on whether the codes were applied consistently between African-American applicants and white applicants. (TR 1439). He said his team prepared a database in digital format from their review of all the notes and he did not know whether that information was provided to OFCCP in discovery. He provided the notes database to Enterprise's counsel. (TR 1440-1441). Dr. White said the notes were reviewed for background, but they did not explicitly factor into the statistical study. (TR 1441). He said he had personally reviewing 30 to 50 applications to spot check some of the information his team transcribed into the notes database. (TR 1441-1442).

Dr. White said it was his understanding that Enterprise accepted applications for the management training program on a rolling basis, although he saw some applications that were rejected with a "position filled" disposition code during the period around the Great Recession of 2008. (TR 1442-1443). He became aware that some applications were rejected automatically by the computer system for not meeting minimum requirements during a conference call with Mr. Arney, Ms. Morris and Ms. Hardesty on January 12, 2018. (TR 1444-1445). He recalled specifically during that conversation being informed that the ICIMS computer program screened applications to see if applicants met the minimum requirements. (TR 1445). He did not recall who was speaking at the time. He knew he reviewed Ms. Hardesty's deposition testimony prior to writing his report because he cited it in the report. (TR 1444-1445). Dr. White was deposed on May 4, 2018. (TR 1446; PX 58). In his deposition, Dr. White said he had reviewed Ms. Hardesty's deposition transcript in preparing for his deposition and that he had reviewed her deposition to see whether recruiters knew applicants' races during the stage one process. (TR [*83] 1447-1448). He was aware that some applicants applied in-person at open houses or during campus visits, but he did know what percentage of applicants that constituted and it did not affect his analysis. (TR 1450).

With respect to the table he prepared, Dr. White said it examined the shortfall of African-American offers after the disposition codes were applied. (TR 1452). He agreed that his analysis for the S1 (education) code showed a statistically significant racial disparity at two standard deviations in 10 of the 11 fiscal years and at three standard deviations in 8 of the 11 fiscal years he examined. (TR 1453). For the S2 (job stability) codes, he found a statistically significant shortfall for African-American applicants at two standard deviations in 10 of 11 fiscal years and at three standard deviations in 8 of 11 fiscal years. (TR 1453-1454).

Dr. White was asked why the number of statistically significant years decreased moving down his table because the pool of African-Americans shrinks along the way. (TR 1454-1455). He agreed that as the pool gets smaller it takes a larger difference to be statistically significant and it is consistent with the declining standard deviations shown in his table. He said the purpose of his study was not to determine whether Enterprise discriminated against African-American applicants for the management trainee program. (TR 1455). Dr. White tried to present the numbers from the different stages in the process showing how the shortfalls evolved, but he said he leaves it up to others to decide whether the numbers are consistent with discrimination. He agreed that his analysis found statistically significant racial disparities at many points in the process. (TR 1456). He said when he applied the sales and customer service experience disposition code to the lines excluded above it there is statistical significance in eight of 11 years at two standard deviations and the shortfall drops from 174.3 to 105.2 and at three standard deviations there are four years that show statistical significance and the shortfall drops to 65.4. (TR 1456-1457). Dr. White also found a significant drop in the shortfall between the I6 (results driven/work ethic) and I7 (compatible career direction) disposition codes. At two standard deviations the shortfall fell from 35.9 to 7.3 and at three standard deviations it fell from 21.0 to zero. (TR 1457).

Dr. White reiterated that pseudo R-squared is a percentage that is the explanatory power of the regression shown in Dr. Madden's Table 3. He said he did not run his own regression and to the extent he believed there were other criteria that should have been controlled for in the study, he elected not to run any alternative regressions. (TR 1460). He believed that the analysis depicted in the table with sequential removals by disposition codes adequately addressed the issue. (TR 1460-1461; PX 106 at 20).

Dr. White clarified that Dr. Madden's regression was not on job offers, it was a regression on applicants successfully passing through the first stage of the hiring process. Dr. Madden's Table 1 and Dr. White's table address job offers. (TR 1462). He agreed that in his analysis of the disposition codes and the telephone screenings, branch observations, and in-person interviews he did not check to see if the codes were applied in a race-neutral manner. (TR 1463-1464). In response to my request for clarification, Dr. White confirmed that a bigger drop-off in the shortfall numbers meant a bigger impact on African-Americans applicants. (TR 1465).

[*84] Dr. White reviewed his deposition transcript, PX 58, and specifically his deposition testimony about the telephone conference call he had with Mr. Arney, Ms. Morris and Ms. Hardesty. He said Enterprise had not instruct him to withhold information at the deposition about applications being screened by computer to see whether applicants met the minimum requirements. (TR 1466-1467; PX 58 at 112-114).

On re-direct examination by Enterprise, Dr. White said it would not be appropriate for he or Dr. Madden to substitute their judgments for the evaluations made by Enterprise since the company is more knowledgeable about the process. (TR 1468). Dr. White said he did not receive a written request from OFCCP for the notes database he prepared. (TR 1471). He agreed that when he applied the reasons Enterprise gave for rejecting applicants the shortfall was reduced to zero. (TR 1473).

On re-cross examination by OFCCP, Dr. White said "I don't have strong objections to" Dr. Madden's statement that the only two explanations for the hiring results is that African-Americans as a whole are less qualified than white applicants or Enterprise applied its criteria in a racially discriminatory manner. (TR 1476-1477). Dr. White said he would not testify whether or not the reasons Enterprise gave for rejecting applicants were valid reasons. (TR 1477).

DR. JANICE FANNING MADDEN

(REBUTTAL TESTIMONY)

OFCCP recalled Dr. Janice Fanning Madden as an expert witness. Dr. Madden said that she revised the numbers in Tables 1, 2, 3 and 5 and prepared new tables, which were admitted into evidence without objection as PX 114. (TR 1491-1493). She testified that during the course of the hearing she learned that some of the decisions on applications were made by a computer program, which is the neutral standard she advocated for in assessing applicant qualifications. She revised her initial testimony to remove the applications that were screened out by computer for failing to meet the minimum requirements rather than being rejected by human beings. The first page of PX 114 is Table 1A, which is the same as Table 1 except for the removal of applicants screened out by computer.19 (TR 1494-1495). The number of shortfalls decreased [*85] from 220 to 196 and the standard deviation decreased from 13.4 to 12.7. (TR 1495). Dr. Madden said that in her earlier testimony she said the probability of the racial disparity being by chance was a decimal with 42 zeros in front of the numeral and now it would be the same but with 36 zeros in front of the numeral. (TR 1495-1496).

Dr. Madden had no objection to calculating shortfalls by year as Dr. White had done, but she said the overall pattern — considering that some years were positive and some years were negative — was clearly negative. (TR 1496-1497). Dr. Madden said the shortfall for 2012 was 25 and the standard deviation was 5.2. (TR 1497-1498).

Comparing Table 2 with Table 2A, Dr. Madden said there was no impact on 2007 and 2008 because there was no computer program screening applications during those years.20 Computer screening began to have an impact in 2009 where taking those rejected by computer out of the equation reduced the number of white applicants from 417 to 374. (TR 1499). The number of African-American applicants remaining in the pool in 2009 decreased from 351 to 289. Dr. Madden did the same reductions for all of the years through 2017 and it decreased the shortfall by five people, which Dr. Madden said is still statistically significant. (TR 1500). The standard deviation in Table 2A was 16.4 compared to 17.6 in the original Table 2. Dr. Madden said the calculations for stage two and stage three in the original Table 2 should not change as a result of subtracting computer screenings in stage one since her original calculations included 14 applicants that had advanced after correcting their original submissions.

Dr. Madden recalculated Table 3A using the same logit regression she used for the original Table 3. (TR 1501). In Table 3 she looked at what happens if you control for credentials versus not controlling for credentials. She said that with no controls for credentials, the probability of failing is lower overall because the unqualified applicants were removed from the pool. When controls for credentials are added, the probability of African-American applicants failing increases, which indicates that decision-making by human beings had a much bigger racial effect. The results indicate that African-American applicants were "substantially better situated than the whites, were more qualified, overall." (TR 1502). When controls for sales and customer service experience were added the results for African-American applicants were a little better, but the failure rate was still well above the expected rate. (TR 1503). Dr. Madden said she did not make any judgments about the criteria Enterprise chose; she only [*86] measured how well the data showed African-American applicants and white applicants met those criteria. She did not control for communication skills because there was no objective information to assess, but the huge racial difference calls it into question. (TR 1505-1506).

With respect to damages, Dr. Madden said removal of the applicants the computer determined failed to meet the minimum requirements reduced the damages from $7,693,184.00 to $6,989,439.00. That figure includes interest from December 31, 2017 to June 27, 2018, but it does not include lost earnings from August 2017 to the hearing in June 2018 because Dr. Madden did not have data for that period. (TR 1506-1507). The figure was calculated using a hiring shortfall of 171.8, which was based on a decrease in the offer shortfall from 220 in her original rebuttal report to 196 when the computer rejected applicants were removed. (TR 1506-1507).

On cross-examination by Enterprise, Dr. Madden said she was present at Dr. White's deposition and she heard him explain that he removed the applicants who did not meet the minimum requirements, but he did not say the applications were screened by computer. (TR 1510). She agreed that her Table 3A is not an analysis of job offers. She said that she could mathematically assess offers if she was given more time, but she had not done so. She believed it would show that African-Americans that received job offers were 50 percent more qualified on average than white applicants who received job offers. (TR 1511-1512).

Dr. Madden agreed that the reduction in the number of African-American applicants and white applicants in Table 1 in her rebuttal report and the current Table 1A was due solely to the removal of individuals the ICIMS computer system screened out for not meeting the minimum requirements. She said the figure included the 14 applicants she referenced earlier who proceeded on in the process despite initially failing to pass all of the minimum requirements. She agreed that her shortfall figure fell from 220 to 196 and that reflected about a ten percent decrease. (TR 1513). She agreed that the standard deviation improved from 13.6 to 12.6, although she noted that it went "from absolute certainty to absolute certainty." (TR 1513-1514).

Dr. Madden agreed that African-American applicants were about 59 percent of the individuals the ICIMS computer system knocked out, but she noted that they represented 55 percent of the total applicant pool, so the difference was not statistically significant. (TR 1514). She agreed that she had not conducted studies to determine if Enterprise was racially biased in its evaluations of whether an applicant had a DUI conviction, a felony conviction, held two or more non-temporary jobs in the prior year, or had relevant work experience in sales and customer service. (TR 1515-1517). She did not agree that Enterprise's employees did not know the race of applicants during the application review and telephone screening stage. (TR 1518). She agreed that her team did a word search for job titles to assess the sales and customer service experience criteria. (TR 1519). She agreed that she only did a regression analysis for the first stage of the hiring process. Removing the applicants who had been screened out by computer was the only revision made in order to prepare Table 3A. (TR 1520).

Dr. Madden agreed that the only thing Table 3A analyzed was the probability of failing at the application review and telephone screening stage and not moving on to the interview stage. (TR 1521-1522). Her regression analysis did not consider the reasons applicants were rejected in [*87] the second or third stages because appropriate data were not available. (TR 1522). She agreed that Enterprise had evaluated job experience information applicants provided in their applications as well as information provided in their resumes and during the telephone screening. (TR 1523).

Dr. Madden reviewed DX 176, an email sent by OFCCP employee Robert LaJeunesse to several other Department of Labor employees on December 19, 2014 with subject line "R-squared for Enterprise Regression." (TR 1524; DX 176). She agreed that the first row in the chart in the email had an R-squared of roughly 17 percent and that the regression analysis was less than 17 percent. She said the pseudo R-squared for her Table 3A went down, meaning there was "more discretion going on among human beings than among the computer." (TR 1524). The pseudo R-squared for the first row of Dr. Madden's table 3A (race only) was one percent.

Dr. Madden reviewed DX 7363, which is the computer analysis of the data her team ran to compile Table 3A. She had sent the data to counsel for OFCCP, but she did not review it after it was provided to her by her team. The document shows the steps the computer ran to do the regression analysis. (TR 1526).

