Rhodes v. Cracker Barrel Old Country Store, Inc.

Full title: KELVIS RHODES, STEPHEN WILSON, CONNIE REGAN, JERRY BARBEE, SANDRA KEEL…

Court: United States District Court, N.D. Georgia, Rome Division

Date published: Dec 31, 2002

Civil Action No. 4:99-CV-217-HLM (N.D. Ga. Dec. 31, 2002)

Opinion

Civil Action No. 4:99-CV-217-HLM

December 31, 2002

ORDER FOR SERVICE OF REPORT AND RECOMMENDATION

WALTER E. JOHNSON, United States Magistrate Judge

Attached is the Report and Recommendation of the United States Magistrate Judge made in accordance with 28 U.S.C. § 636 (b)(i)(B), Federal Rule of Civil Procedure 72(b), and the Court's Local Rule 72.1B. Let the same be filed and a copy, together with a copy of this Order, be served upon counsel for the parties.

Pursuant to 28 U.S.C. § 636 (b)(i), each party may file written objections, if any, to the Report and Recommendation within ten days of the receipt of this Order. Should objections be filed, they shall specify with particularity the alleged error(s) made (including reference by page number to any transcripts if applicable) and shall be served upon the opposing party. The party filing objections will be responsible for obtaining and filing the transcript of any evidentiary hearing for review by the District Court. If no objections are filed, the Report and Recommendation may be adopted as the opinion and order of the District Court, and any appellate review of factual findings will be limited to a plain error review. United States v. Slay, 714 F.2d 1093, 1095 (11th Cir. 1983).

The Clerk is directed to submit the Report and Recommendation with objections, if any, to the District Court after expiration of the above time period.

REPORT RECOMMENDATION

I. INTRODUCTION

Seven putatively representative African-American plaintiffs accuse defendant, Cracker Barrel Old Country Store, Inc. (hereafter "Cracker Barrel" or the "Company"), of racial discrimination against them and the class they seek to represent. Am. Compl. [92]. This action is before the Court on Plaintiffs' Motion for Class Certification [82]. For the reasons stated below, the undersigned recommends that the Motion be DENIED.¹ II. PLAINTIFFS' ALLEGATIONS AND PROPOSED CLASS DEFINITION

1 Two other related pending motions — Defendant's Motion to Strike Plaintiffs' Declarations in Support of Motion for Class Certification [100] and Plaintiffs' Motion to Strike Cracker Barrel's Employee-Witness Declarations [111] — are addressed in an Order issued concurrently herewith.

In their Amended Class Action Complaint [92], the seven putatively representative plaintiffs (hereafter collectively "plaintiffs") allege that they bring this action on behalf of themselves and similarly situated persons of their race to redress Cracker Barrel's alleged continuing systemic racial discrimination in employment through the use of discriminatory selection and compensation procedures. Am. Compl. [92] ¶ 1. They seek a declaratory judgment that defendant has engaged in systemic racial discrimination in violation of Title VII of the Civil Rights Act of 1964, as amended, 42 U.S.C. § 2000e et seq. (hereafter "Title VII"), and 42 U.S.C. § 1981 (hereafter "§ 1981"). Id. ¶ 2. Plaintiffs allege that they and the class have been subjected to a pattern and practice of intentional discrimination and a battery of practices having unlawful disparate impact on their employment opportunities. Id. ¶¶ 140, 142-43.

Plaintiffs seek equitable relief necessary to make the members of the class whole, to undo the effects of defendant's alleged racial discrimination, and to prevent such discrimination from continuing. Such relief includes promotions, back pay, attorney's fees, prejudgment interest, expenses, affirmative restructuring of defendant's selection and compensation procedures, training, and establishment of a task force on equality and fairness. Am. Compl. [92] ¶¶ 2-3, 26-27, 146-48. However, plaintiffs seek no compensatory or punitive damages.

Plaintiffs allege that employment opportunities at Cracker Barrel are structured in terms of pay scale and prestige. At the bottom tier are those employees who work behind the scenes or in the kitchen in so-called "back-of-the-house" job classifications, which include (1) dishwasher; (2) night maintenance; (3) prep cook; (4) back-up cook; and (5) grill cook. Am. Compl. [92] ¶ 8. The middle-tier positions are those that have regular interaction with customers in so-called "front-of-the-house" job classifications, which include (1) cashier; (2) host or hostess; and (3) server. Id. The top tier positions are those in store management, which include store general manager and various assistant and/or associate store managers. Id.

Plaintiffs allege that Cracker Barrel denies African Americans desirable job assignments, promotional opportunities, training, management positions, compensation, and other benefits of employment on the same terms applied to Whites.² Am. Compl. [92] ¶ 24. Specifically, plaintiffs allege that Cracker Barrel "channels" African Americans into positions with limited opportunity for promotion, such as dishwasher; deters African Americans from seeking promotions, management positions, and desirable job assignments; fails to provide African-American employees with regular evaluations of their skills; ignores, and in some cases actively supports, racism among its staff; and fails to enforce policies prohibiting racial discrimination. Id.; see also id. ¶ 141 (listing seven discriminatory practices, the first being "channeling" to back-of-the-house positions).

2 The Court uses the term "White" instead of Caucasian because plaintiffs use that term in their filings.

According to plaintiffs, Cracker Barrel routinely advertises for front-of-the-house positions like server or hostess. Am. Compl. [92] ¶ 25. Both Whites and African Americans apply. However, White applicants are routinely offered the advertised jobs, but African-American applicants are routinely offered back-of-the-house positions as dishwashers. Id. Because of this illegal policy, plaintiffs allege that Cracker Barrel has two workforces. Id. One workforce, which is primarily White, enjoys preferential treatment, better job opportunities, and a swift path to advancement into management. Id. The other workforce, which is primarily African American, holds a disproportionate share of the lowest-level positions, is denied equal terms and conditions of employment, and, with few exceptions, has not been allowed to advance.Id.

Plaintiffs allege that they meet the four requirements for class certification under Fed.R.Civ.P. 23(a): commonality, Am. Compl. [92] ¶¶ 28-30; typicality, id. ¶ 31; numerosity, id. ¶ 32; and adequacy of representation. Id. ¶ 33. In addition to satisfying the requirements of Rule 23(a), plaintiffs claim that they satisfy the requirements for certification under Fed.R.Civ.P. 23(b)(2). Id. ¶¶ 35-37. Therefore, they seek to be certified as representatives of the following class:

All African-Americans who applied for employment or were employed by Cracker Barrel Old Country Store, Inc. at any time since April 9, 1994, or who would have applied during such period in the absence of the employment practices challenged in this lawsuit.

Pls.' Mot. for Class Certification [82] at 1. (In the alternative, plaintiffs seek certification with subclasses. Id. at 2 ¶ 2.)³ III. THE PARTIES A. Defendant Cracker Barrel 1. Organizational Structure and Workforce

3 Plaintiffs' Motion for Class Certification does not request an evidentiary hearing. Both sides have filed designations of the evidence they would present if an evidentiary hearing were held, including depositions, appendices of documents, declarations, and expert reports. The Court has reviewed those designations and finds that an evidentiary hearing, even if one had been requested, is not needed.

Cracker Barrel owns and operates a chain of approximately 454 restaurants located in approximately forty-one states, primarily in the Southeast, Midwest, Mid-Atlantic, and Southwest. Am. Compl. [92] ¶ 6. Each restaurant consists of a rustic, country-store-style building, which houses a dining room, a retail gift shop, a kitchen, and storage areas. Id. ¶ 7. Each restaurant (or store) employs approximately 100 people over two shifts. Id. ¶ 9. Overall, the Company employs approximately 50,000 people. Id. ¶ 10.

Cracker Barrel's stores are grouped into districts, with the districts grouped into regions. Store assignments to districts and regions changed during the relevant time period, as well as the size of the districts and regions. Over the period 1995 through 2001, the average number of stores in a region was 100, while the average number of stores in a district was twelve. See Report submitted by defendant's experts, Drs. Joan Haworth and Janet Thornton (dated March 22, 2002, filed in defendant's Statistical Evidence Appendix [hereafter "H/T Report"]) at 9-10.

Although the African-American composition of Cracker Barrel's employees was relatively stable (averaging 12% between 1995 and 2001), the presence of African-American employees in each store varied significantly, generally reflecting the local labor market's racial composition.⁴ H/T Report at 10. Stores also differ with respect to their local labor markets, pay rates, and the employment and promotion opportunities available to applicants and employees. Id. In the average store, the primary hourly positions include fifty servers, nineteen dishwashers, nine cashiers, nine hosts/hostesses, eight grill cooks, six retail-shop employees, three back-up cooks, three prep cooks, and two night maintenance employees. Id. There is a variation in the relative mix of jobs among stores and in store size. Id. at 10-11. The average store has 115 employees, but the largest has 235 and the smallest has 59. Id. The retail space available also varies from 1,300 to 2,500 square feet. Id. at 11.

4 For example, the percent African American ranged from 0% (stores no. 8, Corbin, Kentucky, and no. 443 in Corydon, Indiana) to 66% (store no. 455, Albany, Georgia). These percentages closely mirrored the percent African American among those 18 and older in those labor markets (i.e., 0.1% in Corbin, Kentucky; 1.3% in Corydon, Indiana; and 60.4% in Albany, Georgia). H/T Report at 10.

The Company has a large workforce with high turnover. In 1998, for example, the average size of the workforce was 39,305 employees, but there were over 74,000 terminations and approximately 80,000 hires and re-hires. H/T Report at 22. The average hourly employee worked seventeen weeks. Id. The average employee working in PAR 0 (the PAR system is described below) worked eight weeks, while the average employee in PAR IV worked thirty-five weeks. Id. Dishwashers worked the fewest average number of weeks (11) while servers worked the most (19).⁵ Id. The average hourly employee worked only 28.7 hours per week. Id. at 23. Employees in PAR 0 worked an average of 24.1 hours per week compared to 29.9 hours per week for those in PAR IV. Id. Hosts/hostesses worked the fewest hours per week (18.2), while grill cooks worked the most (30.1).Id. 2. Cracker Barrel's Employment Policies

5 Dishwasher had the highest turnover rate in 1998 — 266%. H/T Report at 43. Thus, this job has the most opportunities for placement of new hires. The jobs of server, prep cook, host, and grill cook had average turnover rates ranging from 179% to 192% in 1998. Id. The lowest turnover rate was among retail shop employees (116%). Id.

Plaintiffs allege that Cracker Barrel maintains either actual or constructive control over the operations, including employment practices, of each restaurant. Am. Compl. [92] ¶ 6. In response to that allegation, defendant submitted the following Appendices detailing aspects of its employment policies: (1) Hiring (both at new and existing stores), Initial Job Assignments, and Training; (2) PAR Movement and Cross-Training; (3) Compensation; (4) Promotion to Management; and (5) Equal Employment Opportunity. The contents of these Appendices are summarized below.

a. Hiring/Initial Job Assignment/Training (1) Hiring at New Store Openings

From January 1995 to December 2000, the New Unit Openings Department (hereafter "NUO") supervised new store openings. NUO first contacted an advertising agency before the store's opening date to obtain information about the labor market and to buy advertising to reach potential employees. Maidl Decl. ¶¶ 5-6 (part of defendant's Hiring and Initial Job Assignment Appendix). The new store's District Manager also conducted a study of the relevant labor market. NUO used the labor study to determine the wages to use in the advertising. Id. ¶¶ 7-8.

Before mass hiring began, the NUO's Opening Supervisors (restaurant and retail) spent two to four hours reviewing Company policy with store management to ensure proper questioning and screening of applicants. NUO used a manual entitled "Interviewing and Hiring Guidelines" (Ex. H-1). Maidl Decl. ¶ 9. Three weeks before opening, the interviewing and hiring process began with scheduling appointments to take applications and conduct interviews. Id. ¶ 10. When an applicant arrived, he filled out an application and underwent a brief pre-screening interview.Id. During this interview, an Associate Manager asked preliminary questions to determine if an applicant would receive a second interview. These questions came from the "Interviewing and Hiring Guidelines" handbook's segment entitled "New Store Opening/Mass Hiring Recommended Job Interview Format." Questions focused on the applicant's availability, scheduling, permission to contact previous employers, qualifications for the position sought, and behavior-based questions.Id. ¶¶ 11, 13; Ex. H-1 at 1-2, 7-9, 11-13. Associate Managers could ask their own questions with prior General Manager approval. Maidl Decl. ¶ 13.

Managers also reviewed the applications to see if applicants sought specific positions. Maidl Decl. ¶ 12. During new store openings (whether currently or in the past), Cracker Barrel policy is to pre-qualify an applicant for the position indicated on the application, with two exceptions. Id. First, if an applicant seeks a position that is already filled, and a manager determines after an interview that the candidate is excellent, the manager is to explain that the position sought is filled, and ask the applicant if he is interested in any other position. Id. Second, if an applicant seeks a cook position, the manager is to determine whether he is qualified for any cook position (i.e., prep cook, back-up cook, or grill cook). Id.

The information obtained during the pre-screening interview was used to either disqualify the applicant or schedule him for a second interview with the General Manager or Retail Manager. Maidl Decl. ¶ 15; Ex. H-1 at 9. During the second interview, the General Manager and Retail Manager focused on behavior-based questions. Like the Associate Managers, General Managers and Retail Managers were to use the "Interview and Hiring Guidelines" handbook as a guide on appropriate questions to ask. Id. ¶ 17; Ex. H-1 at 2-6, 8, 35-37.

Beginning in January 2001 and continuing to the present, the advertising process for new store hiring is the same as described above, except responsibility has been shifted from NUO to the Employment Department. Maidl Decl. ¶ 19; Ex. H-2 at 9-11. Managers of the new store come to the Home Office five weeks before store opening for New Store Meetings, where Field Human Resources gives a one-day presentation on staffing and retention and instructs Managers on interviewing and hiring practices using the "Unit Opening-Manager's Handbook" (filed as Ex. H-2, CBR 24020-36). Maidl Decl. ¶ 21. This Handbook is taught in conjunction with the document entitled "Best Practices-Hitting the Target Through Staffing and Retention." Id.; Ex. H-3.

Associate Managers now use the "Seven Steps to Effective Interviewing," when questioning the applicant using the format outlined in "Best Practices-Hitting the Target Through Staffing and Retention." Maidl Decl. ¶ 24; Ex. H-2, CBR 24028-29; Ex. H-3 at 28-34. Associate Managers also review the applications, ask skill questions specific to the position sought, ask behavior-based questions, and explain pay rates. Maidl Decl. ¶¶ 24-25; Ex. H-3 at 29-31. If an applicant is chosen for a second interview, he is then asked to read Cracker Barrel's Equal Opportunity Statement. Maidl Decl. ¶ 26.

During the second interview, the General Manager and the Retail Manager ask behavior-based questions. Maidl Decl. ¶ 27. Like the Associate Managers, General Mangers and Retail Managers are to use the "Best Practices-Hitting the Target through Staffing and Retention" as a guide for appropriate pre-employment inquiries. Id. ¶¶ 27-28; Ex. H-3 at 30-31, 63, 69-70. General Managers and Retail Managers at the local stores make the final hiring decision, which is to be made on the basis of the applicant's ability to perform the essential functions of the position, experience, availability, attitude, and communication skills. Maidl Decl. ¶¶ 18, 29.

(2) Hiring at Existing Stores

In 1993, the Company's Management and Development Department began a training program concerning interviewing and hiring practices entitled "Bringing People In." Pate Decl. ¶ 5; Ex. H-S. This program provided managers with the "Interviewing and Hiring Guidelines," which contained interviewing guides, rules of the interview, sample interview questions, and information regarding appropriate pre-employment inquiries. Pate Decl. ¶¶ 6-8; Ex. H-i.

"Bringing People In" training was administered first to District Managers in the Company's Home Office. Pate Decl. ¶ 9. The District Mangers then presented the material to their store General Managers, who in turn trained their Associate Managers. Id. From 1994 to 1999, during their training at the Home Office, Managers in Training (hereafter "MITs") were also given a day of instruction on the "Bringing People In" and "Interviewing and Hiring Guidelines" materials. Id. ¶ 11.

From 1999 through 2000, the Management and Development Department rolled out anew guidebook entitled "Staffing and Retention-Resource Guide." Pate Decl. ¶ 13. The Guide contained information about creating a staffing plan, advertising, interviewing, hiring, and retaining employees. Id. Sample questions were provided, but managers were encouraged to customize their questions to match their stores' specific needs. Id. ¶ 15. The materials also instructed managers on how to conduct interviews, how to make a good hiring decision, and on the corporate values of diversity and compliance with the law. Id. ¶ 16. The Company forwarded the Guide to all General Managers and Associate Managers. Id. ¶ 17. From 1999 through the beginning of 2001, MITs received a day of training at the Home Office on the Guide. Id. ¶ 18.

Beginning in 2001, the Company began training managers using the "Best Practices-Hitting the Target through Staffing and Retention" handbook (discussed supra). Pate Decl. ¶ 19; Ex. H-3. This handbook contains information on creating a staffing plan, handling applicant flow, advertising, interviewing, retention, and training. Pate Decl. ¶ 19. The material was first presented to the Regional Vice Presidents, with further training administered training down the chain of command. Id. ¶ 24. From 2001 to the present, MITs have received a day of training at the Home Office on the "Best Practices" materials (which is revised periodically). Id. ¶¶ 25-26. Among other things, the "Best Practices" materials teach that it is unlawful to discriminate on the basis of race. Id. ¶ 28.

During the time period at issue in this case, Company policy has been that the decision to hire is based on objective and subjective factors regarding the applicant: (1) responses to questions about his availability; (2) ability to perform the essential functions of the job; (3) prior work experience; and (4) overall "presentation." Pate Decl. ¶ 29. Decisions to hire for an hourly store position have always been made by store-level management, either by an Associate Manager, the store General Manager, or the Retail Manager. Id. ¶ 30.

(3) Initial Job Assignments

When a store needs to fill hourly positions, managers typically advertise those employment opportunities. Pate Decl. ¶ 36. When an applicant fills out an application, he is directed to indicate on the application the position(s) for which he is applying. Id. A manager then interviews an applicant for one or more of the positions indicated. Id. If the applicant is qualified for the job sought, and if the job is available, it is presumably offered to him. Id.

If more than one position is available, Cracker Barrel instructs its managers to ask introductory questions, gather information, explain the positions that are available, and ask the applicant to read the essential functions of each to determine the position(s) in which the applicant is interested. Pate Decl. ¶ 37. The manager is also to determine if the applicant is minimally qualified on the basis of his responses to questions about availability, the scheduling needs of the store, and the applicant's prior experience. Id. If the applicant is interested in the available position(s) and the availability needs of the particular store can be met, then the manager will interview the applicant "on the spot" or schedule a future interview. Id. If a store is not looking to fill any particular position, but an applicant desires to complete an application, a manager is to explain the store's lack of staffing needs, but accept the application and keep it for possible future use. Id.

(4) Training Programs

From 1995 to 1996, the Company's Employee Relations Department trained from the "Management Trainee Guidebook: Issues in Employment Law" during MIT training at the Home Office. Pate Decl. ¶ 31; Ex. H-7. From 1997 to 1998, the Employee Relations Department trained using handouts from that Guidebook, including the importance of hiring the most qualified candidate for the position. Pate Decl. ¶ 31. Employee Relations Specialists purportedly discussed anti-discrimination laws and Cracker Barrel's anti-discrimination policies and stressed the importance of not considering race in making employment decisions. Id. During 1999, the Employee Relations Department trained using the "Leader's Guide-Human Resources Policy Review (also known as the "Patchwork Quilt"). Id. ¶ 32; Ex. H-8. This guide was sent to the stores and included the same anti-discrimination policies mentioned above. Id. From 2000 to the present, the Employee Relations Department trains MITs using the "Management Reference Guide." Pate Decl. ¶ 33; Ex. H-9. This Guide includes the Company's anti-discrimination policies and stresses that employment decisions should be made without regard to race. Id.

b. PAR Movement and Cross-Training (1) PAR Program

There are five PAR levels for each of the nine hourly store positions. The lowest level is 0 and the highest is IV.⁶ Stinson Decl. ¶ 7. An employee can progress from one PAR level to the next by working a specified time in his existing PAR level, receiving a passing score on his evaluation (the passing score increases at each PAR level), and passing a written test. Id. An employee receives a pay increase and enhanced benefits when he advances in PAR level.⁷ Id. Each PAR level for each position has a maximum pay rate. Id. ¶ 8. To exceed the maximum pay rate for a position, an employee must either advance to the next PAR level for that position or receive corporate management approval. Id. Movement within PAR is not a promotion; however, by advancing from one PAR level to the next, an employee acquires a level of achievement that evidences an increase in knowledge and skill within a position. Id. ¶ 9. Store management does not have the discretion to refuse or prevent an employee from participation in PAR. Id. ¶ 10. If that occurred, it would be against Company policy. Id. ¶ 10 n. 3.