[Dr. Madden had not reviewed DX 7363 prior to being questioned on it at the hearing and she was not prepared to answer detailed questions about it. An error in the computations was detected. The parties agreed that Dr. Madden and her team would re-run the data and produce a new table 3A. (TR 1545-1547)].

OFCCP introduced PX 115, notepad data for a newly revised Table 3A, and PX 116, the revised Table 3A. Both were admitted without objection. (TR 1548-1549). On re-direct examination by OFCCP, Dr. Madden said she conferred with her team and there had been an error in entering a signature variable with a proper "I" code that caused some applicants to be improperly removed from the regression. (TR 1549). Dr. Madden said PX 115 is the output from a software package called Stata that is used to run statistical analysis. The top of page one shows a table with four rows and four columns. The row that begins with "0" and shows a frequency of 4,444 is a count of applicants who made it through stage one. (TR 1550-1551). The row that begins with "1" and shows a frequency of 7,040 is a count of applicants that did not make it to the interview stage. The row that begins with "8888" and shows a frequency of 2,919 is a count of applicants who self-selected out of the hiring process. The row that begins with "9999" and shows a frequency of 935 is a count of applicants where there was no record they were hired, interviewed or self-selected out. The rows total 15,535. The chart at the bottom of page one shows by fiscal year from 2007 through 2017 the numbers that fall into each of the four categories by year. (TR 1551; PX 115 at 1).

On page 2 of the Stata output, there were 2,929 people removed because they were screened out by computer rather than by human beings. (TR 1552). The 2,919 applicants who self-selected out and the 28 applications that were still pending hiring decisions were also removed leaving a total of 9,462 in the study. (TR 1553). Dr. Madden said that the pseudo R-squared value of 0.1497 means that about 15 percent of the overall variation in selection of individual applicants is explained by the characteristics the study controlled for: race, college degree, over age 18, valid driver's license, authorized to work in the U.S., full-time employment, [*88] work visa and then application of the appropriate fiscal year. Adding the sales and customer service experience variable results in a pseudo R-squared value of 0.1656 or about 17 percent. (TR 1555).

Dr. Madden said that PX 116 is the same as Table 3A, but with the removal of the applications that were screened out by computer.21 It shows that African-American applicants were significantly more likely to fail to progress past stage one than white applicants. The standard deviation was 11.84 and demonstrates that "race is a huge predictor of this outcome." (TR 1556). The next column compares what happens when the characteristics Enterprise identified as important were added in, except for sales and customer service experience. (TR 1556-1557). The effect of race increased to 0.634 and the standard deviation rose to 13.52. (TR 1557; PX 116). Dr. Madden said that the revised data did not cause her to change the opinions she expressed earlier and if anything the result was stronger than she had calculated before. (TR 1558).

On re-cross examination by Enterprise, Dr. Madden said the only change in Table 3A and Table 3A-revised is the removal of a variable that was not explaining anything in the analysis and that caused the Stata program to kick applications out. She said the "9999" line in PX 115 reflected applications where Enterprise had not assigned a code showing where they were rejected, there was no interview or background check information, and there was no other basis to determine where in the hiring process the applicants were kicked out. She agreed that she assumed they were rejected in stage one and they remained in as rejections in the regression analysis. (TR 1559-1560). She said self-select outs were in line "8888" and were not included in the 935 applications shown on line "9999." (TR 1561). Dr. Madden agreed that if any of the 935 applicants had self-selected out that would be an error in her data. (TR 1567). She said in the data that she ran, line "8888" was 2,191 self-select outs and line "9999" was the 935 applications that were rejected but the application flow log did not show where in the process they were rejected. She agreed that she relied solely on the application flow log information and that the regression analysis in Table 3A-revised only looked at stage one. (TR 1570-1571). She assumed the 935 applications on line "8888" had been rejected in stage one because there was no information on the application flow log reflecting that they got interviews or background checks. (TR 1574-1575).

[*89] Dr. Madden said that the selection process Enterprise used from 2007 through 2017 was essentially the same, although where in the process they checked for certain credentials moved in some instances. Table 1A looked at the entire selection process and not individual stages. (TR 1581). She said Table 3A-revised looked at how application characteristics predict whether an applicant made it through to an interview. (TR 1582-1583). Table 2A looked at whether an application moved to the interview stage and on to a background check. (TR 1585). She agreed that Table 5A adopts Dr. White's methodology and calculation with respect to back pay and mitigation for alleged shortfalls. (TR 1586). Dr. Madden said that if any of the 935 applications in line "9999" were self-select outs in the first stage they should have been in line "8888" instead. She was not able to determine from the data in the application flow logs where they were rejected and, since they were predominately white, she left them in the analysis. Had they been removed, since they were predominately white applicants, it would have made an even stronger case for racial disparity. (TR 1587-1588).

On redirect examination by OFCCP, Dr. Madden said that as an economist and statistician she was comfortable with the assumption she made to leave the applications on line "9999" in her analysis. (TR 1589-1590). Table 1A assumes that African-American applicants were equally qualified as white applicants and Table 3A tests that assumption. Dr. Madden said that when she controlled for what African-American applicants and white applicants had on their applications and indexed the criterion that were used consistently over the entire period, there is a higher failure rate for African-American applicants. (TR 1590).

In Table 3A-revised, Dr. Madden removed 14 applicants who had been determined not to meet the minimum requirements but later revised their applications and proceeded on in the process. She said as a statistician whether the 14 applications stay in or are taken out does not change the overall result. (TR 1591). African-Americans accounted for about 55.6 of the total applicant pool and about 58.6 percent of the applicants determined not to meet the minimum requirements, which is a slight marginal increase and is not statistically significant. (TR 1592).

I asked Dr. Madden to clarify whether the 900-plus applicants alleged to have self-selected out (shown as part of line "9999" in the PX 114 table) were shown as self-select outs on the application flow logs. She said they were, but there was no information on where in the process they self-selected out. (TR 1596-1597).

On re-cross examination by Enterprise, Dr. Madden agreed that if any of the "9999" self-select outs that I asked about had selected out in stage one they should have been excluded from the regression analysis, but if they selected out later in the process they should have been included. She said leaving the "9999" group in the pool counted a disproportionate share of white applicants as rejections and had the ones who self-selected out in stage one been removed it would have produced even stronger results showing a racial difference. (TR 1597-1598).

On re-direct examination by OFCCP, Dr. Madden said the roughly 900 self-select outs that were in line "9999" were identified as "self-selected out before an offer" in Dr. White's data. (TR 1598).

[*90] DR. PAUL F. WHITE

(REBUTTAL TESTIMONY)

Dr. Paul F. White was recalled as an expert witness by Enterprise. Dr. White said that his conclusions with respect to Dr. Maddens Table 1A did not change from his previous testimony and that the only difference is a decrease in the shortfall. (TR 1600). One of his previous criticisms was that Dr. Madden included applications that were determined not to meet the minimum requirements by the ICIMS computer system and her revised analysis resolved that issue. (TR 1601). Dr. White had concerns about the methodology Dr. Madden used to identify sales and customer service experience and the education requirement. With respect to education, some of the applications that indicated applicants did not meet the education requirement included additional documentation showing that the applicants met the requirement by the time Enterprise hired them. (TR 1602).

Dr. White said Dr. Madden failed to account for the reasons Enterprise rejected applicants. (TR 1603). In the analysis he conducted, Dr. White showed that accounting for the reasons applicants were rejected results in a shortfall of zero. Dr. Madden's revised tables did not change Dr. White's conclusion in that respect. (TR 1604).

Dr. Madden ignored all of the information about the individuals contained in their applications, resumes, telephone screens and interview evaluations. (TR 1605). Her Table 1A shortfall analysis looks at the end result of the selection process. Dr. White said that he analyzed all of the reasons Enterprise gave for rejecting applicants and Dr. Madden did not. (TR 1607). He said the goal was to create a statistical model that reflects the decision-making process that distinguishes between individual applicants and ignoring all of the individual characteristics is not consistent with the decision-making process model. (TR 1607-1608).

Dr. White said it was his understanding that Enterprise did not know the races of applicants during the application review and telephone screening process, and did not know an applicant's race until there was an in-person interview. He said he did not do a study of African-American sounding names in the application database. (TR 1609). Dr. Madden's Table 2A (PX 114) shows that at stages two and three, when Enterprise was aware of race, the disparity is smaller than at stage one where presumably the screeners had no racial information. (TR 1611). Dr. Madden's Table 1A shortfall analysis is based on her Table 3A-revised that only did a regression analysis of stage one. Dr. Madden did not do a regression analysis of offers and, had she done so, the shortfall might be different. (TR 1612).

Dr. White's concerns about Dr. Madden's regression analysis (Table 3A-revised) included: (1) it only analyzed stage one; (2) there are qualifications she did not consider; (3) she did not account for all of the decisions made throughout the hiring process; (4) her methodology for identifying sales and customer service work experience by searching job titles ignored other job information that was available in the application packages; and (5) her methodology for identifying education by looking solely at the application ignored other information available in the application packages that showed completion of education requirements by the time of hiring. (TR 1615-1616). Dr. White believed each period of time should have been analyzed by the factors Enterprise considered during that period and if the factors later changed in other [*91] periods then there should have been controls for the changes. (TR 1617). Instead of doing that, Dr. Madden put all of the data for the entire period into a single model. (TR 1618). His concerns about Dr. Madden's methodology remains the same for all of her tables — original and revised — because the methodologies she employed did not change. (TR 1620).

Dr. White said the regression analysis depicted in the middle of page one of PX 115 matches the first row of Table 3A-revised (PX 116). The pseudo R-squared was 0.0108, which means race accounts for 1.08 percent of the story being told on who passes stage one; it shows one piece of a 100 piece puzzle. (TR 1621). The next regression analysis (PX 115 at 3) includes the age, driver's license, full-time work, visa, authorized to work, plus year factors and it shows that now 15 percent of the story behind who passes through to the next stage is explained. (TR 1622). Dr. White said Dr. Madden is taking a model that explains 15 percent of the story and drawing conclusions that discrimination exists. (TR 1622-1623). The third regression analysis (PX 115 at 3-4) adds controls for sales and customer service experience as Dr. Madden defined them and the pseudo R-squared is 16.56 percent. (TR 1623).

Dr. White's team looked at the 935 applicants in the "9999" line on page one of PX 115 and they found that most were self-select outs at some point in the process. Dr. Madden counted all 935 as rejections in stage one of the hiring process. (TR 1624). In Dr. White's opinion, the ones who self-selected out before stage one should not have been in the analysis and the ones who self-selected out later in the process should have been counted as having passed through stage one rather than as having failed to advance. (TR 1625). Those in the "9999" line who self-selected out in stage one should have been added to the 2,919 applicants in line "8888" who were removed as self-select outs. Dr. White's team was able to determine that 905 of the applicants in the "9999" line were self-select outs by looking at information in the applicant flow log data. (TR 1626).

Dr. White's team prepared a spreadsheet that showed the 905 self-select outs that were in Dr. Madden's "9999" line. (TR 1627; DX 7364). Dr. White said he thought his team was able to identify the applicants listed in the spreadsheet from Dr. Madden's backup data for her Table 3A-revised, but he said he did not know the exact process his team used to compile the data in the spreadsheet. [OFCCP objected to DX 7364, but the objection was withdrawn after it was agreed that OFCCP would have post-hearing time to submit rebuttal evidence if it chose to do so. (TR 1629-1633)]. Dr. White said there were 935 applicants included in the "9999" line and his team found that 905 of those were self-select outs before an offer. (TR 1633). There were 16 applicants who declined interviews, which meant they had progressed beyond stage one, and 14 more who declined screening. (TR 1633-1634). Dr. White said removing the 935 applicants in the "9999" line would change Dr. Madden's regression; he explained that any change — even removing one application — would cause a change in the results. (TR 1634).

On cross-examination by OFCCP, Dr. White said there were 509 white applicants and 396 African-American applicants in the 905 self-select outs shown in the "9999" line. (TR 1635). He would not predict how the majority of the "9999" self-select outs being white would impact the regression analysis. (TR 1637-1638).