The PAR Department has used the following documents in administration of the program: (1) Pleasing People Through PAR; (2) the PAR Program Manual; and (3) various PAR Manuals related to each hourly position. Stinson Decl. ¶ 11. Part of the Pleasing People Through PAR Manual is the "Introduction to PAR," which is given to new employees at the new-hire orientation and describes the program and requirements for each PAR level. Id.

The Pleasing People Through PAR Manual (excerpts filed as Ex. PAR-1) includes training information and materials used by the PAR department, store management, Employment Training Specialists (hereafter "ETSs"), and Employment Training Coordinators (hereafter "ETCs") in administering the PAR program. Stinson Decl. ¶ 11(a). The PAR Program Manual includes information about testing schedules, testing procedures and logistics, as well as current versions of the actual tests. Id. at ¶ 11(b); Ex. PAR-2. The PAR Manuals (samples found at Ex. PAR-3) include information about the skills and knowledge required for each of the nine hourly store positions. Each position has its own PAR manual. Stinson Decl. ¶ 11(c).

From 1994 until 1996, the corporate PAR Department administered the PAR program at the store level through the use of corporate PAR representatives, also called "PAR testers." Stinson Decl. ¶ 12. In 1996, the PAR testers were phased out as the Company began utilizing ETSs. Id. ¶¶ 13-14. The Company's intent in assigning an ETS to work with each store was to ensure continued corporate office involvement in and supervision of PAR at the store level, as well as to improve PAR training. Id. ¶ 14. This corporate involvement precluded local management from exercising discretion regarding testing and other administrative tasks. Id.

In late 1998 and early 1999, Cracker Barrel began implementing a system in which the ETC, who is an hourly employee of the store, provided support for the administration of PAR. Stinson Decl. ¶ 15. The ETC receives training at the corporate headquarters on the PAR program and store administration of PAR. Presently, there are about 250 ETCs in the field. Id.

Employees study PAR training materials during study sessions held in the store while "on the clock." Stinson Decl. ¶ 16. Home study of PAR materials is prohibited. Id. Throughout its existence, the PAR program has undergone continual improvement through updated materials and different methods of training. See, e.g., Ex. PAR-S. Most recently, the Company has introduced computer-assisted PAR instruction and training (also called "e Learning"). Stinson Decl. ¶ 17.

With regard to PAR level movement, through 1999, all employees were required to test for PAR level 1.⁸ Stinson Decl. ¶ 19. After passing PAR level I, an employee was not required to participate in PAR any further. Id. The PAR I test was considered a minimum-knowledge-based test. Id. Employees who chose not to participate further were still required to maintain a minimum level of knowledge and skill, which were measured by evaluations every six months. Id.

8 The PAR I test is still considered a minimum-knowledge-based test, but those employees who choose not to participate in PAR are still required to demonstrate a minimum level of knowledge by completing courses in e Learning. Stinson Decl. ¶ 24.

After a set, continuous period at a PAR level, employees then had options regarding PAR level movement and/or cross-training. The following chart summarizes PAR requirements and rewards for advancing in PAR level: PAR Requirement to Benefits of that Min. Time Options Level Reach that Level Level Req'd in PAR Level

0 30days • Test for PAR I [After — — 1999] I • Min. score of • Pay increase 3 months • Test for PAR II in same 80% on • Gold star position performance • 20% discount • pre-1999: Cross-train evaluation card for retail into another position at • Pass skill test store or PAR I [post-1999, restaurant must reach PAR II to cross-train] II • Min. score of • Pay increase 4 months • Test for PAR III in same 85% on • Gold star position performance • Certificate of • Cross-train into another evaluation achievement position at PAR I [preMm. • Min. score of • Eligibility for 1999]/ PAR 0 85% on skill decrease in [post-1999] test insurance • Test for higher PAR premiums level in other position already cross-trained to perform III • Min. score of • Pay increase 7 months • Test for PAR IV in same 90% on • Gold star position performance • Certificate of • Cross-train into another evaluation achievement position at PAR I [preMm. • Min. score of • Eligibility for 1999] / PAR 0 90% on skill decrease in [post-1999] test insurance • Test for higher PAR premiums level in other position already cross-trained to perform IV • Min. score of • Pay increase None, but — 95% on • Gold star must score performance • Certificate of 95% on evaluation achievement evaluation • Min. score of • Eligibility for every 6 95% on skill decrease in mos. to test insurance maintain premiums PAR IV • Yearly stock options See Stinson Decl. ¶¶ 20-24, Exs. PAR-6 through PAR-9.

(2) Cross-Training

Cracker Barrel views cross-training as an opportunity to help each store run more efficiently by employing a staff of versatile and flexible employees. Stinson Decl. ¶ 26. An employee may cross-train into any position as long as he meets the minimum time requirements in his current primary position. Id. In considering when to cross-train, the employee and local management consider the objective business needs of the unit, such as staffing needs and availability of staffing hours, in conjunction with the flexibility, skill level, and availability of the employee. Id.

Cracker Barrel does not require employees to have prior work experience or training in a front-of-the-house position to cross-train or test for front-of-the-house PAR levels. Stinson Decl. ¶ 27. Company policy has always provided that an employee may cross-train in' any skill position.Id. ¶ 28. The only prerequisites have been the attainment of PAR I (pre-1999) or PAR II (post — 1999) in their primary skill level.Id.; see Ex. PAR-4, July 26, 1999, memorandum announcing change. For example, after 1999, a dishwasher could cross-train to hostess, server, grill, or any other position in the store as soon as he became a PAR II dishwasher. Stinson Decl. ¶ 28.

However, before testing in the cross-trained position of interest, an employee should work in that position long enough to attain a working knowledge of the position sufficient to enable him to pass the evaluation. Stinson Decl. ¶ 29. The amount of time necessary to develop that level of knowledge varies, depending upon how often the staffing needs of the store permit the employee to work in the cross-training position, the availability of the employee, and the learning ability of the employee. Id. To be fully cross-trained in a position, an employee must pass the evaluation and skill test for that position. Id.

Historically, an employee received the highest of his current rate of pay for his primary skill or the minimum of the PAR 0 wage range for the position for which he was cross-training, whichever was higher. Stinson Decl. ¶ 30. Once the employee passed the corresponding PAR test for the position for which he was training, he earned wages for the corresponding position and PAR level. Id. Thus, it was possible for an employee to earn different wages in the different skill areas depending upon the position in which the employee was working Id.; see also Jones Decl., filed in Defendant's Compensation Appendix, ¶ 33.⁹

9 For example, if a PAR III Dishwasher making $5.75 per hour was cross-training to become a grill cook, he was paid $5.75 per hour during training because the $5.75 wage rate was higher than the $5.15-per-hour rate earned by a PAR 0 grill cook. Once the employee passed the corresponding PAR level (in this example a PAR III grill cook), the employee earned wages in the PAR III grill cook range whenever he worked that position. Stinson Decl. ¶ 31; Jones Decl. ¶ 34.

However, effective November 2001, an employee cross-training into a new position earns the rate applicable for that position at PAR 0, as if newly hired into the position, as opposed to higher of his primary job or that of the cross-trained position. Stinson Decl. ¶ 32. For instance, if a PAR IV hostess cross-trains as a cashier, she will earn whatever the pay rate is for the PAR 0 cashier. Id. Upon reaching PAR I status, the employee then receives the pay rate for PAR I cashier whenever she works in that position. Id.; Jones Decl. ¶ 35.

c. Compensation of Hourly Employees

Each Cracker Barrel store differs with respect to its own local labor market, business needs, and the employment opportunities available. Jones Decl. ¶ ¶. Recognizing these local business situations and individual store needs, Cracker Barrel examines local labor markets before a store opens and creates a labor market study to determine the wage levels that will be sufficiently competitive to allow the new store to hire the necessary employees. Id. ¹⁰ Cracker Barrel also uses its labor market studies to develop pay grids, which it then uses to establish wage ranges for each of the following hourly positions: (a) night maintenance; (b) dishwasher; (c) back-up cook; (d) prep cook; (e) grill cook; (f) cashier; (g) host/hostess; (h) gift shop; and (i) server. Jones Decl. ¶ 7. A store's pay grid specifies the minimum and maximum wage rates allowed for employees in each job and PAR level. Id. ¶ 12. A store's pay grid may change as local market conditions change. Id. ¶ 11.

10 An example of a local labor market study is found in the Compensation Appendix at Ex. C-1. The information in a study falls into four categories: (a) a description of the store's location; (b) competition in the local labor market; (c) sources of labor; and (d) local wage rates. Jones Decl. ¶¶ 8-9. See also H/T Report at 11-13. A store's District Manager reviews the study, recommends a pay grid, and forwards a copy of the study and his recommendation to the Regional Vice President for approval. Jones Decl. ¶ 10.

Cracker Barrel has used three different pay grid systems since 1995. Jones Decl. ¶ 13. Before the fall of 1995, each store was assigned to one of four general Areas (or grids) referred to as Areas 1, 2, 3, and 4. Id. In addition, Cracker Barrel developed supplemental variations in response to local market conditions. Id. (Copies of these pay grids are found at Ex. C-2.)

Under the pay grid system established in 1995, Cracker Barrel developed new pay grids based upon state and local wage laws for tipped positions (server) and the competitive wages required for the highest paying position at Cracker Barrel (grill cook). Jones Decl. ¶ 14. The Company created nine pay scales for servers and twelve pay grids for grill cooks. Id. In total, Cracker Barrel created thirty-one different basic pay grids. Id. The Company created even more grids when it added stores in new states or laws changed in states where it already had stores. Id. ¹¹

11 Under this 1995 system, all stores assigned to a given general pay grid had identical wage ranges for the server and grill cook positions. In contrast, the wage ranges for the other hourly positions varied from store to store, even for stores within the same general pay grid, so that individual stores could remain competitive in their local labor markets for other positions. For example, store no. 42 in Columbia, South Carolina, and store no. 70 in Pelham, Alabama, were both assigned to the general wage grid "GE." Both stores had identical wage ranges for the server and grill cook positions. However, their wage ranges for dishwashers varied. Store no. 42 had a wage range for PAR 0 dishwashers of between $4.30 and $5.25 per hour, while store no. 70 had a wage range for PAR 0 dishwashers of between $4.30 and $5.50 per hour. Jones Decl. ¶ 15; Ex. C-3 (1995 pay grids); see also H/T Report at 24-27.

In August 1998, Cracker Barrel changed its pay grid system to the one now in use. Initially, there were five general grids (lettered A-E). Jones Decl. ¶ 17. These grids listed wage ranges for all hourly jobs except server. Id. As under the prior system, server wages reflected state minimum wage laws for tipped employees, as well as local labor market conditions. Id. When the server wage ranges were combined with the five basic pay grids for the other hourly jobs, there were twenty different pay grids. Id. The Company subsequently added a sixth pay grid for Connecticut (grid f), which was also applied to stores opened in Rhode Island. Id.; Ex. C-4 (1998 pay grids).

Once a store is assigned to a pay grid, the actual wage paid to an individual store employee is determined by the store's General Manager, provided that the wage is within the range set by the pay grid for the employee's position and PAR level. Jones Decl. ¶ 18. However, highly experienced or skilled individuals may earn above the maximum for a given PAR level with the approval of the District Manager and Regional Vice President. Id. ¹² The wage rate for any particular hourly employee depends upon a number of factors, including the following: (1) local labor conditions; (2) the stores's pay grid; (3) the employee's experience level; (4) the employee's position; (5) the employee's PAR level; (6) whether the store General Manager requested a variance for the employee's position and PAR level; and (7) whether the requested variance was approved or denied. Id. ¶ 21.

12 In response to changing labor market conditions, a manager can request a change in his store's pay grid or adjust the wage range within a pay grid. When market pressure is particularly intense for a specific job, store managers may make use of "ad hoc market adjustments" for paying employees outside the maximum allowable range for a particular job and PAR level for their store's pay grid. H/T Report at 30. Defendant's experts examined, by PAR level and job, the wage rates of employees working in November 2000 compared to the maximum allowable wages of the March 1999 wage grid assigned to their store. Approximately 27% of the employees were off the grid. In 10% of the stores, over 50% of the employees were off the March 1999 grid, and in 36.5% of the stores, over 30% were off it. Id.

Back-of-the-house positions generally have a higher wage rate range than front-of-the-house positions. Jones Decl. ¶ 22. For example, grill cook has the highest hourly wage rate of any job; dishwashers generally have a higher wage rate range than cashiers and hostesses.Id.; Exs. C-3, CA. This lay testimony by Mr. Jones is supported by expert analysis. See H/T Report at 27-28 Fig. 2.

To achieve the objectives of Cracker Barrel's salary administration policy,¹³ the Company established salary administration guidelines for all hourly positions. Jones Decl. ¶ 25. Store management must adhere to these guidelines or face disciplinary action. Id. These guidelines establish, inter alia, the minimum length of employment required for an employee to be eligible to take a PAR test; the required passing score for each PAR test; the percentage wage increase for passing a PAR test; and the percentage wage increase for merit raises. Id. ¶¶ 25-26. (A copy of Cracker Barrel's Field Hourly Salary Administration Guidelines, effective August 1, 1998, is found at Ex. C-5 of the Compensation Appendix; the 1999 Guidelines are found at Ex. C-6.)

13 The salary administration policy's objectives are to achieve the following: (a) pay salaries that fairly reflect the value of each job relative to other jobs in the Company; (b) recognize individual contributions and learned skills through participation in the PAR program; (c) pay salaries that are competitive with others in the local marketplace; (d) provide capability to adapt to changing market conditions; and (e) ensure that the policy conforms to the law. Jones Decl. ¶¶ 23-24.

PAR testing is completed bi-monthly for all units in a region, and tests are given one week of every other month. Jones Decl. ¶ 27. The standard hourly rate increase for employees passing a PAR test before March 1999 was up to three percent, approved by the store General Manager, and up to four percent after March 1999. Id. ¶¶ 28, 33. With District Manager approval, a deserving employee could receive a five-percent increase. Id. ¶ 28. Any increase that would result in the employee being paid above the maximum for his PAR level required the advance approval of the Regional Vice President. Id.

Employees are eligible for merit raises based on performance every six months from the date of their last increase (which includes PAR increases). Jones Decl. ¶¶ 29-31. Under the Guidelines effective August 1, 1998, a store General Manager could give a merit increase of up to three percent. id. ¶ 32. (This increased to four percent in March 1999.) Advance approval from the District Manager was required to give an increase between 3.1 and 5.0%. Id. Advance approval of the District Manager and Regional Vice President was required for (1) increases that exceeded 5.0%; (2) increases that were more frequent than six months from the last increase; and (3) increases that resulted in the employee being paid an hourly rate that exceeded the maximum for his PAR level. Id.

The Company has separate salary administration guidelines for the server position. Servers are compensated at a set wage per PAR level and receive a fixed dollar increase as they attain the next PAR level. Jones Decl. ¶¶ 36-37. There is no exception to these rules without prior written approval of both the District Manager and Regional Vice President. Id. Servers become eligible for merit increases six months after attaining PAR IV. Id. ¶ 38. Before March 1999, merit increases were limited to a maximum of three percent with General Manager approval and five percent with District Manager approval. Id. Anything higher than five percent required approval from the Regional Vice President. Id. Merit increases given more frequently than six months or before attaining PAR IV also required the advance approval of the Regional Vice President. Id. Effective March 1999, merit increases were limited to a maximum of four percent with store General Manager approval and five percent with District Manager approval. Id. Anything higher than five percent still requires Regional Vice President approval. Id.

d. Promotion to Management

In 1994, the Company created its "Management Internship Program." Pate Decl. ¶ 5 (filed as part of Promotion to Management Appendix). This program seeks to identify hourly store employees who have the potential to move into restaurant or retail management, and to provide those employees with extensive training for such a promotion. Id. Since its inception, the program has consisted of four phases: (1) selection; (2) in-store training; (3) management training; and (4) reassignment to a store.Id. ¹⁴

There are several prerequisites to the Internship Program. Pate Decl. ¶ 6a-c. First, a candidate must have a high school diploma or its equivalent. Id. Second, a candidate must have a certain level of work experience at Cracker Barrel. Id. At the beginning of the program, candidates were required to meet each of the following work experience requirements:

1. Must have been a PAR IV employee in any area;

2. Must have held the Lead Trainer, New Store Opener, or Merchandise Assistant position for a minimum of six months;

3. Must have been a Cashier PAR Level II or higher;

4. Must have been a PAR level II or higher backup cook or grill cook; and

5. Must have been a PAR level II or higher server or host.

Id. ¶ 6c.¹⁵ Third, a candidate must take the initiative to complete an application.¹⁶ See Ex. M-3. Finally, although Cracker Barrel policy does not mention this requirement, several plaintiffs testified infra that a candidate must be willing to relocate.

15 Over the years, the positions that satisfy the experience prerequisite have changed. The positions of POS Technical Trainer, Employee Training Coordinator ("ETC"), Skill Trainer (in stores with an ETC), and Shift Leader (if the applicant worked at least 16 hours per week in that position) were added to the list of non-PAR-job prerequisites, all of which an employee must have held for six months to qualify for the program. Pate Decl. ¶ 6c.

16 Initially, the application had to be signed by the General Manager of the applicant's store before submission to corporate headquarters. In September 1998, that requirement was modified to require local management's approval after the application was received by corporate headquarters. If the General Manager of the applicant's store does not recommend him, the reason is documented. Pate Decl. ¶ 6a.

If an applicant met the minimum requirements, his name would be forwarded to the Regional Management Recruiter ("RMR") for the region in which the candidate was employed. Pate Decl. ¶ 7. The RMR would contact the candidate and schedule an assessment. Id. ¶ 8. The assessments varied over time, but each one was developed by an outside professional consultant who determined that it was a valid, job-related selection device. Id. If a candidate successfully completed the assessment, he was then interviewed by the RMR. Id. ¶ 9. The Company provided RMRs with a structured interview guide that identified the questions to be asked during the interview. Id. The RMR made the decision whether to accept the candidate into the Internship Program. Id. ¶ 10.

In 1998, District Managers assumed this responsibility from the RMRs. Like the RMRs, District Managers followed an "action checklist" and adhered to structured interview questions. Pate Decl. ¶ 11. Before finalizing a decision, a District Manager consulted with his Regional Vice President. Id. On March 15, 1999, the responsibility for interviewing candidates returned to the Recruiting Department. Id. ¶ 12. However, the decision whether to accept an applicant into the Internship Program became the responsibility of the Regional Vice President. Id. The structured interview format also changed in 1999 to include questions in additional areas. Id. ¶ 13; see also Ex. M-5 of Promotion to Management Appendix. In the vast majority of cases, candidates have been selected if they satisfied the minimum requirements, passed the assessment, and passed a mandatory drug screen. Pate Decl. ¶ 14. In only a few instances has a candidate been rejected because of a poor interview. Id.

The Company also publishes the "Internship Program Handbook" to provide information to persons selected for the program as well as those interested. Pate Decl. ¶ 15; Ex. M-2 (current handbook). Since September 1998, the Company has also informed employees of the Management Internship Program through posters in store break areas. Pate Decl. ¶ 16; Ex. M-4 of Promotion to Management Appendix.

e. Equal Employment Opportunity Policies

Since 1995, the Company has promulgated the following written policies:

Equal Opportunity Statement (Ex. E-1)

Harassment and Discrimination Policy (Ex. E-2)

Rules of Conduct (Ex. E-3)

Open Door Policy (Ex. E-4)

Pleasing People Mission Statement (Ex. E-5)

Patchwork Quilt Policy (Ex. E-6)

Tramel Decl. ¶¶ 4-9 (filed in Defendant's EEO Appendix). Cracker Barrel's official policy during the relevant time period has been that all hiring decisions are governed by the Company's Equal Opportunity Statement, which requires that all decisions be made without regard to race. Pate Decl. ¶ 12; Maidl Decl. ¶ 28.

The Company has made its employees aware of these policies through their publication in the Employee Handbook and through new-hire orientation sessions. Tramel Decl. ¶¶ 10-11. The Company provides its Pleasing People Mission Statement to employees upon hire and in PAR manuals, and disseminated this Mission Statement to all employees in September 2001. Id. ¶ 12. EEO posters are displayed in store break areas. These posters include a toll-free number for complaints that is ultimately directed to Employee Relations. Id. ¶ 13.¹⁷ The Patchwork Quilt Policy poster is displayed in all store break areas. All employees received a pamphlet about the Patchwork Quilt Policy with their paychecks. All new hires watch a video about the Patchwork Quilt during orientation. Id. ¶ 14.

17 Employees can report discrimination or harassment by complaining personally to any level of management or to an Employee Relations representative by using the toll-free number or e-mail. Complaints may be made anonymously. Tramel Decl. ¶ 30. Although management training includes information on investigating such complaints, managers are asked to report them to Employee Relations, who can receive the complaint, assist with the investigation, maintain records, and ensure that responsive action is taken when necessary. Id. ¶ 32. Employee Relations uses standard forms for reporting incidents and obtaining witness statements and has a checklist for conducting a proper investigation. Id.