[*92] Dr. White agreed that the data entry for the review of all the applications was contracted out to a third party and that his team then checked the data entry. He and his team relied on the information Enterprise reported in the applicant flow logs to do their analysis and only used the database of application information for background information purposes. (TR 1640). He relied on the disposition codes Enterprise assigned to prepare his table entitled "Effects of Sequential Removal by Applicant Flow Code on Offer Shortfall Calculations" that is part of his March 9, 2018 report. (TR 1641; PX 106). Dr. White said his table was sequential and a fact-finder could pick and choose among the disposition codes shown in his table and decide which ones to accept or reject. (TR 1642-1643). He agreed that his table assumed the reasons Enterprise gave for rejecting applicants were legitimate business reasons. (TR 1645).

Dr. White's report showed a shortfall of 190.1 African-American hires and statistical significance at the two standard deviation level in every year except fiscal year 2013.22 (TR 1648-1649). He prepared his calculations by taking the data from the applicant flow logs and excluding the self-select outs, those who did not meet the minimum requirements, and the fiscal year 2017 applications that were still pending. He excluded the 905 self-select outs that were in Dr. Madden's "9999" line. His table is comparable to Dr. Madden's Table 1A, except that his table was prepared in January 2018. (TR 1649-1650). He and Dr. Madden both did a pools analysis, but then she did Fisher's exact test while Dr. White did a test of proportions. He did not calculate a standard deviation for the entire 11-year period. (TR 1650). He said he would not be surprised if the standard deviation for the entire 11-year period was higher than the standard deviation for any of the individual years because of the law of large numbers — more coin flips — means it is more likely a disparity will be statistically significant. (TR 1650-1651).

[*93] Dr. White testified that he did not have an expert opinion on whether the screeners knew the races of the applicants. He was familiar with a study entitled "Are Emily and Greg More Employable than Lakisha and Jamal? A Field Experiment in Labor Market Discrimination," dated July 2003 from the National Bureau of Economic Research. (TR 1651). He believed one of the conclusions was that similarly situated white applicants and African-American applicants were treated differently. He recalled that in the study identical resumes were sent out, except one contained an African-American sounding name like Lakisha and the other contained a white sounding name like Emily. He did not recall the exact statistic, but he was aware that the resumes with African-American sounding names received fewer offers for job interviews than those with white sounding names. (TR 1652-1653). He agreed that in his analysis he did not examine whether screeners at Enterprise might have drawn similar inferences. He agreed that the screeners could see the names on the applications. (TR 1653). Dr. White said he had not looked at the numbers to have an opinion of whether the Baltimore metropolitan area is a racially integrated or segregated. (TR 1654).

Dr. White agreed that Dr. Madden's Table 2 showed a smaller racial disparity as it moved from stage one to stage two and then from stage two to stage three. (TR 1654). He agreed that the number of observations in each stage got smaller, but he said they were still large enough to be statistically significant. He added that if there was a race bias he would expect the numbers to get worse in stages two and three where applicants' races was evident to the Enterprise decision makers. (TR 1655).

Dr. White agreed that one of his criticisms of Dr. Madden's analysis was that she assumed all subjective judgments were biased. He said just because a judgment is subjective does not mean it is necessarily biased. (TR 1657). Dr. White stated:

I believe that if you want to see the effect of something, you run the model with it in, and you run the model with it out. Dr. Madden has not run the model with a lot of things in. She excluded a lot of things from all of her models. And so my interpretation of her testimony is that she thinks they're subjective, and because they're subjective, they may be tainted.

Dr. White agreed that he did not assess what counted as sales and customer service experience. He said he was analyzing job offers, not stage one application review and screening. (TR 1659). He agreed that Dr. Madden's coding decisions on what job titles were or were not included as sales and customer service was not racially biased. He reiterated that he was looking at offers and did not run a regression on stage one as Dr. Madden had done. (TR 1661). Dr. White said he was not saying Dr. Madden was wrong in how she calculated the racial coefficient in her regression (Table 3A-revised), but he was not sure what the percentage meant and he would have converted it to an odds ratio. (TR 1662-1664). Dr. White said including or excluding the 900-plus applications shown on line "9999" of Dr. Madden's table would not have an impact on her shortfall computation. (TR 1664).

[*94] On re-direct examination by Enterprise, Dr. White agreed that the study examining the impact of African-American sounding names and white sounding names on whether identical applicants were invited for interviews used resumes that were exactly the same. He said he did not know if there had been a study of the Enterprise database to determine if any of the application qualifications were exactly the same. (TR 1665).

In response to questions by me, Dr. White agreed that Enterprise would know the races of the applicants its recruiters met through on-campus recruiting, job fairs and open houses. (TR 1668-1669). He said that might change his opinion that stage one was totally race blind, but it would not change his calculations. (TR 1669).

DR. JANICE FANNING MADDEN

(SURREBUTTAL TESTIMONY)

OFCCP recalled Dr. Madden in surrebuttal. Dr. Madden said it is not necessary to model the entire decision-making process to detect racial disparity in hiring. (TR 1673-1674). She said Dr. White's model used disposition codes in sequential order, but in practice they occurred simultaneously rather than in sequence. (TR 1674). She said that in coding job titles as evidence of sales and customer service experience that some job titles may have been improperly included or excluded, but that would be a neutral factor that applied equally to African-American applicants and white applicants. (TR 1674-1675).

Dr. Madden said stage one was where most applicants were rejected and it was the only stage where there was objective information that could be assessed. (TR 1675-1676). She said she was surprised by Dr. White's testimony about pseudo R-squared. Dr. Madden said she was on the review board for several major economics publications and has published herself, and the standard of an R-squared is basically the norm. She said the R-squared values for her equations are quite normal. She said she was not trying to explain why an individual was hired, she was trying to explain the role of race in the process. She said relying on the disposition codes Enterprise entered to assess whether there was racial disparity is "just so unjustifiable" and that assuming the decisions were racially neutral would mean you would never find discrimination anywhere. (TR 1680-1681).

I asked Dr. Madden to explain "racial coefficient" as the term applies to Table 3A-revised. She said the higher the number the worse off African-Americans applicants are and if the number is negative then white applicants are worse off. She said a score of 1.0 would be the maximum. (TR 1682).

TRIAL EXHIBITS

The parties submitted the following documents into evidence.

LEGAL FRAMEWORK

In a case brought under Executive Order 11246, the legal standards developed under Title VII of the Civil Rights Act of 1964, as amended, 42 U.S. Code § 2000e , are applicable and provide that it is unlawful for an employer to discriminate against any individual with respect to employment based on an individual's race. OFCCP v. Bank of America, ARB Case No. 13-099 (Apr. 21, 2016). The plaintiff has the burden of establishing a prima facie case of discrimination. Albemarle Paper Co. v. Moody, 422 U.S. 405, 425 (1975). Proof of discriminatory intent is required, but it may be based upon circumstantial evidence, including statistical evidence. An unlawful motive can be inferred from a showing of a disparity between class members and comparably qualified members who are not part of the minority group. Hazelwood School District v. United States, 433 U.S. 299, 307 (1977). A prima facie case of a pattern or practice of discrimination may be entirely statistical. Hazelwood, supra ; OFCCP v. Greenwood Mills, Inc., ARB Case No. 1989-OFC-00039 , slip op. at 21-2, 45 (Sec'y, Nov. 20, 1995). A statistical disparity in the treatment of minorities may have one of three explanations: (1) it is the product of unlawful discriminatory animus; (2) there is a legitimate nondiscriminatory cause; and (3) it may be the product of chance. Palmer v. Schulz, 815 F.2d 84, 91 (D.C. Cir. 1987). If the disparity is significant enough — which shows that the probability it resulted from chance is negligible — it may be inferred that the disparity is the result of unlawful animus. Hazelwood, supra, 433 U.S. at 307-308 .

Statistical evidence may be used to rule out chance as a likely reason for a significant racial disparity and the courts have consistently found significance in disparities exceeding two standard deviations. Hazelwood, supra, at 308 , n.14; Adams v. Ameritech, 231 F.3d 414, 424 (7th Cir. 2000). Ruling out chance does not automatically demonstrate that race discrimination was a motivating factor, but it does make such a reason a viable factor that can be inferred. Watson v. Fort Worth Bank & Trust, 487 U.S. 977, 988, 994 (1988). The more significant the statistical disparity, the less additional evidence that is needed to establish that the reason was race discrimination. Extreme cases of statistical disparity may allow a trier of fact to find that intentional racial discrimination occurred without the need for additional evidence. Tagatz v. Marquette University, 861 F.2d 1040, 1044 (7th Cir. 1988). The burden is on the plaintiff to present sufficient evidence to convince the fact-finder that racial discrimination was a motivating factor in the employer's hiring decisions.

If the plaintiff establishes a prima facie case, the burden shifts to the employer to rebut it by showing that the plaintiff's statistical evidence is inadequate or inaccurate. Greenwood Mills, supra , slip op. at 22. The employer can do this by attacking the plaintiff's statistical methods or by showing that the racial disparity resulted from legitimate, non-discriminatory factors. Palmer v. Schultz, supra, at 99 . If the employer proffers evidence that the disparity was the result of [*99] legitimate, non-discriminatory factors, the plaintiff can prevail by establishing that the factors were used as a pretext for unlawful discrimination. McDonnell Douglas Corp. v. Green, 411 U.S. 792 (1993).

OFCCP'S PRIMA FACIE CASE

OFCCP has alleged that Enterprise discriminated against African-American applicants for employment as management trainees between August 1, 2007 and July 31, 2017. In order to establish a prima facie case, OFCCP must establish by a preponderance of the evidence that there was a substantial disparity and that race was a cause. OFCCP may satisfy its burden by presenting statistical evidence from which it can be inferred that discrimination occurred and the greater the statistical severity the less additional evidence that is necessary for OFCCP to meet its burden of proof. Where the statistical disparity is extreme, discrimination may be inferred without additional evidence. Bank of America, supra, at 11-13 . As Enterprise's expert witness, Dr. Paul White, explained in his report, two standard deviations provides 95 percent confidence and three standard deviations provides 99 percent confidence in statistical conclusions. (PX 106 at 15).

OFCCP's expert witness, Dr. Janice Fanning Madden, analyzed application flow log data for the management trainee program to assess whether there was racial disparity in hiring. She concluded that the statistical probability of the shortfall in the number of African-American applicants that received job offers for management trainee positions occurred by chance in the absence of consideration of race was one million times less likely than the probability of a single ticket matching the six numbers and winning the Powerball Lottery. (TR 1055; PX 113 at 5). Dr. Madden found statistically significant disparities in each stage of the hiring process that in most instances she concluded constituted a "virtual certainty." (TR 1060). She ran a regression analysis on the application review and telephone screening stage because it was the only stage in the hiring process where there were objective data to evaluate. (TR 1130-1131). She assumed that the group of African-American applicants and the group of white applicants as a whole had comparable qualifications, but she did not assume that individual applicants were comparable. (TR 1187-1188).

Dr. Madden revised her calculations during the course of the hearing to remove applications that had been screened by computer and rejected automatically for failing to meet the minimum requirements, which is the kind of objective evaluation of applications Dr. Madden encouraged. She had been unaware that Enterprise implemented computer screening in 2009. She agreed that the computer screening rejected about 59 percent (58.6 percent) of the African-American applicants assessed during the charged period, but she noted African-Americans were about 55 percent (55.6 percent) of the total applicant pool and the difference was not statistically significant. (TR 1514; 1592).

Dr. Madden compared the number of African-American applicants that received job offers with the number of African-American applicants and white applicants combined that received offers for each year and for the entire period, and she assumed African-American applicants and white applicants as a whole were comparable. That assessment is reflected in Table 1A. As an example, Dr. Madden said that in 2008, African-Americans were 49 percent of [*100] the applicants Enterprise considered, but they received 23 percent of the job offers. (PX 113 at 22-24). She then tested whether her assumption that African-American applicants and white applicants were, as a whole, similarly qualified by controlling for the hiring criteria Enterprise used consistently over the charged period. That assessment is reflected in Table 3A-revised.