In November 1999, the Company began using payroll scrolls on paychecks to further disseminate the toll-free number to Employee Relations and the Open Door Policy. Tramel Decl. ¶ 15. Beginning in January 2000, the Company printed the toll-free number on all Employee Discount Cards. Id. Further, the Company's EEO, non-discrimination, and non-harassment policies are addressed in the Company's newsletter, which is distributed to all stores and placed in break areas. Id. Cracker Barrel also posts its EEO policy on its website and employment application. Id. ¶ 16. In February 2001, the Company added a paragraph to its application form directing applicants/employees to report concerns about discrimination or harassment to Employee Relations. Id. Lead Trainers or ETCs also provide training to store employees on EEO policies after orientation. Id. ¶ 17.

In October 2000, the Company added equal employment opportunity to management performance reviews, requiring verification of the manager's understanding of, and compliance with, Cracker Barrel's EEO policies. Tramel Decl. ¶ 18. In November 2001, the Company added equal employment opportunity initiatives to the Managers' Staffing and Retention Guide. Id. Since at least 1999, management trainees have received information about equal employment and anti-harassment policies in the Employee Relations Guidebook, their Orientation Manual, the Internship Handbook, and the Interview Guide. Id. ¹⁸ In May 2000, Cracker Barrel created a toll-free number to Employee Relations for the exclusive use of managers. Id. ¶ 19. The Company has periodically modified the above-listed policies and procedures to improve them and reflect changes in the law. Id. ¶ 20.

18 Since at least 1995, all management trainees have been required to attend a seminar on EEO/Title VII issues during their initial training at corporate headquarters. Tramel Decl. ¶ 21. Beginning in 1997 or 1998, management trainees were required to participate in an interactive CD ROM training program, which includes EEO issues. In early 2000, Employee Relations began using the Patchwork Quilt training program for management trainees. In September 2000, the Company created the Managers Reference Guide (which incorporates the Patchwork Quilt) and began using it for training on EEO issues with management trainees. id. ¶ 22. Patchwork Quilt training has also been provided to District Managers, who, in turn, were required to train all management employees in their District. Id. ¶ 23. District Managers were assigned the CD ROM training program in August 2001. Id. ¶ 27.

As Associate Managers are promoted, they receive additional Patchwork Quilt training through the Senior Associate Manager program. Tramel Decl. ¶ 24. Since at least 1995, Senior Associate Managers have attended a one-week training course at corporate headquarters before promotion to store General Manager. During this week, they receive 1.5 hours of training on EBO issues. Id. All General Managers who have been promoted to District Manager receive a corporate orientation that includes another hour of training with employee relations to review Patchwork Quilt materials. Id. ¶ 25. Since 1999, Patchwork Quilt materials have been presented at the annual General Managers meeting, which is attended by all General Managers, District Managers, and Regional Vice Presidents. Id. ¶ 26.

In the fall of 2001, Cracker Barrel continued employee relations training for new-unit-opening employees using a format similar to the management training program. Tramel Decl. ¶ 24. ETCs also attended Employee Relations training beginning in the fall of 2001 and continuing, which included information on EEO issues. Id. ¶ 29. In addition to the formal training mentioned above, each level of management provides training on EBO issues on an as-needed basis to their subordinates. Id. The corporate Employee Relations Department is also available to assist at any time. Id. B. The Putative Class Representatives

The Court draws the following summary of each putative representative's claims from the Amended Complaint, his/her Declaration, and his/her deposition transcript.

1. Kelvis Rhodes (current employee)

Mr. Rhodes alleges that he worked for Cracker Barrel at its League City, Texas, location from July 1994 through September 1998 as a reserve grill cook. From October 1998 to the present, Mr. Rhodes has been employed at Cracker Barrel's Dalton, Georgia, location as a grill cook. Am. Compl. [92] ¶¶ 11, 38.

In his January 3, 2002, Declaration, Mr. Rhodes states that, when he applied for work at the League City store in 1994, he sought a position as either a grill cook (first choice), back-up cook, or dishwasher. Rhodes Decl. at 1; see also Rhodes Dep. at 23-24. (All are back-of-the-house jobs.) However, he was told that the only position available was dishwasher, which he accepted. Rhodes Decl. at 1; Rhodes Dep. at 27. When he moved to Dalton in 1998, Mr. Rhodes sought either the grill cook or back-up cook positions. Rhodes Dep. at 55-57. Thus, Cracker Barrel could not have channeled Mr. Rhodes to back-of-the-house jobs because he sought only those type of jobs.¹⁹

19 Mr. Rhodes rated the Company's best jobs in the following order: grill cook, back-up cook, prep cook, server, and dishwasher. Rhodes Dep. at 53-54. Only one of those jobs — server — is in the front of the house.

Mr. Rhodes also claims that he informed his supervisors of his interest in being promoted to positions with better opportunities for advancement, but he never received any instruction or encouragement regarding applying for the Management Training Program or cross-training for front-of-the-house positions. Am. Compl. ¶ 39.²⁰ In his deposition, Mr. Rhodes testified that he talked to a manager named Micah about a management position. Rhodes Dep. at 59; Rhodes Decl. at 3. He claims that Micah responded that it would be best if he went back to school. Id. He claims this was discriminatory because some White managers have only high school diplomas. Rhodes Decl. at 3.

20 It is apparent, however, that Mr. Rhodes was cross-trained in back-of-the-house jobs and that he had no real interest in front-of-the-house jobs. Also, he must have meant to allege lack of information about the Management Internship Program, as that was the option available to current employees.

However, upon further examination of the deposition, Mr. Rhodes admitted that he was not interested in relocating (which was a requirement to be considered for management), and that, although Micah did tell him to go back to school, he could not remember if Micah's comment had anything to do with promotion into management. Rhodes Dep. at 59-61.²¹ Thus, Mr. Rhodes could not have been denied a management position because he never sought one.

21 In his Declaration, plaintiff asserts that an unidentified manager (perhaps Micah) told him to go back to school after plaintiff asked about getting a more highly paying job. No mention is made of promotion to management. Rhodes Decl. at 3.

Plaintiff also alleges that, when he bypassed the PAR II and III tests and passed the PAR IV test on back-of-the-house jobs, he was not given the standard pay increase. Am. Compl. ¶ 39. Mr. Rhodes further contends that he was denied the opportunity to take a PAR test for back-up cook.²² Mr. Rhodes testified that, whenever the list of names of those eligible to take this PAR test was posted on the door, his name was not included. Rhodes Dep. at 90. He added that on one occasion some time ago he mentioned to an unidentified manager that his name was not on the list. Id. at 90. However, he did not recall the manager's response, and he had not, in the year preceding his deposition, expressed his desire to take the PAR test for back-up cook (whether level III or IV) to anyone in management. Id. at 91. He feels that he should not have to do so, because everyone else who is eligible for a test is listed. Id. at 91.

22 The record is conflicting on this point. In his deposition, Mr. Rhodes claims that he was denied the opportunity to take the PAR III test. Rhodes Dep. at 89-90. In his Declaration, he claimed denial of the PAR IV test. Rhodes Decl. at 4. However, Mr. Rhodes's PAR testing history, discussed infra, shows that he took three PAR tests for back-up cook and is at PAR IV in that job now.

Finally, Mr. Rhodes asserts that the Company has routinely promoted White employees to vacant positions over African-American employees of comparable or more extensive experience. Rhodes Dep. at 91. He claims that Ms. Serena McDermott, an African American, has requested promotion to cashier, but her store's management has always responded that no cashier positions were available. Am. Compl. ¶ 39.

Mr. Rhodes has the following PAR testing history: Date of Test Skill Par Level of Test Results

10/12/94 Dishwasher 1 Pass 01/17/95 Dishwasher 2 Pass 06/06/95 Dishwasher 3 Pass 01/25/96 Dishwasher 4 Fail 03/29/96 Back-up Cook 1 Pass 05/23/96 Dishwasher 4 Pass 09/13/96 Back-up Cook 2 Pass 01/15/97 Grill Cook 1 Pass 11/12/98 Grill Cook 4 Pass 08/01/99 Back-up Cook 4 Pass H/T Report at 75. This plaintiff had the opportunity to cross-train in other jobs and take PAR tests.

The hourly rate paid to Mr. Rhodes increased from $6.04 in 1995 to $12.53 in 2001, an increase of 108%. H/T Report at 101. In 1995, his rate was one cent less than the other PAR III dishwasher, who was non-African American, and higher than any other employee who worked the majority of his hours as a dishwasher. Id. The majority of Mr. Rhodes's hours in 1996 was as a grill cook, a job for which he had yet to pass a PAR test. Id. His rate of pay was lower than six of the other employees who spent the majority of their hours as a grill cook, five of whom were PAR III or IV grill cooks, and one of whom was African American. Id. His rate was higher than the rate for the remaining seventeen other employees who worked primarily as grill cooks. Id.

After moving to Dalton in late 1997, plaintiffs rate of pay was less than one non-African-American PAR I grill cook, but it was higher than the remaining four PAR 0 and PAR I grill cooks, three of whom were non-African American. H/T Report at 101. In 1998, Mr. Rhodes bypassed the grill cook PAR II and PAR III tests and passed the PAR IV test for grill cook. Id. at 102. At the end of 1998, Mr. Rhodes's hourly rate was lower than the seven other PAR III and PAR IV grill cooks who worked primarily in that job. Id. One of these was plaintiff Connie Regan (discussed below) and the other six were non-African Americans. Id. He had a higher rate of pay than all grill cooks at PAR levels below III, with the exception of one African American. Id. At the end of 1999, his hourly rate was less than the other three PAR III and IV grill cooks who worked primarily in that job, one of whom was Mr. Regan and the other of whom were non-African American. Id. In 2000 and 2001, Mr. Rhodes became the most highly paid hourly employee in the Dalton store. Id. 2. Stephen Wilson (former employee)

Mr. Wilson alleges that he worked for Cracker Barrel as a grill cook at its Dalton, Georgia, location from September 1997 through December 1998. Am. Compl. ¶ 12. In contradiction of the Amended Complaint, Mr. Wilson's Declaration asserts that he began his employment with Cracker Barrel in June 1994 as a dishwasher, became a grill cook in August 1994, and quit in January 1999. Wilson Decl. at 1. At his deposition, Mr. Wilson testified that he started working at Cracker Barrel in April 1994, left in January 1997, returned in January 1998, and quit again in December 1998. Wilson Dep. at 45-46, 129. He alleges constructive discharge. Am. Compl. ¶ 41.²³

23 A review of Mr. Wilson's file by defendant's experts shows the following employment history. He applied for and was hired as a dishwasher at the Dalton store and began work on June 8, 1994. He voluntarily terminated on June 5, 1995, and was rehired as a PAR I grill cook on August 14, 1995. He voluntarily terminated again on June 26, 1997, and was rehired again on January 14, 1998, as a prep cook. He terminated again on January 12, 1999. H/T Report at 102.