Dr. Madden found that there was a shortfall of African-American job offers in each year from fiscal year 2007 through fiscal year 2017, and the standard deviations for those years ranged from a low of 1.8 in 2013 to a high of 6.2 in 2008. The standard deviation was greater than two in 10 of the 11 years and it was greater than three in 7 of the 11 years. She calculated a total shortfall of 196 offers to African-American applicants for the 11-year period and a standard deviation of 12.7 for the entire period. (PX 114, Table 1A). Most applicants as a whole, and most African-American applicants in particular, were rejected in the application review and telephone screening stage. Dr. Madden found a shortfall of African-American applicants that progressed through that stage to in-person interviews in each of the 11 years with standard deviations that ranged from a low of 2.5 in 2011 to a high of 8.2 in 2008, and a standard deviation of 16.4 for the entire 11-year period. (PX 114, Table 2A). Finally, Dr. Madden did a regression analysis to see if the assumption that African-American applicants and white applicants were comparable as a whole on the key characteristics Enterprise identified as minimum requirements (college degree, met minimum age, had a driver's license, seeking full-time employment, and had a visa or was legally authorized to work in the U.S., and then adding sales and customer service work experience) and she found that African-American applicants were comparable to if not better distributed than white applicants on the minimum requirements. (TR 1286-1287). After controlling for all of the key characteristics, Dr. Madden calculated a standard deviation of 12.53 for the 11-year period. (PX 116).

I find that OFCCP has established a prima facie case that Enterprise unlawfully discriminated against African-American applicants for its management trainee program in the period from fiscal years 2007 through fiscal year 2017 with the exception of fiscal year 2013 where the standard deviation was less than two. The calculations by Dr. Madden and Dr. White were nearly identical and found statistically significant racial disparities greater than two standard deviations in each year in the charged period except for fiscal year 2013. (PX 114 at Table 1A; PX 106 at 17). Their statistical findings are buttressed by the testimony of the African-American applicants who were rejected for the management trainee program, the testimony of the Enterprise employees involved in the screening and interview process, and the review of application files showing inconsistent treatment of white applicants and African-American applicants.

ENTERPRISE'S REBUTTAL

The burden shifts to Enterprise to show that OFCCP's statistical analysis is insufficient or that the racial disparity was the result of legitimate, non-discriminatory reasons. Enterprise raised a variety of arguments challenging OFCCP's prima facie case as summarized below:

• (1) OFCCP failed to prove by a preponderance of the evidence that Enterprise engaged in a pattern or practice of discrimination. (Defendant's Post-Hearing Brief at 52).
  

[*101] Enterprise contends that Dr. Madden's statistical analysis is irrelevant and insufficient for OFCCP to sustain its burden to establish a pattern or practice of discrimination. Enterprise argues that: (1) Dr. Madden did not consider the core competencies Enterprise established for management trainees; (2) Dr. Madden did not consider Enterprise's "race-blind" evaluation of sales or customer service experience; (3) Dr. Madden did not consider information Enterprise recorded in the application files; (4) Dr. Madden created her own evaluation approach and ignored the approach Enterprise created; and (5) Dr. Madden ignored Enterprise's six-step hiring process. As Enterprise states in its closing brief, "[t]his case turns almost entirely on Enterprise's subjective evaluations of candidates' work experience and the Core Competencies ..." (Defendant's Post-Hearing Brief at 53).

There was significant common ground between Dr. Madden and Dr. White. They both agreed that the largest number of applicants were rejected in the application review and telephone screening stage and that the racial disparity was statistically significant in all of the years except for fiscal year 2013. (TR 1101, 1388; PX 58 at 154). Dr. Madden said she did a regression analysis of the application review and telephone screening stage because that was the only stage in the hiring process where objective information was available for analysis. She did not do a regression analysis of the latter stages because in those stages Enterprise personnel made subjective judgments about applicants and whether those judgments were influenced by race was the fundamental question under evaluation. (TR 1130-1131).

Dr. Madden found a statistical significant disparity in the first stage of the hiring process in each year of the relevant time period and Dr. White did not dispute her findings. When asked if he was able to find any year that lacked statistically significant racial disparity, he responded, "No. This disparity is significant for the first stage for each of the years." (PX 58 at 154). Dr. White said that in some studies a disparity in one direction in one stage might have an offsetting disparity in the opposite direction in another stage. Here, however, he said the disparity in stage one did not have any offsetting disparity in any of the subsequent stages. (DX 58 at 153). Common sense would suggest that if a group's attrition in the first stage is disproportionate then in the normal course of events it is likely to remain disproportionate in the stages that follow.

I have considered each of Enterprise's arguments concerning alleged deficiencies in the way Dr. Madden conducted her statistical analyses and I find that those arguments lack merit. Dr. White testified that he did not take issue with Dr. Madden's mathematics, but he questioned some of the premises she used in doing her calculations. (DX 58 at 154, 239, 272). For instance, he questioned whether the keywords she identified to search for job titles in assessing sales and customer service experience captured all of the relevant job titles that were reflected in the applications. (DX 58 at 7-9). He testified at the hearing: "We saw examples in the data where it looked like she did not count job titles that one could argue are sales and customer service related," although he admitted he had no reason to believe she purposefully included or excluded particular job titles and that he did not do any testing to assess whether her keyword titles were racially biased. (DX 58 at 165-167). Dr. White did not cite any job titles that were included or excluded that had a material impact on Dr. Madden's conclusions or that he believed might alter the outcome of her analysis. Dr. Madden said that reasonable people could disagree about which job titles should or should not be included in the analysis, but as long as the list was reasonable [*102] and it was not racially biased a study of the racial effects after controlling for job titles would not be biased. (PX 107 at 8). She said she would be happy to add or delete job titles and analyze the results again, but no one had asked her to do so. In the end, Enterprise has not identified or established that any deficiencies could reasonably call into question the validity of Dr. Madden's statistical conclusions.

The experts' discourse on job titles underscores the reasonableness of Dr. Madden's decision to do a regression analysis on the application review and telephone screening stage where there were objective data to assess: Assessing subjective judgments requires applying subjective judgments. When Dr. White was asked if he had gone back to the data in the application files to investigate factors that appeared to have had a greater impact on racial disparity, he said "[t]hat would have involved going through hundreds if not thousands of applications and making some kind of judgment as to what's fair and what's not." (DX 58 at 177; TR 1439). He testified that he accepted the disposition codes as Enterprise entered them into the system and he did not refer back to the applications to check whether the codes had been applied fairly because "[w]e weren't going to impose our judgment over the judgment of those who were actually doing the process." (TR 1439). It is important to remember that the judgments made by those engaged in the process is the core issue here. Dr. Madden said in the application review and telephone screening stage, she had the information the applicants entered on their applications rather than the subjective impressions of Enterprise's hiring personnel. She assumed that as a whole African-American applicants and white applicants were equally likely to provide accurate information in their applications, and I note that there was no evidence to the contrary.

Enterprise's claim that the application review and telephone screening stage of the hiring process was "race-blind" and that members of the recruiting team "did not know the race of the candidates whom they phoned" is contrary to the weight of the evidence. (Defendant's Post Hearing Brief at 55). Ms. Lichter acknowledged that sometimes "[y]ou could make a guess" that an applicant was African-American from the information contained in the application, like the applicant's name or whether he or she attended a historically black college or university, and she said that sometimes from the applicant's voice on the telephone "I could guess" whether the applicant was African-American. (TR 931). Mr. Wucher denied that he would know an applicant's race from information contained in an application, but when asked if he might infer that an applicant named "Da'Shawn" was African-American he said, "[s]ure, I could make an educated guess. Yes." (TR 826). He also said it was "[m]ore likely than not" that an applicant who graduated from high school in Nigeria was African-American. (TR 829).

That is not to suggest that Enterprise's screeners deliberately sought out information about applicants' races or that the information contained in the applications always signaled an applicant's race to the screeners, but a reasonable person applying common sense and knowledge of the ways of the world who surveys the names on the applications that are in the record (DX 223-7361) could often reasonably infer an applicant's race as well as other personal characteristics, such as age and gender even though those details are not specifically documented in the application. I find the assertion that the application review and telephone screening stage of the hiring process was totally race-blind is inaccurate.

• (2) [*103] Dr. Madden's statistical tables do not establish a prima facie case of a pattern or practice of intentional discrimination. (Defendant's Post-Hearing Brief at 62; Defendant's Post-Hearing Reply Brief at 10 and 27).
  The establishment of a prima facie case was discussed above.
  
• (3) Enterprise produced statistical evidence showing non-discrimination. (Defendant's Post-Hearing Brief at 72; Defendant's Post-Hearing Reply Brief at 30).
  

Enterprise claims that Dr. White's statistical analysis demonstrates that it did not discriminate against African-American applicants.

Dr. White and Dr. Madden agreed that they would expect the percentage of African-Americans in the applicant pool and the percentage of all applicants offered jobs to be approximately equal, and they agreed on the expression of statistical significance of the difference between the actual versus expected rates by standard deviations.23 (PX 106 at 13). Dr. White's table showing shortfalls and standard deviations for each year (PX 106 at 18) and Dr. Madden's Table 1A (PX 114 at 1) are nearly identical in their calculations of African-American shortfalls and the standard deviations for the disparities; Dr. White calculated a shortfall of 190 African-American job offers while Dr. Madden calculated a shortfall of 196 for the 11-year period. Dr. White did not disagree with Dr. Madden's statement that the two possible explanations for the statistically significant disparity between the expected number of job offers extended to African-American applicants and the actual number of job offers they received were (1) African-American applicants are less qualified as a whole than white applicants or (2) Enterprise applied its hiring criteria in a racially discriminatory manner. (TR 1474-1477).

The additional analysis Dr. White conducted that Enterprise argues demonstrates statistically that it did not engage in a pattern or practice of intentional discrimination was to sequentially apply the disposition codes Enterprise assigned for rejecting applicants. By the end of the sequence, Dr. White calculates that the shortfall is reduced to zero. (PX 106 at 20).

Enterprise cites Wards Cove Packing Co. v. Atonio, 490 U.S. 642, 659-660 (1989), and notes that its burden is only to produce evidence of its non-discriminatory reasons for its hiring decisions and that the burden of persuasion is on OFCCP to establish intentional discrimination. (Defendant's Post-Hearing Reply Brief at 30-31). Nonetheless, Enterprise argues that by presenting Dr. White's analysis it has "proffered its own far more accurate statistics" that trumps OFCCP's statistical evidence produced by Dr. Madden. (Defendant's Post Hearing Brief at 73).

Recognizing that Enterprise only has a burden of production, not of persuasion, I find that its view of the weight that should be afforded Dr. White's statistical evidence concerning the sequential application of the disposition codes is overstated.

[*104] First, Dr. White identified three steps where the alleged bases for rejections had a statistically significant impact on African-American applicants: work experience (S3),

communication skills (I1) and compatible career and direction (I7). (PX 106 at 19-20). When asked if his analysis showed "statistically significant racial disparities at many points in the process," Dr. White responded, "Yes, it does." (TR 1456). There were no non-discriminatory reasons offered for the statistically significant racial disparities in the application of the three dispositions codes Dr. White identified.

Second, while for purposes of his analysis Dr. White accepted at face value the disposition codes Enterprise entered, the personnel responsible for applying and entering those codes testified that some were clear yes or no binary choices, but others called for the application of their personal subjective judgments. As Mr. Wucher testified: "Almost all of the phone screen is essentially black or white. They either have a driver's license or they don't." (TR 790). With respect to the work experience requirement — the disposition code Mr. Wucher stated "is easily the one that I use probably the most, S3" to reject applicants (TR 806) — he said:

And what I'm looking for in this is just them being themselves. They don't even need at this point really to talk necessarily about their job. It helps. But if they can sell themselves, if they can be that confident leader that we want to hire them for — because again, we're not just trying to hire management trainees. We're trying to hire for future, you know, people that are going to run stores and run areas. And if they can have that sort of confidence and leadership ability, and be persuasive in that nature, then absolutely, come on in.