At initial hire, plaintiff alleges that the only job available was dishwasher. Wilson Dep. at 36-37, 41. Thus, he accepted this back-of-the-house job, and was quickly promoted to grill cook. Wilson Decl. at 1. Plaintiff points to no evidence suggesting that front-of-the-house positions were actually available when he was hired. Thus, he has no channeling claim at hire. Like Mr. Rhodes, Mr. Wilson testified that the Company's best non-management job was back-up cook, which was in the back of the house. Wilson Dep. at 121-22.

Later, plaintiff asserts that he requested to be cross-trained as a server (so that he could position himself for a move into management), but his request was either ignored or he was told to remain as a grill cook. Am. Compl. ¶ 42; Wilson Dep. at 149-50; Wilson Decl. at 3. At the same time, a White grill cook, Devante Delinco, was allowed to cross-train as a server. Wilson Dep. at 150-51; Wilson Decl. at 3.

Mr. Wilson also asserts that he was discouraged from seeking a management position because of the Company's failure to post vacancies for the Management Internship Program,²⁴ by not being cross-trained as a server, and by not being made aware of the opportunity to take the PAR IV grill cook examination. Am. Compl. ¶ 42; Wilson Decl. at 3; Wilson Dep. at 149-51. With regard to the PAR system, Mr. Wilson claims that his managers denied him the opportunity to PAR test for back-of-the-house jobs on three occasions, and that, after he believed that he had passed a PAR II test in 1996, he did not receive the automatic raise. Wilson Dep. at 77-78, 91-92, 120.

24 Mr. Wilson admitted that he was unwilling to relocate (a prerequisite for management) and that he never sought entry into any management training program. Wilson Dep. at 193, 204.

Mr. Wilson alleges that he was underpaid in comparison to similarly situated White employees. Wilson Decl. at 4-5. He alleges that Mr. Delinco held the same PAR level as he did, but Mr. Delinco was paid $2.00 per hour more. Am. Compl. ¶ 43; Wilson Dep. at 124-26. He asserts that African-American employees were routinely denied merit pay increases. Am. Compl. ¶ 43. Mr. Wilson alleges that, in 1998, he acquired seniority over some White cooks, but he was often directed to work the back grill, which is more difficult and typically assigned to persons with less seniority. Id. White cooks were more often assigned to the easier front grill station. Id.

Mr. Wilson also claims that Cracker Barrel has promoted White employees to vacant positions over African-American employees who had comparable or superior experience. Am. Compl. ¶ 43. For example, he points out that, although he had three years' experience working at a janitorial maintenance company, he was denied cross-training to the night maintenance position. Id. When a night maintenance position became available, he sought it, but the job went to a White employee. Id. Plaintiff claims that the store General Manager did not want African Americans on night maintenance because he believed that they could not be trusted inside the store after hours. Id.

Mr. Wilson also alleges that White employees were given preferential treatment in scheduling, with African-American employees almost always required to work on holidays. Am. Compl. ¶ 44. When plaintiff and a White employee were both scheduled to work on July 4, both elected not to show up. Id. However, plaintiff claims that he was suspended and reprimanded, but no action was taken against the White employee.²⁵ Id. Mr. Wilson also maintains that, in his store, certain White managers frequently referred to African Americans as "boy" or "you people" or made other disparaging remarks. Management took no steps to discipline them.Id. Wilson Decl. at 5-6.

25 This allegation of unequal discipline is based upon plaintiffs claim that he knew Mr. Gary Giles (White) was scheduled to work, but he saw Mr. Giles on the street. Later, Mr. Giles bragged that he had not shown up for work, but had not been disciplined. Mr. Wilson does not know if Mr. Giles was ever disciplined. Wilson Dep. at 108-09, 111.

Mr. Wilson's PAR testing history is as follows: Date of Test Skill Par Level of Test Result

08/16/94 Grill Cook 1 Pass 03/06/95 Grill Cook 2 Fail 05/22/95 Grill Cook 2 Fail 04/23/96 Grill Cook 2 Pass 09/25/96 Grill Cook 3 Pass 02/25/97 Prep Cook 1 Fail 02/10/98 Prep Cook 1 Pass H/T Report at 75-56.

With regard to hourly rate of pay, in 1995 there was one PAR I grill cook, an African American, who had a higher hourly rate than Mr. Wilson, and four PAR I grill cooks, all of whom were non-African American, who had lower hourly rates. H/T Report at 103. At the end of 1996, Mr. Wilson was one of four PAR III grill cooks, three of whom were non-African American. Id. He had a lower hourly rate than two of them and the same rate as one of them. Id. at 104. In 1997, Mr. Wilson began working as a prep cook and earned the highest hourly rate of any employee who worked primarily in that job that year. Id. At the end of 1998, Mr. Wilson's hourly rate was greater than all PAR I prep cooks and all PAR 0 through PAR II grill cooks. Id. 3. Connie Regan (current employee)

Mr. Regan began working as a grill cook at Cracker Barrel's Williamsburg, Virginia, location in 1987. Am. Compl. ¶ 13. In 1991, he transferred to the Lumberton, North Carolina, location as a grill cook. Id. In 1995, Mr. Regan transferred to the Dalton store as a grill cook and remains in that position at this time. Id. He is a PAR IV employee. Regan Dep. at 42.²⁶

26 It appears that Mr. Regan actually began work at Williamsburg in 1990, not 1987. Regan Dep. at 28; Regan Dep. Ex. 1; Regan Decl. at 1; H/T Report at 98. Although the above-cited paragraph of the Amended Complaint alleges that plaintiff currently works in Dalton, elsewhere plaintiff alleges that he works in Calhoun, Georgia. Am. Compl. ¶ 46. However, his Declaration states that he worked at Dalton and then at Calhoun. Regan Decl. at 1. The Declaration appears to be more accurate, as the H/T Report states that plaintiff transferred to Dalton in July 1994 and to Calhoun in January 2000. H/T Report at 98-99.

Mr. Regan testified that the only jobs for which he applied at the Williamsburg, Virginia, Cracker Barrel were back-of-the-house jobs that involved cooking. Regan Dep. at 29. These were the only jobs he wanted.Id. When he moved to Lumberton, North Carolina, and later to Dalton, Georgia, he sought to remain a cook because he liked cooking. Id. He never pursued a front-of-the-house position. Id. at 88, 175. Thus, he has no claim that he was channeled to a back-of-the-house position.

Mr. Regan testified that he was never denied the opportunity to cross-train at Lumberton. Regan Dep. at 140. He knew of no other African-American employees who were denied cross-training opportunities at Lumberton. Id. at 140-41. The same holds true when he moved to Dalton. Id. at 176, 233.

Mr. Regan contends that defendant's discriminatory selection procedures adversely affected him. For instance, plaintiff claims that he stated his interest in the position of Lead Store Opening Trainer to his supervisors while at Lumberton. However, he was not afforded the opportunity to apply for that job because the Company failed to announce, post notice of, or accept applications for, vacancies in that position. Am. Compl. ¶ 47; Regan Dep. at 75-76, 136-37; Regan Decl. at 2.

Mr. Regan contends that the defendant's compensation procedures adversely affected him. Am. Compl. ¶ 47. At the Lumberton store, plaintiff claims that White employees were paid higher salaries and received regular pay increases, whereas similarly situated African-American employees were underpaid. Id. He even claims that, when he arrived at Lumberton, he was forced to take a reduction in pay. Id. Mr. Regan asserts that neither he nor any other African-American employee at Lumberton ever received an annual raise in excess of 20 cents per hour. Id.

Mr. Regan contends that White employees receive more frequent raises and promotional opportunities. Am. Compl. ¶ 47. For instance, at the Dalton store, he claims that a White kitchen employee, David Fridele, received a substantial pay increase despite withdrawing from a scheduled PAR examination. Id. However, he believes that co-plaintiff Kelvis Rhodes, who passed three PAR examinations, was given a smaller pay increase.²⁷ Id. However, Mr. Regan was unable to identify any similarly situated White employee who was paid more than he at Dalton. Regan Dep. at 199.

27 However, plaintiff admitted in his deposition that he was unaware of what any other employee at Dalton earned. Regan Dep. at 199.

Mr. Regan admits that he did not seek a promotion to management while at Williamsburg. Regan Dep. at 75. When he moved to Dalton, he worked under African-American managers and considered that path for himself (even talking to a manager about the Management Internship Program), but became discouraged about moving into management when those managers were removed for what plaintiff believed were discriminatory reasons. Id. at 166-68, 175-78, 217; Regan Decl. at 2-3. He never knew of anyone who was excluded from the Management Internship Program, but he never pursued it. Regan Dep. at 232.

This plaintiff also contends that Cracker Barrel promoted White employees into vacant positions over African-American employees of comparable or superior experience. Am. Compl. ¶ 48. In Lumberton, White kitchen employees filled all management vacancies, and, to plaintiffs knowledge, no African Americans were informed of or considered for vacancies. Id. He also complains that there were no African-American managers in Lumberton and only one (a trainee) in Williamsburg. Id. In the kitchen, Mr. Regan asserts that White employees are generally assigned choice positions, regardless of tenure or experience. Id. For example, African-American cooks are relegated to the back grill while Whites occupy the front grill. Id. When an African-American cook, Stephen Wilson, complained about that, management reacted with hostility, leading him to quit. Id.

Mr. Regan claims that, of the 150 persons employed at the Lumberton store, only six or seven African Americans worked there at any given time, and they normally worked in the kitchen. Am. Compl. ¶ 49. Over a four-year period, the Lumberton store had only one African-American hostess. Id. Of the sixty to seventy servers there, at any one time only three were African American. Id. In Dalton, Mr. Regan contends that similar disparities existed of 180 employees at that location, only seven were African American. Id. Five of them worked in the kitchen, despite a number of applicants for front-of-the-house positions. Id. He adds that two African-American employees applied for transfers to cashier positions but were rejected. Id.