Mr. Wucher's testimony about the subjective judgment screeners applied in deciding whether an applicant met the sales and customer service work experience requirement was consistent with the testimony of Ms. Hardesty and Ms. Lichter. Ms. Lichter said two applicants could have held the exact same job, had the exact same job title, and worked for the exact same company and one could get credit for sales or customer service experience and the other rejected for a lack of the same depending upon how they articulated their experiences during their telephone conversations with her. (TR 918). Ms. Hardesty testified that an applicant who had worked as a tutor might get credit for sales or customer service experience (TR 615), but Ms. Lichter rejected Dr. Worrell and entered an S3 code because she believed "the primary job function of a tutor is not to sell, it's to tutor, so it wasn't, according to the screening guidelines, relevant sales experience." (TR 906).

[*105] The way Enterprise defined its minimum requirement for experience in sales or customer service in its screening criteria appeared on paper at least to be an objective factor throughout the period, although it evolved over time. (PX 10A-10C). In other words, applicants either had the required amount of experience or they did not. The testimony showed, however, that in actual application the screeners exercised substantial subjective judgment in assessing whether applicants would proceed on or be eliminated from the hiring process based upon the work experience criteria. Rather than being an objective yes or no decision like a driver's license, whether an applicant was deemed to meet the minimum requirement for sales or customer service work experience depended in large part on the subjective impression he or she made on the screener during a five to ten minute telephone conversation and the screener's personal sense of whether the applicant was likely to be successful in the future in Enterprise's sales and customer service oriented environment. As Mr. Wucher testified, when he was screening for relevant work experience he was looking for applicants to "basically persuade me to have them come in." (TR 791). This reflects a subjective assessment of an applicant's future potential based on a present sense impression rather than an objective verification of an applicant's past employment.

Third, Dr. White's analysis showed the significant difference between objective and subjective judgments in the assessment of African-American applicants for the management trainee program. Whether applicants met the minimum age requirement or had driver's licenses had the same impact on both African-American applicants and white applicants, but whether they met the minimum requirement for sales and customer service experience had a significantly greater impact on African-American applicants. (PX 106 at 20, 56-57). Dr. White said he accepted the reasons Enterprise assigned for rejecting applicants and he did not attempt to determine whether they were applied in a discriminatory or non-discriminatory manner. He said to do so would require reviewing all of the applications and making judgments on whether the disposition codes were applied consistently to African-American applicants and white applicants. (TR 1439, 1456, 1463). Dr. White said his goal was to provide the fact-finder with a way to understand the steps in Enterprise's decision-making process and the fact-finder could then decide whether each of the steps was legitimate. He acknowledged that he did not analyze whether Enterprise applied its criteria in a race-neutral manner. (TR 1420-1421, 1429-1430; PX 58 at 249-250).

It is clear that Enterprise did not apply all of the disposition codes consistently. That is patently obvious with respect to the S3 code that was used frequently and had a disparate impact on African-American applicants. Dr. White noted a statistically significant racial impact at both the two and three standard deviation levels for the S3 code. The testimony of Ms. Morris, Ms. Hardesty, Ms. Lichter and Mr. Wucher showed the highly subjective and seemingly arbitrary nature of what Enterprise counted as relevant work experience; where potentially two applicants could have held the same job and one gets credit for it and the other does not. Mr. Wucher described it best when he said he was looking for applicants to "sell themselves" at that moment and that it was helpful, but not necessary, for them to even talk about their prior work experience. The testimony of each screener was summarized in detail above, including their reviews and discussions of a number of the applicant files in which they were involved. It showed that what on paper purported to be an objective minimum standard based upon relevant past work experience was in application a subjective judgment based upon whether applicants [*106] made a favorable impression on the telephone screeners at that moment about their future potential for success at Enterprise. That point was also highlighted by three of the four unsuccessful applicants who testified at the hearing. Dr. Worrell, Ms. O'Hara and Ms. Lowery were all rejected and assigned S3 disposition codes despite reporting work experience that was comparable to the past work experience white applicants reported who were deemed to meet the minimum requirement.24

Accordingly, I find that Enterprise's claim that it "proffered its own far more accurate statistics" is not supported by the evidence of record.

• (4) Enterprise articulated legitimate, non-discriminatory reasons for its decisions. (Defendant's Post-Hearing Brief at 73).
  

There is no dispute that Enterprise assigned a reason for every application it rejected. In Dr. White's analysis, he accepted those reasons at face value and did not test to see whether they were applied consistently despite acknowledging that there was a statistically significant racial disparity in job offers in 9 of 11 years at the two standard deviation level and in 7 of 11 years at the three standard deviation level. (PX 106 at 18). Dr. Madden performed a regression analysis on the application review and telephone interview stage and, when she controlled for the minimum qualifications Enterprise articulated, she found a statistically significant racial disparity and that African-American applicants were "substantially better situated than whites, were more qualified, overall." (TR 1502, PX 114).

The 60-plus individual applications the parties introduced and discussed at the hearing also called into question whether the reasons for Enterprise's decisions were legitimate and non-discriminatory. For instance, Enterprise did not credit Dr. Worrell's work experience coaching students on mathematics, chemistry and physics, but credited M.H.'s work experience coaching students on softball. (TR 476-477, PX 90, PX 108). Enterprise did not give T.J. credit for four years of work experience as a realtor selling and marketing houses, but gave J.R. credit for work experience painting houses. (Revised Attachment A to Defendant's Post-Hearing Brief, dated Sep. 28, 2018, at 5, 40, 47, 52). Dr. Worrell and T.J. are African-American and were rejected for lack of relevant work experience. M.H. and J.R. are white, were deemed to satisfy the work experience requirement, and were offered management trainee positions.

Examining four applications from one fiscal year out of thousands of applications submitted over an 11-year period, and observing that standards were applied in an inconsistent manner that was more lenient for the white applicants than the African-American applicants, in and of itself might not cast substantial doubt on the legitimacy of the hiring process. Here, however, that point has to be considered in the context of the statistical evidence from both experts that showed a significant racial disparity and the testimony of the Enterprise personnel involved in the hiring process that showed a substantial difference in how Enterprise defined its minimum requirements in writing and how they were actually applied in day-to-day practice.

• (5) [*107] OFCCP failed to sustain its ultimate burden of showing a pattern or practice of intentional discrimination. (Defendant's Post-Hearing Brief at 73; Defendant's Post-Hearing Reply Brief at 35, 41 and 42).
  

Enterprise contends that OFCCP did not disprove the legitimacy of the criteria it established as qualifications for the management trainee program. (Defendant's Post-Hearing Brief at 73-74). It also contends that OFCCP "cherry-picked" 13 applicants to highlight in its brief, yet presented no evidence that any white applicant was hired instead of an equally qualified African-American applicant. (Defendant's Post-Hearing Reply Brief at 35).

Enterprise's arguments miss the mark. There was no dispute that the criteria Enterprise established for management trainees appeared to be legitimately tailored to its business needs, but the evidence was clear that the criteria as they appeared on paper were not applied in a consistent manner in actual practice. That point stood out most vividly on the work experience criteria. Requiring applicants for entry level positions involving sales and customer service to demonstrate that they have successfully performed sales and customer service work before makes logical sense. That, however, was not how the standard was applied. As has been discussed several times before, Enterprise's screeners rejected applicants and applied S3 disposition codes based upon the screeners' subjective present sense impressions of applicants rather than objective assessments of applicants' past work experiences. That is how in the context Ms. Lichter described two applicants could have identical work histories and one would move on to the interview stage and the other would be rejected for a lack of relevant past work experience.

In International Brotherhood of Teamsters v. United States, 431 U.S. 324, 360 (1977), the Supreme Court said that "the Government is not required to offer evidence that each person for whom it will ultimately seek relief was a victim of the employer's discriminatory policy. Its burden is to establish a prima facie case that such a policy existed. The burden then shifts to the employer to defeat the prima facie showing of a pattern or practice by demonstrating that the Government's proof is either inaccurate or insignificant." In Bank of America , the Administrative Review Board said that the more extreme the statistical disparity the less additional evidence that is necessary to establish racial discrimination as a cause and that in cases of severe statistical disparities discriminatory intent may be inferred. Bank of America, supra , slip op. at 13-14.

This case is in many respects akin to Bank of America . As the Board noted there, "this case is not a case of a rosy version of the statistics competing with a troubling version; all the versions are troubling." Id. at 15 . Dr. Madden and Dr. White both found statistically significant racial disparities in every year of the charged period except for fiscal year 2013, and their calculations are virtually identical. The experts' statistical calculations should come as no surprise to Enterprise. Annual impact ratio analyses ("IRAs") repeatedly warned Enterprise that African-Americans were being hired at rates that were significantly less than white applicants, yet Mr. Arney and Ms. Morris testified that year after year those warnings led to some internal discussions, but no changes to their processes. Dr. White's sequential removal of shortfalls by disposition codes was the only potentially rosy version of the statistics, but his syllogism fails [*108] because one of his premises — that disposition codes were fairly and consistently applied — failed. The statistical results were buttressed by the testimony of African-American applicants Enterprise rejected, the testimony of Enterprise personnel involved in the hiring process, and the review of applications that showed inconsistent outcomes when Enterprise personnel made subjective judgments. As an example, a white applicant was deemed to have sales and customer service experience for painting houses while an African-American applicant was rejected for a lack of sales and customer service experience despite several years of experience selling houses. For all of these reasons, I find that OFCCP met its burden to prove there was a pattern or practice of intentional discrimination.

• (6) Enterprise used a structured behavioral interview process that constitutes the gold standard. (Defendant's Post-Hearing Brief at 74; Defendant's Post-Hearing Reply Brief at 31).
  

As Enterprise states in its brief, OFCCP has not challenged the structured behavioral interview process it employed, which it says is the process the Office of Personnel Management recommends Federal agencies use. OFCCP did not dispute that the minimum requirements Enterprise established to screen applicants for the management training program were legitimate standards for the position nor did OFCCP dispute that the core competencies Enterprise articulated were reasonably related to the job. The issue here is not the process as it existed on paper, it is the process as it was applied in actual practice. Wall v. Mississippi Dept. of Public Welfare, 730 F.2d 306, 313, 319 (5th Cir. 1984) (discrimination occurred despite the state using a multiple stage hiring process implemented to comply with federal funding requirements — "fair in form, but discriminatory in effect."). Additionally, Dr. Madden and Dr. White both found statistically significant racial disparities in the application review and telephone screening stage in each year of the 11-year period, which results in a smaller share of African-America applicants making it to the structured behavioral interview stage.

Enterprise submitted summary descriptions of 31 applicants that were discussed during the hearing. (Revised Attachment A to Defendant's Post-Hearing Brief, dated Sep. 28, 2018). It is impossible to compare each of the thousands of management trainee applicants to one another. Likewise, it is not useful to attempt to do a point by point comparison of all 31 of the applicants listed in Attachment A; however, a few comments are warranted.

With respect to Ms. Evans, under column "G" where Enterprise presents its arguments against OFCCP's allegations of race-based discrimination, it states:

In the 2nd Interview notes Brian McVickers rejected Ms. Evans because he felt she was not "results-driven" and also commented that he felt Ms. Evans was "looking at ERAC for experience" rather than a career and therefore was not a good candidate for the Management Trainee position at ERAC.

Applicant T.J., who is African-American, had a degree in marketing from Temple University and was working as a realtor at the time she applied for a management trainee position. She had been in real estate sales for four years at that point and had been involved in 170 real estate transactions, including the sale of 89 residential properties. She said she was involved in the sales and marketing of apartments, commercial properties and residential real estate, and she noted that she worked with lenders, home inspectors, title companies and appraisers. The telephone screening was conducted by Ms. Lichter who rejected T.J. for a lack of relevant sales or customer service experience and entered an S3 disposition code. On the Phone Screen form, Ms. Lichter wrote: "Real Estate since 2003 — solicit sales by marketing homes to buyers. Provide cust with information & meet needs." (PX 82). T.J. did not advance to the structured behavioral interview stage and was not offered a management trainee position.

Applicant B.M. was white and had a degree in mass media communications from Salisbury State University. He reported on his application that he worked nearly four years as an office assistant at a law firm where he organized and maintained files and did general office upkeep. In his resume, B.M. also reported that he had co-hosted a radio program once a week for 11 months; produced and directed five episodes of a university television program; and was the night manager at a gas station where he closed up and secured the money, oversaw shift employees and ordered inventory. Ms. Morris met B.M. on-campus and there was no telephone screening. Enterprise offered B.M. a management trainee position. (PX 99; DX 4567).

Applicant M.H. was white and had a degree in psychology from Florida Institute of Technology. On her application, M.H. reported that she had worked as an assistant softball coach at a college for five months, as a treatment counselor for emotionally and behaviorally challenged children for four months, and as a personal assistant to a mentally and physically challenged elementary school student for nine months. Her resume also noted that she was a private softball instructor, had worked in sales at a produce store for two months, and was a secretary/carpenter at a concrete and construction company for a year and a half. Enterprise states in Attachment A that M.H. "successfully moved past the phone screen based on her discussion of her previous sales/customer service experience as a softball coach as explained during the phone screening." (Attachment A at 47). Enterprise offered M.H. a management trainee position. (PX 108). I thoroughly reviewed PX 108 as well as DX 3070, which contain the same documents pertaining to M.H. There is no telephone screening form in either exhibit; therefore, I have no way to determine whether M.H. was screened by telephone or how her work experience was described by whoever conducted her telephone screening and found that she satisfied the sales and customer service work experience requirement.

Applicant K.M. was white and had a degree in political science from Frostburg State University. On his application he listed a three month internship working with the Department of Homeland Security installing chemical and biological detection equipment and a job as a [*110] laborer installing insulation on plumbing and heating and air conditioning equipment. Enterprise states in Attachment A that K.M. "successfully moved past the phone screen based on his previous sales/customer service experience as explained during the phone screen." (Attachment A at 51). Enterprise offered K.M. a management trainee position. (PX 108). I thoroughly reviewed PX 108 as well as DX 4103, which contain the same documents pertaining to K.M. There is no telephone screening form in either exhibit; therefore, I have no way to determine whether K.M. was screened by telephone or how his work experience was described by whoever conducted his telephone screening and found that he satisfied the sales and customer service work experience requirement.

Applicant J.R. was white and had a degree in marketing from Niagara University. On his application he listed work as a maintenance manager at Wysocki's Lake Park Manor and as a house painter. He met an Enterprise recruiter at an on-campus event and did not have a telephone screening. Enterprise offered J.R. a management trainee position. (PX 108).

The actions taken on these few applications alone do not establish that Enterprise intentionally discriminated against African-American applicants. They do, however, illustrate the sweeping discretion Enterprise applied in making subjective judgments about what they did or did not count as relevant sales and customer service work experience, and this disproportionally prevented African-American applicants from reaching the structured interview stage.

• (7) OFCCP presented no evidence that Enterprise's interviewers intentionally discriminated. (Defendant's Post-Hearing Brief at 80).
  

The cases cited in Enterprise's brief are not persuasive because the context here is distinctively different. For example, in Anderson v. Douglas & Lomason Co., Inc., 26 F.3d 1277, 1298 (5th Cir. 1994), cert. denied, 513 U.S. 1149 (1995), the court held that the collective belief of some plaintiffs that the company refused to give them employment applications because of their race was mere speculation where the facts showed that other African-Americans received applications during the same period and the employer established that it had a policy of not issuing applications when there were no positions to fill or when it had enough applications to fill its vacancies. In Carroll v. Sears, Roebuck & Co., 708 F.2d 183, 195-96 (5th Cir. 1983), the court held that the plaintiff's belief that he was not hired for a full-time position was mere speculation when he admitted that he was hired for a part-time position because there were no full-time positions available.

The facts in Wall v. Mississippi Dept. of Public Welfare, 730 F.2d 306, 322 (5th Cir. 1984), are more comparable to the facts in the present case. There, several Mississippi state agencies used a multi-stage hiring process implemented to comply with requirements mandated in order to receive federal funding. Individuals seeking employment submitted applications to the state merit system agency where they were screened to determine whether the applicants met the educational requirements for the positions for which they applied. Applicants who met the educational requirements then took a written test and all of those who achieved a certain score were placed on a registry of eligible candidates for employment. When a vacancy arose, the county director where the vacancy existed received a list of the top five eligible candidates [*111] seeking employment in the county. The county director interviewed the candidates who were on the list and then submitted written evaluations and recommendations to the state welfare commissioner who made the formal selections, which invariably followed the county directors' recommendations. While on paper the hiring process looked reasonable, African-American candidates were about half as likely to get hired as white candidates. The court said:

Discriminatory intent may be established by either direct or circumstantial evidence. Statistical evidence may be used in a disparate treatment case to prove both motive and a pattern or practice of racial discrimination if gross statistical disparities in the composition of an employer's work force can be shown. To the extent that statistical proof is not sufficiently strong, the plaintiff's prima facie case cannot be made without additional evidence of the defendant's intentional discrimination. In the present case, the trial court considered a statistical analysis performed by the plaintiffs' expert statistician, Dr. Peter Kolesar, together with evidence of the paucity of blacks employed in the contested positions, and the anecdotal evidence of eight black individuals who had unsuccessfully applied for Department positions. The court found that the plaintiffs successfully made out a prima facie case upon proof that the state Department treated some people less favorably than others because of their race, and that the defendants had not successfully shown that the statistics relied on by the plaintiffs were either "inaccurate or insignificant." (citations omitted).

Likewise, the cases Enterprise cites involving studies related to the visual identification of race are not particularly pertinent in the context of this case. Here, Ms. Lichter and Mr. Wucher said they could sometimes infer an applicant was African-American based upon information contained in the application, such as the applicant's name or attendance at a historically black college or university. Dr. Madden testified that based upon her 45 years of experience in labor economics and statistics she was aware that racial inferences could be drawn from names and Dr. White testified that he was familiar with a study where identical resumes were sent out and those with names that suggested applicants were African-American were significantly less likely to get invited to come for interviews than those with names that suggested the applicants were white. That same study was cited by Justices Ginsburg and Souter in a 2003 Supreme Court decision involving race.26

[*112] Here, discriminatory intent is established by statistical evidence bolstered by anecdotal evidence that included the testimony of African-America applicants Enterprise rejected, the testimony of Enterprise personnel involved in the screening and interviewing process, and the review of applications. The evidence as a whole establishes a pattern or practice of racial discrimination.

• (8) OFCCP failed to prove its disparate impact claim by a preponderance of the evidence. (Defendant's Post-Hearing Brief at 82; Defendant's Post-Hearing Reply Brief at 45).
  

In Griggs v. Duke Power, 401 U.S. 424, 431 (1971), the Supreme said that with Title VII Congress intended to remove "artificial, arbitrary, and unnecessary barriers to employment when the barriers operate invidiously to discriminate on the basis of racial or other impermissible classification." The Court explained:

The Act proscribes not only overt discrimination but also practices that are fair in form, but discriminatory in operation. The touchstone is business necessity. If an employment practice which operates to exclude Negroes cannot be shown to be related to job performance, the practice is prohibited.

Enterprise contends that OFCCP failed to prove that any of its practices discriminated against African-Americans and that it demonstrated that it had legitimate, non-discriminatory business reasons for its hiring practices. Accepting those contentions requires taking a very myopic view of the record.

As noted earlier, the standards and processes Enterprise articulated on paper for hiring management trainees look fine. Put into practice, however, those processes take on an entirely different patina.

The U.S. Court of Appeals for the D.C. Circuit addressed a disparate impact claim in a recent decision involving the District of Columbia Child and Family Services Agency. Davis v. District of Columbia, ___ F.3d ___ (June 7, 2019). In that case, the plaintiffs were long-time agency employees who were terminated during a reduction-in-force ("RIF") necessitated by budget cuts. Plaintiffs claimed that the job categories the agency selected for elimination had a disproportionate adverse impact on African-Americans. The agency argued that its RIF decisions were not actionable because they did not involve an objective test or requirement, but instead were a series of subjective, contextual judgments made in "multiple individual decisions by the agency leadership," which it claimed could not be challenged on a disparate-impact theory. Id. at 4. The district court granted summary judgment for the agency saying that the plaintiffs failed to identify a specific employment practice that had a racially disparate impact.

The D.C. Circuit disagreed. The D.C. Circuit held that a disparate impact claim could include "the processes by which the Agency identified plaintiffs' jobs for elimination as "particular employment practice[s]." Id. at 7. The Court said:

[*113] A disparate impact claim contends that an observed disparity caused by a particular employment practice cannot be justified as necessary to the employer's business. The purpose of disparate impact analysis under Title VII is to permit plaintiffs to challenge "practices, procedures, or tests" that may be "neutral on their face, and even neutral in terms of intent," but that disproportionately harm members of a protected class. Griggs, 401 U.S. at 430 , 91 S.Ct. 849 . Whereas disparate treatment requires a showing of discriminatory motive, disparate impact supports liability in the absence of proof of invidious intent, based on evidence that the challenged practices have a disproportionately adverse effect on the plaintiffs that cannot be justified as necessary to an employer's business.

In this case, the evidence shows that the seemingly race-neutral standards Enterprise articulated for hiring management trainees had a disproportionate adverse impact on African-American applicants. The statistical evidence presented by Dr. Madden and Dr. White showed that the racial disparity in job offers to African-American applicants was well in excess of two standard deviations in every year of the charged period except for fiscal year 2013. That was consistent with Enterprise's own internally-generated data from its IRAs that repeatedly warned of an adverse impact on African-Americans. Dr. Madden used a regression analysis to examine whether differences in data reported on applications accounted for African-American applicant's failing to progress to the interview stage and she controlled for the characteristics Enterprise identified as key characteristics for management trainees. When all of the characteristics were considered, including customer service and sales experience, the failure rate for African-American applicants was 67 percent greater and the analysis showed that the differences in the characteristics of African-American applicants and white applicants did not account for the racial disparity in who was selected to move on to the interview stage. (PX 107 at 14-15; PX 114 at Table 3A). Dr. White's sequential removals from the shortfall tally using Enterprise's disposition codes highlighted three codes that had a disproportionate adverse impact on African-American applicants: customer service and sales experience (S3), communication ability (I1), and compatible career and direction (I7). (PX 106 at 19-20). There was no evidence introduced that African-American applicants as a whole were less likely than white applicants to have had jobs involving sales and customer service, were less effective in communicating, or were less likely to desire a career in sales and customer service.

SUMMARY OF FINDINGS

I found by a preponderance of the evidence that OFCCP met its burden to establish a prima facie case that Enterprise intentionally and unlawfully discriminated against African-American applicants for its management trainee program in 10 of the 11 years in the charged period.27 Dr. Madden and Dr. White both found that the statistical significance of the racial disparity in fiscal year 2013 was less than two standard deviations; therefore, fiscal year 2013 is not included. OFCCP's statistical evidence was buttressed by anecdotal evidence that included [*114] the testimony of African-American applicants that Enterprise had rejected, the testimony of Enterprise personnel involved in screening applications and interviewing applicants, and comparisons of applications that led to inconsistent outcomes. Dr. White did not dispute Dr. Madden's assertion that the only two explanations for the outcomes in hiring were that African-Americans as a whole were less qualified than white applicants or Enterprise applied the criteria in a racially discriminatory manner.

The burden shifted to Enterprise to show that OFCCP's proof was inaccurate or insignificant. Enterprise could not rebut the statistical evidence by simply arguing that its decisions were legitimate or non-discriminatory. See International Brotherhood of Teamsters v. United States, et al., 431 U.S. 324, 360 n.46 (1977). Instead, Enterprise had to show that the pattern or practice of discrimination established by the statistical evidence did not exist either by successfully attacking Dr. Madden's statistical methodology or by showing that all of the statistical disparities were explained by legitimate, non-discriminatory reasons. Enterprise could not meet its burden by arguing that its hiring personnel did not intend to discriminate. A finding of ulterior motives is not a prerequisite to a finding of intentional discrimination. See Greenwood Mills, supra , slip op. at 13.