Plaintiff claims that, in both Lumberton and Dalton, he observed that African-American employees were disproportionately assigned to work holidays and weekends. Am. Compl. ¶ 50. He alleges that African-American employees in both cities have been continually referred to as "boy" by both managers and co-workers. Id. He asserts that a White co-worker referred to an African-American cook using the "N" word, but a manager, with knowledge of the incident, refused to issue discipline.Id. Finally, Mr. Regan was unable to rank the hourly jobs at Cracker Barrel according to their desirability. Regan Dep. at 176. The logical inference is that this plaintiff does not consider back-of-the-house jobs to be inferior.

Mr. Regan's PAR testing history is as follows: Date of Test Skill Par Level of Test Result

07/10/90 Grill Cook 1 Pass 11/14/90 Grill Cook 2 Pass 08/20/91 Grill Cook 3 Fail 11/05/91 Grill Cook 3 Pass 06/09/92 Grill Cook 4 Fail 11/10/92 Grill Cook 4 Fail 01/12/93 Grill Cook 4 Pass H/T Report at 75. This plaintiff thus had the opportunity to take PAR tests. However, he testified that, while at Williamsburg and Lumberton, management failed to give him advance notice of PAR tests for back-of-the-house jobs, so that he was delayed in taking some tests. Regan Dep. at 120-22, 126-27.

With regard to compensation, Mr. Regan worked exclusively as a grill cook from 1995 through 1999. H/T Report at 99. Beginning in 2000, he also began working as a back-up cook. Id. By the end of 2001, most of his hours were in that position. Id. In 1995, Mr. Regan's hourly rate was $9.55, but by the end of 2001, it had risen to $14.39, an increase of almost 51%. Id. In 1995, among PAR IV grill cooks who worked primarily in that job, his hourly rate was less than two non-African Americans and greater than two non-African Americans and one African American. Id. In 1996 and 1997, his hourly rate was between two other PAR IV grill cooks, both of whom were non-African American. Id. In 1998 and 1999, Mr. Regan was the most highly paid hourly employee at the Dalton store, and in 2000 and 2001, he was the most highly paid hourly employee at the Calhoun store. Id. at 100.

4. Jerry Wayne Barbee (current employee)

Mr. Barbee began working for Cracker Barrel at its Athens, Alabama, location in 1991 as a dishwasher. Am. Compl. ¶¶ 14, 52. He left Cracker Barrel's employment in January 1994, but returned in May 1994 as a dishwasher. Id. He currently works as a back-up cook.²⁸ Id.

28 The H/T Report states that Mr. Barbee was hired in July 1991 and terminated in October 1993. He was rehired and terminated in February 1994. He was rehired again in June 1994. H/T Report at 92.

Mr. Barbee alleges that, despite his interest in a front-of-the-house position, he has never been encouraged to apply for cross-training for such positions. Am. Compl. ¶ 53. However, plaintiff admitted in his deposition that he asked to be hired as a dishwasher and has never sought a server, cashier, or retail employee (i.e., front-of-the-house) position since that time. Barbee Dep. at 47-49, 224-25. Thus, this plaintiff can make no claim that he was channeled at hire or denied cross training to a front-of-the-house job.

Plaintiff also asserts that he was denied the opportunity to participate in the Management Internship Program. Am. Compl. ¶ 53. However, he again admitted in his deposition that he never applied for it or expressed any interest in the Program to any manager because he did not think "it was cut out for me." Barbee Dep. at 337-38. Although plaintiff complains about not being offered a lead trainer position, Barbee Decl. at 4, there is no evidence that he ever applied for it. Similarly, plaintiff complains that he was denied training as a shift leader. However, he admitted that he never applied for that program as well. Id. at 264-65.

Mr. Barbee also alleges that very few of the employees at the Athens store were African American, and most who were worked in back-of-the-house jobs as dishwashers. Am. Compl. ¶ 53. The store had no African-American managers Since 1991, no African-American dishwasher or prep cook became a server or hostess. Id. No African-American dishwasher has ever been promoted to grill cook, but five White dishwashers have. Id.; Barbee Dep. at 122-23; Barbee Decl. at 4. Although local management promised plaintiff that he would be cross-trained for grill cook, this did not occur despite the hiring of six Whites for the grill cook position. Id.

Mr. Barbee claims that Cracker Barrel has promoted White employees to vacant positions over African Americans who have comparable or superior experience. Am. Compl. ¶ 54. After four years, he was promoted from dishwasher to back-up cook. However, a White dishwasher, Jeremy Beauchamp, who worked about six months as a dishwasher and had three less years with the Company than plaintiff, was promoted over Mr. Barbee to grill cook. Id.

He adds that African-American employees are routinely denied information regarding the availability and timing of PAR examinations. Am. Compl. ¶ 54. For example, he was told at hire that he would be given a PAR level examination after thirty days on the job, but he was not tested until ninety days had passed. Id. He was also told that he would be given subsequent PAR examinations every couple of months. Id. However, it took eighteen months for him to receive and pass all four PAR level examinations in his job. Id. He claims that he has only received one merit pay increase during his eight years with the Company. Id.; Barbee Dep. at 275-76; Barbee Decl. at 4.

Plaintiff also alleges that he has witnessed and been subjected to disparate treatment. Am. Compl. ¶ 55. For example, in the area of scheduling, African Americans are not given set schedules and thus have to work different times each week. Id. One of the White back-up cooks, Judy Barrian, asked for and receives every Thursday off. Id. When plaintiff mentioned this schedule to his assistant manager and sought to have a set "off" day, his request was denied. Id. Although this plaintiff makes the general allegation that pay discrimination occurs at Cracker Barrel, he admitted in his deposition that he has no facts to support such a claim, and that he bases it only on his "feeling" that Whites are paid more than African Americans. Barbee Dep. at 115, 135, 218-19.

Mr. Barbee's PAR testing history is as follows: Date of Test Skill Par Level of Test Result

08/27/91 Dishwasher 1 Pass 01/28/92 Dishwasher 2 Fail 06/17/92 Dishwasher 2 Pass 03/18/93 Dishwasher 3 Fail 05/06/93 Dishwasher 3 Pass 12/20/94 Dishwasher 4 Pass 06/19/95 Back-up Cook 1 Pass 09/07/95 Back-up Cook 3 Fail 11/20/95 Back-up Cook 2 Pass 05/14/96 Back-up Cook 3 Pass 10/30/96 Back-up Cook 4 Fail 11/21/96 Back-up Cook 4 Pass H/T Report at 74. This plaintiff thus had the opportunity to cross-train in other jobs and take PAR tests.

The majority of Mr. Barbee's hours was in the back-up cook position for 1995 to 2001. H/T Report at 93. In 1995, he was the most highly paid PAR II back-up cook. Id. From 1996 to 1999, he was the most lowly paid PAR TV back-up cook, but the other two PAR IV back-up cooks were more tenured.Id. Despite their tenure, in 2000 and 2001, Mr. Barbee became the most highly paid PAR IV back-up cook. Id. at 93-94. In fact, in 2000, he was the most highly paid hourly employee in the store, and in 2001, he was the second most highly paid hourly employee (the most highly paid being an African-American grill cook). Id. at 94.

5. Sandra Keel (current employee)

Ms. Keel is a resident of Lake City, Florida, and has worked for Cracker Barrel at its store in that area since 1989. Am. Compl. ¶¶ 15, 57. She alleges that she began work as a dishwasher, and, after seven months, was promoted to prep cook. Id. at ¶ 15. In 1998, she was demoted to dishwasher. Id. After surgery in October 2000, plaintiff returned to work with a restriction on heavy lifting. Keel Decl. at 1. Thus, Cracker Barrel placed her in a front-of-the-house hostess position. Id.

Ms. Keel testified that, when she applied at Cracker Barrel in 1989, she sought a back-of-the-house position as a dishwasher. Keel Dep. at 54-55; Keel Dep. Ex. 1. Moreover, although she claimed in her Declaration, at 4, that Cracker Barrel channels African Americans who apply for front-of-the-house jobs to back-of-the-house jobs as dishwashers, she admitted that she has no information about the positions that applicants seek. Keel Dep. at 114. Ms. Keel added, however, that numerous times she has witnessed unidentified African Americans apply for cashier or hostess or gift shop jobs. Id. Those persons were informed that the store was not hiring, but a few days later Whites were hired for those jobs. Id. at 114-15. She knew nothing of the relative qualifications of the persons hired and those who were not. Id. at 118-19. Ms. Keel alleges that, in 1999, she sought a front-of-the-house position (cashier or hostess). Am. Compl. ¶ 58; Keel Dep. at 263-64; Keel Decl. at 4. Her store General Manager replied that there were no openings. About three days later, two Whites were hired as hostesses. Id. Plaintiff claims that it was only after she was injured and no longer able to work in a back-of-the-house job that Cracker Barrel made her a hostess.²⁹ Id. Ms. Keel never sought a management position and was not willing to relocate (a prerequisite to management). Keel Dep. at 58, 260.

29 Ms. Keel testified that she asked her store manager about moving to a server job in 1996 but never heard back from him. Keel Dep. at 269.

Ms. Keel claims that, over the ten years in which she has worked at Cracker Barrel, she has received only one salary increase. Am. Compl. ¶ 58. In contrast, a White grill cook with less tenure received three salary increases over one year and earns more than she. Id. She adds that White dishwashers are often permitted to work as back-up cooks, but she has never been allowed to do so, although she has passed all four PAR levels as a prep cook. Id. She admits, however, that she could not identify a single similarly situated White employee who was paid more than she. Keel Dep. at 277. She asserts discrimination by her general Store Manager in his application of the PAR requirements for her back-of-the-house positions. Keel Decl. at 2-4.

Ms. Keel alleges that she has observed and been subjected to disparate treatment. During her ten years of employment, she worked with only one African-American manager and claims that she has had few African-American co-workers and few African-American servers. Am. Compl. ¶ 59. Most of the African Americans worked in the back of the house as dishwashers.Id. Despite these allegations, Ms. Keel admitted in her deposition that her store has had an African-American cashier and "plenty" of African-American hosts. Keel Dep. at 268.

She also alleges that White employees are allowed to use the telephone in the store manager's office, but that African-American employees must use the pay phone in the break room. Am. Compl. ¶ 59. African-American employees are routinely scheduled to work weekends, holidays, and closing shifts, while White employees routinely receive weekends and holidays off. Id. African-American employees are often told to end their breaks early to assist in the dish room, whereas White employees are allowed to finish their breaks. Id. Finally, she asserts that, as an African American, she is "written up" for minor infractions for which Whites are not. Id.; Keel Decl. at 5; Keel Dep. at 303-04.