Dr. White identified what he perceived as deficiencies in Dr. Madden's statistical analysis. (PX 106). Some of his concerns were addressed during the course of the hearing, such as Dr. Madden's inclusion of self-select outs and computer screened applications in her initial calculations. Dr. Madden revised her figures as reflected in PX 114-116. There was no showing that Dr. White's concerns about the job titles Dr. Madden chose to assess sales and customer service experience and whether applicants had bachelor's degrees meaningfully influenced the results. Dr. White said he did not believe there was any racial bias in the way Dr. Madden classified job titles for her study. (PX 58 at 267-268). As Dr. Madden explained, if some things were missed or added, if they were examined overall in a neutral manner, there was no reason to believe they had a racially disparate impact on the results. In other words, including or excluding a job title should not have a materially greater or lesser impact on a particular race and there was no evidence presented to the contrary.

Dr. White argued that the pseudo R-squared values for Dr. Madden's regression analysis of the application review and telephone screening stage had a low explanatory power. Dr. Madden countered that pseudo R-squared measures the overall variation in selection of individual applicants, not groups, and it is not a measure of the reliability of a study. She testified that the pseudo R-squared values here were "quite normal" and that her objective was to examine the role of race in the selection process as a whole, not to explain why an individual applicant was or was not hired. (TR 1678-1679). I find that the evidence on what inferences can be drawn from the pseudo R-squared values is in equipoise. Since the burden is on Enterprise to demonstrate that OFCCP's statistical evidence is inaccurate, it does not assist Enterprise in meeting its burden.

Overall, I found that Enterprise connected a few jabs in its effort to knockout OFCCP's statistical evidence as inaccurate, but it did not score a TKO or score more points for punches landed. In Brown v. Nucor Corp., 785 F.3d 895, 904 (4th Cir. 2015), the Fourth Circuit observed: "The critical question is thus not whether the data used is perfect but instead whether [*115] it is reliable and probative of discrimination. To that end, a court must examine whether any statistical assumptions made in the analysis are reasonable." OFCCP's statistical evidence was reliable and probative, and the assumption made were reasonable.

Dr. White's analysis showed there was statistically significant racial disparity at the two standard deviation level in all 11 years of the charged period if no applicants were excluded from the pool. He found statistically significant racial disparity at the two standard deviation level in every year except fiscal year 2013 when he excluded self-select outs, applicants that did not qualify, and applications from fiscal year 2017 that were pending at the time of his analysis. He did a shortfall analysis where he removed applicants sequentially from the pool based upon the reasons Enterprise gave for rejecting the applicants. Using that approach, Dr. White reduced the shortfall in job offers to African-American applicants down to zero. (PX 106 at 20). Dr. White accepted that the disposition codes Enterprise applied were all used and applied properly. In other words, Dr. White accepted that the disposition codes reflected legitimate, non-discriminatory business reasons for rejecting applicants. When asked if his analysis showed any evidence of racial bias in the application of the independent variables by Enterprise's personnel, he answered, "I've not seen evidence one way or the other on that." (TR 1430).

I found that the sequential removal of African-American applicants from the shortfall based upon the disposition codes Enterprise assigned carries little evidentiary weight. Dr. White determined that there were three disposition codes that had a large impact on African-American applicants: sales and customer service work experience (S3), communication ability (I1) and compatible career and direction (I7).28 (PX 106 at 19). Unlike whether an applicant did or did not meet the minimum age requirement or had a valid driver's license, the three criteria that Dr. White found had a significant adverse impact on African-American applicants were subjective and no race-neutral justifications for the disparities were offered.

Mr. Wucher testified that he never used some of the disposition codes, other he used infrequently and that the reason he assigned for rejecting applicants most often was the S3 sales and customer service work experience disposition code. (TR 806). Ms. Hardesty also testified that some disposition codes were used more often than others. Mr. Wucher testified that in assessing work experience he was not as interested in hearing about whether applicants met the stated objective criteria based on past employment as he was in them subjectively selling him on having what it took to warrant moving on to the interview stage. He admitted that it was a subjective judgment based on the impression an applicant made on the telephone rather than whether he or she satisfied the objective past work experience requirement, which he did not treat as a hard and fast requirement.

Ms. Lichter determined that a white applicant who had worked at Target in loss prevention met the sales and customer service work experience requirement, but an African-American applicant who had worked at Target handling customer service transactions did not [*116] and was assigned an S3 disposition code. (TR 896-897; PX 20A, PX 20B, PX 75, PX 76). The capricious manner in which Enterprise assigned S3 disposition codes to African-American applicants was also illustrated through testimony and review of multiple applications, which showed, for instance, that white applicants who tutored individuals on softball or painted houses met the sales and customer service work experience standard while African-American applicants who tutored individuals on chemistry and physics or sold houses did not.

For all of these reasons, I found that the assumption that the reasons Enterprise gave for rejecting African-American applicants were legitimate and non-discriminatory was not an assumption that was supported by the weight of the evidence. Accordingly, Enterprise failed to successfully attack the methodology or significance of OFCCP's statistical evidence showing a significant adverse impact on African-American applicants, or that the statistical disparities are explained by legitimate, nondiscriminatory reasons; therefore, Enterprise has not carried its burden to rebut OFCCP's prima facie case of unlawful discrimination against African-American applicants for the management trainee program.

DAMAGES

The legal standards the courts developed under Title VII of the Civil Rights Act of 1964 in employment discrimination cases apply to employment discrimination cases brought under EO 11246. See Greenwood Mills, supra , slip op. at 5; OFCCP v. Cleveland Clinic Foundation, 1991-OFC-00020 , slip op. at 3 (ARB July 17, 1996); U.S. Dep't of Labor v. Honeywell, Inc., 1977-OFC-00003 , slip op. at 10 (Sec'y June 2, 1993). A remedy is appropriate "to make persons whole for injuries suffered on account of unlawful employment discrimination." Albemarle Paper Co. v. Moody, 422 U.S. 405, 418 (1975). Back pay is one element of the "make whole" relief that can be provided to victims of employment discrimination. Individuals need not have filed a complaint as a prerequisite to OFCCP seeking such relief on their behalf. Interest on back pay is calculated from the date of the loss and compounded quarterly at the percentage rate established by the Internal Revenue Service for the under-payment of taxes. 41 C.F.R. § 60-1.26(a)(2) .

A class-wide formula can be used to calculate a back-pay award rather than attempting to assess each victim individually in an effort to calculate the damages they sustained separately. See McClain v. Lufkin Indus., Inc., 519 F.3d 264, 280-281 (5th Cir. 2008); Segar, 738 F. 2d at 1289-1291 ; Greenwood Mills, supra , slip op. at 5-6. Such a class-wide approach may be necessary in cases that are complex, the class is large, or the illegal discrimination extended over a long period of time. McClain, 519 F.3d at 280-81 ; Pettway v. American Cast Iron Pipe Co., 494 F.2d 211, 261 (5th Cir.1974).

A class-wide approach is warranted in this case. After excluding fiscal year 2013 applications and applications that did not qualify by computer screening or where applicants self-selected out, there were well over 6,500 applications in the pool for the relevant years. (PX 106 at 18; PX 114 at 1). A single application file can be as few as five pages or more than 105 pages in length (see DX 2005 and DX 4406). While it is not literally impossible to review the tens of thousands of pages in the application files, it is clearly impractical and unlikely to permit meaningful individualized determinations. Brown v. Colegio de Abogados de Puerto Rico, 613 F.3d 44, 53-54 [*117] (1st Cir. 2010), cert. denied, 562 U.S. 1200 (2011). This is the type of case where attempting to craft individualized determinations would produce the "quagmire of hypothetical judgments" the court warned of in Pettaway v. American Cast Iron Pipe Co., 494 F.2d 211, 260 (5th Cir. 1974), cert. denied, 439 U.S. 1115 (1979).

The two experts were in general agreement on the methodology for the calculation of damages. Dr. Madden used figures provided by Dr. White and she concluded that the assumptions he drew from government reports in doing his calculations were "quite reasonable." (TR 1137). She said she had "quibbles" with some of the adjustments he made to the database, but she found them to be reasonable. (PX 107 at 2).

In his report, Dr. White calculated potential lost earnings by year for applicants who were employed and those who were unemployed when they applied for the management trainee program, and he calculated a weighted average for any single African-American applicant included in the shortfall.29 (PX 106 at 22-25). He applied those figures to the annual shortfalls as determined by Dr. Madden.30

[*118] There are some discrepancies in the experts' calculations. The two are in agreement on the numbers of African-American applicants and white applicants in the pool after did not qualify and self-select outs were removed except for fiscal years 2007-2009. Dr. Madden has higher figures for both groups in all three of those years. As a result, Dr. Madden's shortfall in offers is higher (196 total) than Dr. White's shortfall (190.1). (PX 106 at 18; PX 114 at Table 1A). I am unable to determine why the experts' figures differ in the first three years of the charged period. Since Dr. White's calculation is the more conservative — his total shortfall is included in Dr. Madden's total shortfall, but not vice versa — I will use Dr. White's calculation of 190.1 as the total shortfall in offers to African-American applicants for the entire charged period. Since I found that fiscal year 2013 was below the two standard deviation threshold and discrimination was not established for that year, the 2013 offer shortfall of 8.1 is subtracted and leaves a total offer shortfall of 182.

Mr. Arney testified that about 100 to 120 management trainees are hired each year. (TR 226). According to Dr. White, between 81 and 100 percent of the offers Enterprise made for positions as management trainees were accepted during the years included in the charged period and in the most recent year — fiscal year 2017 — the offer acceptance rate was 90 percent. (PX 106 at 25). Common sense suggests that applicants in the early years of the charged period are likely to be a decade or more into other careers and probably many would not be inclined to leave their current jobs to start anew in an entry level position. Applicants in more recent years, on the other hand, may still have an interest in starting a career at Enterprise. It is impossible to predict with any degree of certainty what percentage of the class would accept offers for positions as management trainees. There is at least a reasonable probability that if 182 offers were extended all at once it could result in more class members accepting offers than Enterprise has openings to fill and an entire year group of management trainees being wholly African-Americans.

I recommend that offers for management trainee positions be extended to 182 members of the class.31 To avoid potentially having more class members accept offers than there are existing requirements, half of the offers will be extended in fiscal year 2020 and the second half of the offers will be extended in fiscal year 2021. Any management trainee vacancies that remain open during the period after offers are made to class members will be available to current applicants who have applied for the management trainee program. The 10 years where discrimination occurred can be divided into three periods where the offer shortfalls are relatively equal. Using Dr. White's figures, in 2007-2009 there was an offer shortfall of 59.5, which I round up to 60. In 2010-2012 the shortfall was 63.1, which I round down to 63. In 2014-2017 [*119] the shortfall was 59.3, which I round down to 59. The first 91 offers will be made to 60 class members in the 2007-2009 group, 16 class members in the 2010-2012 group, and 15 class members in the 2014-2017 group. The second 91 offers will be made to 47 class members in the 2010-2012 group and to 44 class members in the 2014-2017 group. I find that this is an equitable way to allocate job offers over the two fiscal year period. Enterprise will determine the rates of pay and benefits for members of the class who accept offers based upon the dates their applications were rejected. Seniority will be established using the same practice.

As the Supreme Court made clear in Albemarle Paper and Nucor, supra , the goal is to make persons whole if they suffered from unlawful employment discrimination. In the context of this case, that would mean putting class members into the positions they would have been in now had they been hired as management trainees in the years in which they applied and were rejected. The normal progression would be from management trainee to assistant branch manager to branch manager to area manager and to group manager. According to Ms. Morris, at any given point in time Enterprise Group 18 has 50 to 60 branches, 11 to 15 areas and 3 groups. Class members in the early part of the charged period, had then been hired when they applied and had they remained at Enterprise, would likely be group managers or in comparable positions by now. There are only three group manager positions in total. There was no indication that Group 18 had any vacant group manager positions at the time of the hearing. Similarly, class members in the middle of the charged period, had then been hired when they applied and had they remained at Enterprise, would likely be area managers or in comparable positions by now and those in the latter part of the period would likely be assistant branch managers or branch managers by now. There are only 11 to 15 area manager positions. There was no indication that Group 18 had vacant area manager positions. There are 50 to 60 branches, so there are more branch manager and assistant branch manager positions, and likely there are some vacancies now or there will be some vacancies in the foreseeable future.

Regardless of the level within the organization, directing Enterprise to place class members into the positions they likely would hold now but for unlawful employment discrimination is not a realistic option. Success for individuals as they move up the organizational ladder is predicated in large part on the knowledge and experience gained starting with frontline customer interactions as management trainees. Without that grounding in real-world experience, the individual is set up for failure and the consequences for the company are likely to be adverse. Accordingly, I do not recommend ordering Enterprise to make direct appointments of class members into senior-level positions. Perhaps as part of a jointly developed remedial plan the parties can fashion a path for class members who accept management trainee offers to be considered for advancement in an expedited manner or to award front pay, but it is beyond my purview to define that path or to craft the remedy in this recommended decision.

Dr. Madden and Dr. White were in agreement on the formula for calculating mitigated lost earnings, benefits and interest, although they applied those financial calculations to slightly different hiring shortfall numbers, which produced slightly different totals in lost earnings and benefits. The figures in the chart below represent the shortfall in management trainee job offers to African-American applicants in the relevant years as calculated by Dr. White applied to the offer acceptance rates and annual financial losses that Dr. Madden and Dr. White accepted as [*120] accurate. The monetary calculations are as of December 31, 2017 and the parties will need to have their experts verify the figures and bring them forward to present value.

Lost Earnings and Benefits with Interest

OFCCP requests a declaration that Enterprise failed to preserve and maintain all personnel and employment records for a period of two years as required by § 202 of Executive Order 11246 and 41 C.F.R. § 60-1.12(a) and (e) . (Plaintiff's Post-Hearing Brief at 77-78). Enterprise responds that applications for 23 applicants in the fiscal year 2006 to fiscal year 2008 period could not be produced because the electronic files were corrupted in an electronic database application it stopped using in 2008. (Defendant's Post-Hearing Reply Brief at 54). Enterprise acknowledged the same in its response to OFCCP's request for admissions (PX 45 at 22-24) and in Mr. Arney's testimony (TR 172, 223-224), and there was no evidence introduced to the contrary. While Enterprise failed to preserve and maintain the 23 applications — a fact that it readily admits — there is no indication that those 23 records out of the thousands of records that were produced were intentionally corrupted in order to hide information that would be adverse to Enterprise. Additionally, the standard deviations for fiscal years 2007 and 2008 were both substantially greater than two, so there is no indication that the 23 records would have statistical significance even if all 23 contained information that was adverse to Enterprise. Accordingly, I decline to apply the presumption permitted by 41 C.F.R. § 60-1.12(e) .

OFCCP requests a declaration that Enterprise failed to conduct an adverse impact analysis of its total selection process for all positions in accordance with 41 C.F.R. §§ 60-3.4 and 60-3.15(A)(2)(a) , and failed to develop an auditing system to periodically measure the success of its affirmative action program in accordance with 41 C.F.R. § 60-2.17(d) . (Plaintiff's Post-Hearing Brief at 77-78). Enterprise responds that OFCCP failed to offer any evidence to support its claims and, in fact, admitted that Enterprise conducted adverse impact analysis of its management trainee program and that it monitored the program. (Defendant's Post-Hearing Reply Brief at 54-55).

[*121] The record shows that Enterprise prepared affirmative action plans and conducted impact ratio analyses. (TR 145-175, TR 193-204, TR 416-432; PX 23-24, PX 110-112, DX 85-103). Mr. Arney testified that when Enterprise submitted its affirmative action plans to OFCCP they were never returned for corrective action. (TR 225). He also testified, and Enterprise confirmed in its response to interrogatories, that neither he nor anyone else at Enterprise took any actions when the data showed an adverse impact on African-American applicants. (TR 174-175; PX 47 at 74-75). The burden is on OFCCP to establish the alleged violations.

From my review of the record, it appears that Enterprise, at a minimum, went through the motions of analyzing its hiring processes and measuring its affirmative action program, even if doing so only resulted in some brief internal discussions that led to no actions to address the warning signs the data provided. Perhaps Enterprise's efforts were deficient, but OFCCP has not established that to be the case. I note that in 138 pages of post-hearing briefing OFCCP devoted less that one page to these issues and that limited discussion is general rather than focused on specific deficiencies. Accordingly, I do not find that Enterprise failed to conduct an adverse impact analysis or failed to develop an auditing system to measure the success of its affirmative action program.

Finally, OFCCP requests that Enterprise's current federal government contracts be canceled and that it be barred from future government contracts until it remedies the harm done by its past discriminatory hiring practices and comes into compliance with Executive Order 11246. (Plaintiff's Post-Hearing Brief at 78, Plaintiff's Reply to Defendant's Post-Hearing Brief at 60).

41 C.F.R. § 60-1.27(b) states:

Debarment. A contractor may be debarred from receiving future contracts or modifications or extensions of existing contracts, subject to reinstatement pursuant to § 60-1.31 , for any violation of Executive Order 11246 or the implementing rules, regulations and orders of the Secretary of Labor. Debarment may be imposed for an indefinite term or for a fixed minimum period of at least six months.

Debarment is intended to induce a company to come into compliance with the Executive Order, not as punishment for violating the Executive Order. Uniroyal, Inc. v. Marshall, 482 F.Supp. 364, 376 n.27 (D.D.C. 1979). Having listened to and observed the current and former Enterprise employees that testified at the hearing, it was not my impression that any of them were bad people with bad motives. On the other hand, it was clear that Enterprise was on notice from the results of the impact ratio analysis that it generated on a regular basis that there were disproportionately bad outcomes for African-American applicants for the management trainee program, yet no changes were implemented to address the disparities. The racial disparity continued unabated and unaddressed even after OFCCP issued the Notice of Violation in 2013. Debarment from current and future government contracts is warranted for an indefinite term until these issues are addressed and rectified.

[*122] The U.S. District Court for the District of Columbia noted in Uniroyal that "a party must be told more than that, in order to be removed from a government blacklist, it must satisfy an administrative official of its compliance with unknown, unspecified, and unpredictable conditions." Uniroyal, supra, at 375 . In this case, in order for Enterprise to be removed from the debarment list, it must agree to implement the steps set forth herein to address the effects of discrimination in the past and to prevent discrimination from occurring in the future. I am not equipped to define in detail what must be done to prevent discrimination from continuing going forward, but at a minimum the steps should include preventing screeners from exercising subjective judgment in assessing whether applicants satisfy the objective criteria Enterprise establishes as minimum requirements for new hires and for both screeners and interviewers to adequately document and justify their reasons for rejecting applicants.

The parties are in a better position to determine a concrete course of action for preventing discrimination from occurring in the future; therefore, I recommend that an order indefinitely barring Defendant from current and future government contracts be held in abeyance for a period of 90 days to give the parties an opportunity to confer on a plan of action. If the parties agree to a plan of action, I recommend that an order of debarment be suspended through the end of the company's fiscal year 2021 at which time it will expire unless Plaintiff has previously filed a motion for an order of debarment with the Office of Administrative Law Judges, which will retain jurisdiction for that purpose.

RECOMMENDED FINDINGS AND ORDER

1. Plaintiff's expert, Dr. Janice Fanning Manning, and Defendant's expert, Dr. Paul F. White, both found a disparity in the hiring of African-American applicants for management trainee positions with statistical significance in excess of two standard deviations in each fiscal year from 2007 through 2012 and 2014 through 2017.

2. The statistical evidence was buttressed by the testimony of four African-American applicants who had been rejected for management trainee positions, three Enterprise employees that conducted applicant screenings, and the review of more than 60 application files. This anecdotal evidence highlighted the arbitrary manner in which information about applicants' qualifications was assessed resulting in the inconsistent application of criteria that substantially disadvantaged African-American applicants.

3. The statistical and anecdotal evidence established a prima facie case that Defendant unlawfully discriminated against African-American applicants for its management trainee program in the period from fiscal years 2007 through 2012 and fiscal years 2014 through 2017.

4. Defendant did not meet its burden to establish that the statistical evidence was either inaccurate or explained by legitimate, non-discriminatory reasons.

5. Defendant is ordered to immediately comply with the requirements of Executive Order 11246 and the implementing regulations.

[*123] 6. The statistical analysis conducted by Plaintiff's expert and by Defendant's expert established a shortfall of 182 offers for management trainee positions to African-American applicants during the years where there were statistically significant racial disparities. Accordingly, Defendant is ordered to extend job offers to 182 class members. Class members who accept offers for management trainee position will upon beginning work accrue pay and benefits as if they had been hired when they applied and were rejected. Unless the parties mutually agree otherwise, offers will be extended as follows:

7. Defendant's expert and Plaintiff's expert agreed on the methodology for calculating lost earnings, benefits and interest for individuals in each year group where there were statistically significant racial disparities. Applying their methodology results in a total amount of back pay and interest of $6,645,444.00 as of December 31, 2017. The parties are directed to verify the accuracy of this computation and to adjust it to bring it forward to present value. The verified amount will be distributed to members of the class in accordance with the ratios in the Lost Earnings and Benefits with Interest table set forth above.

8. Indefinite debarment is appropriate to insure compliance with Executive Order 11246. An order of indefinite debarment will be held in abeyance for 90 days from the effective date of this order to provide the parties an opportunity to confer on a plan of action. If the parties agree on a plan of action, an order of debarment is suspended through the end of the company's fiscal year 2021 at which time it will expire unless Plaintiff has previously filed a motion for an order [*124] of debarment with the Office of Administrative Law Judges, which retains jurisdiction for that purpose.

SO ORDERED.

MORRIS D. DAVIS

Administrative Law Judge

NOTICE OF APPEAL RIGHTS: To appeal, you must file exceptions ("Exception") with the Administrative Review Board ("Board") within fourteen (14) days of the date of receipt of the administrative law judge's recommended decision.

The Board's address is: Administrative Review Board, U.S. Department of Labor, Suite S-5220, 200 Constitution Avenue, NW, Washington DC 20210, for traditional paper filing. Alternatively, the Board offers an Electronic File and Service Request (EFSR) system. The EFSR for electronic filing (eFile) permits the submission of forms and documents to the Board through the Internet instead of using postal mail and fax. The EFSR portal allows parties to file new appeals electronically, receive electronic service of Board issuances, file briefs and motions electronically, and check the status of existing appeals via a web-based interface accessible 24 hours every day. No paper copies need be filed.

An e-Filer must register as a user, by filing an online registration form. To register, the e-Filer must have a valid e-mail address. The Board must validate the e-Filer before he or she may file any e-Filed document. After the Board has accepted an e-Filing, it is handled just as it would be had it been filed in a more traditional manner. e-Filers will also have access to electronic service (eService), which is simply a way to receive documents, issued by the Board, through the Internet instead of mailing paper notices/documents.

Information regarding registration for access to the EFSR system, as well as a step by step user guide and FAQs can be found at: https://dol-appeals.entellitrak.com. If you have any questions or comments, please contact: Boards-EFSR-Help@dol.gov

[*125] If filing paper copies, you must file an original and four copies of the Exception with the Board, together with one copy of this decision. If you e-File your Exception, only one copy need be uploaded.

Any request for an extension of time to file the Exception must be filed with the Board, and copies served simultaneously on all other parties, no later than three (3) days before the Exception is due. See 41 C.F.R. § 60-30.28 .

On the same date you file the Exception with the Board, a copy of the Exception must be served on each party to the proceeding. Within fourteen (14) days of the date of receipt of the Exception by a party, the party may submit a response to the Exception with the Board. Any request for an extension of time to file a response to the Exception must be filed with the Board, and copies served simultaneously on all other parties, no later than three (3) days before the response is due. See 41 C.F.R. § 60-30.28 .

Even if no Exception is timely filed, the administrative law judge's recommended decision, along with the record, is automatically forwarded to the Board for a final administrative order. See 41 C.F.R. § 60-30.27 .