Ms. Keel's PAR testing history is as follows: Date of Test Skill Par Level of Test Result

08/01/89 Dishwasher 1 Pass 04/25/90 Dishwasher 2 Pass 07/20/90 Prep Cook 1 Pass 09/24/90 Prep Cook 2 Fail 11/28/90 Prep Cook 2 Pass 11/29/90 Dishwasher 3 Fail 02/01/91 Dishwasher 3 Fail 04/09/91 Prep Cook 3 Fail 04/09/91 Dishwasher 3 Pass 06/13/91 Prep Cook 3 Pass 12/18/91 Prep Cook 4 Fail 04/20/92 Prep Cook 4 Pass 08/01/99 Dishwasher 4 Pass H/T Report at 74-75. This plaintiff thus had the opportunity to cross-train in other jobs and take PAR tests.

In the years 1995 through 1998, there were no other PAR IV prep cooks at her store, but Ms. Keel was paid at an hourly rate that was greater than that of the other employees who worked most of their hours as prep cooks. H/T Report at 97. In 1999, Ms. Keel's hourly rate was higher than the other PAR IV prep cook, who was African American, but she worked most of her hours as a dishwasher. Id. When compared to other PAR IV dishwashers, Ms. Keel's hourly rate was lower than other co-workers, one of whom was African American and one of whom was not. Id. In 2000, the majority of Ms. Keel's hours were as a dishwasher, but she also worked as a hostess, shop attendant, and prep cook. Id. In 2001, Ms. Keel worked a majority of her hours as a hostess. Id. Her pay rate was between that of the other two PAR IV hostesses, both of whom were African American. Id. 6. Kebere W. Workue (unsuccessful applicant)

Ms. Workue is a resident of Washington, D.C., who applied unsuccessfully for employment at Cracker Barrel stores in Virginia during 1995. Am. Compl. ¶ 16. Ms. Workue alleges that she has been adversely affected by the pattern and practice of race discrimination at Cracker Barrel by not being considered for traditionally White job classifications when she applied for employment. Id. ¶ 61.

In either July or August 1995, Ms. Workue telephoned the Cracker Barrel in Manassas, Virginia, and was told that it had just opened and was hiring employees for every position. Am. Compl. ¶ 62. When she went to that location either that day or the next to complete an application for a server position, a White manager told her either that the store was hiring servers or that no server positions were available, that she had good experience as a server, and that he would call her back when a server position became available. Id. However, no manager called her back. Id.; Workue Decl. ¶ 2. Thereafter, plaintiff applied for server positions at other Cracker Barrel locations in Virginia from August through November 1995 and was never offered a job, despite the fact that Cracker Barrel was hiring servers. Am. Compl. ¶ 63; Workue Decl. at 1 (listing Manassas, Williamsburg, Newport News, Mechanicsville, and Fredericksburg, Virginia, locations).³⁰

30 Workue Declaration adds that she completed applications at all stores except the one in Fredericksburg, Virginia, where someone told her that the store was out of applications. Workue Decl. at 2. At one of the stores, plaintiff claims that an unidentified White employee asked, "'Why Cracker Barrel? Why don't you go somewhere else and apply?'" Id.

Ms. Workue alleges that, when she applied with Cracker Barrel in 1995, she had worked as a server for ten years (1976 to 1986) at the Marriott cafeteria in the Air and Space Museum in Washington, D.C., and from 1986 to the present as a server at a restaurant located in the Museum of American History in Washington called "Daka." Am. Compl. ¶ 64. She alleges that she had nineteen years' experience as a server when she applied unsuccessfully at these Cracker Barrel locations. Id. ¶ 65.

Plaintiffs Declaration lists, however, only the ten years' experience that she had as a server at the Marriott cafeteria located in the Air and Space Museum. Workue Decl. at 1-2. The apparent reason for this change is because plaintiff testified in her deposition that she worked as a cashier at Daka, not as a server. Workue Dep. at 44-49. However, plaintiff falsely stated on the employment application she submitted to the Manassas Cracker Barrel that she worked at Daka Food Services at 12th Street and Constitution Avenue as a waitress (i.e., server) from 1986 to 1995 and that she had been laid off. Workue Dep. Ex. 4 at Bates no. 000007; Workue Dep. at 157-58.

Plaintiffs deposition also reveals that, on the same day she applied at the Manassas location, she contacted attorney Grant Morris. Workue Dep. at 248-49. He drove her to other Cracker Barrel locations where she submitted applications. Id. at 251. Defendant questions whether Ms. Workue was a serious applicant for these positions because she lived in the District of Columbia. Pls.' Mot. for Class Certification [82] at 12, 35. The stores where she applied were located near Interstate Highway exits in either the northern Virginia suburbs of Washington or further outlying areas (e.g., Newport News is over two hours from D.C.). However, she had no car and no driver's license. Workue Dep. at 139. She claimed in her defense that she wanted to find a job, move her family out of Washington, and find an apartment near the store. Id. However, she had no idea if there were housing that she could afford near any of these Cracker Barrel locations or whether public transportation were available. Id. at 138-40.

The omissions and/or false information that Ms. Workue placed on some of her employment applications is disturbing, especially given that she seeks to represent others. For example, on the application that Ms. Workue submitted at the Manassas location, she did not provide any information about her education. Workue Dep. at 135, 147; Workue Dep. Ex. 4 at Bates no. 000007. She listed as prior employment experience that she had worked at Howard Johnson's for two years (1976 to 1978), but she initially admitted that she had worked there for only two weeks. Workue Dep. at 147-48.³¹ She listed the Howard Johnson's as being located at the Air and Space Museum, but it was actually located near the Watergate complex. Id. at 148. For the Howard Johnson job, plaintiff left off her job title, her supervisor's name, her reason for leaving, her job responsibilities, and a contact number at that former employer. Id. at 150-51, 153-54.

31 Plaintiff later claims to have worked at Howard Johnson's for two years, but seemed unsure of the dates. Workue Dep. at 188-89.

Ms. Workue also listed Marriott Hot Shoppes as former employer, but she admitted that she had only applied for a job there. Workue Dep. at 155-57. She also failed to disclose any of the requested information about another prior employer, Gallaudet College. Workue Dep. Ex. 4.³² Moreover, the application asked plaintiff to list three references other than relatives and employers. Ms. Workue violated those instructions by listing only two: her sister and a former employer. Id.; Workue Dep. at 167-69.

32 The Court mentions these omissions because an employer would have legitimate concerns about hiring someone who failed to complete an application. Indeed, in the "Experience" section, Cracker Barrel's application states in all capital letters: "THIS SECTION MUST BE COMPLETED." Workue Dep. Ex. 4.

Plaintiffs Newport News, Virginia, application has similar defects. Workue Dep. Ex. 5; Workue Dep. at 202. For example, she lists as a former employer on this application one she had not mentioned on her application with the Manassas Cracker Barrel — the Washington Hilton Towers. Workue Dep. Ex. 5; Workue Dep. at 171. However, she testified that she only applied there, despite listing beginning and ending dates of employment (1969 to 1970), beginning and ending salary, supervisor's name, and job title. Workue Dep. at 172. The most serious admission plaintiff made, however, is that she did not even move to Washington, D.C., until 1976. Id. at 172-73, 195.

Also on this Newport News application, plaintiff lists Daka Food Services as a former employer, but states that, she worked there from 1970 through 1980. Workue Dep. Ex. 5; Workue Dep. at. 182-83. of course, these dates are different from the dates on the Manassas application (which were 1986 to 1995) and predates by six years the date when she moved to Washington (1976). Plaintiff also failed to list two employers she had listed on the Manassas application — Howard Johnson and Gallaudet. Workue Dep. Ex. 5; Workue Dep. at 198. Finally, regarding each location where she applied, plaintiff has no knowledge of whether anyone was hired instead of her and no knowledge of the qualifications of anyone who may have been hired. Workue Dep. at 103-04, 114, 250.

7. Willie Alexander (current employee)

Mr. Alexander has worked for Cracker Barrel in Savannah, Georgia, from approximately May 1991 through the present as a dishwasher. Am. Compl. ¶¶ 17, 67. Mr. Alexander claims that, during his employment interview in May 1991, he asked what positions were available. Id. ¶ 68. He was told only dishwasher positions were available. Id. His Declaration elaborates by asserting that, when he applied, he asked to be considered for a grill cook position as his first choice and a back-up cook or dishwasher as his second choice. Alexander Decl. at 1. He was told that the only position available was dishwasher. Id. This plaintiff did not seek a front-of-the-house position at hire.³³ In fact, he testified that he was not discriminated against at hire. Alexander Dep. at 64.

33 Mr. Alexander also has no claim that the PAR program discriminated against him, because he testified that he reached PAR IV and was never denied the opportunity to take a test. Alexander Dep. at 141-44.

Mr. Alexander alleges that, in 1992, he informed store General Manager John Freel and that he was interested in being cross-trained and promoted to grill cook and back-up cook (but not cross-trained for front-of-the-house positions). Am. Compl. ¶ 68; Alexander Decl. at 1-2. Mr. Freeland responded that plaintiff was more valuable in the dish room. Alexander Decl. at 3-4. However, when plaintiff persisted, Mr. Freeland responded that, if he went to grill or back-up, he would start at PAR I. Alexander Dep. at 38-39. Plaintiff, who was a PAR IV dishwasher at the time, told Mr. Freeland that he was not willing to cross-train and take a pay cut. Id. ³⁴

34 Mr. Alexander claims that he was not allowed to cross-train to grill cook, prep cook, and host. Alexander Decl. at 4; Alexander Dep. at 103-04; Am. Compl. ¶ 70. However, Mr. Alexander testified that theonly way that he would be willing to cross-train was if he could make the same hourly rate that he made as a PAR IV dishwasher. Alexander Dep. at 45-46. He was unwilling to work as a PAR I in any other position (either in the front or back of the house) when he was a PAR IV in the dish room. Id. at 109.

Mr. Alexander claims that in 1999, he spoke with a White associate manager named Mary Ellen about his participation in the shift leader training program.³⁵ Am. Compl. ¶ 69; Alexander Decl. at 2-3; Alexander Dep. at 158-62. Ms. Ellen replied that for him to be considered, he must have attained PAR level IV and passed a written test. Id. Mr. Alexander had achieved PAR level IV and requested to take the test. Id. However, plaintiff claims that Ms. Ellen never responded to his request. Id. He asserts, however, that four employees (three White and one African-American, three of whom may have been junior to him) were later selected by the store's White General Manager to participate in that training program. Id.

35 Plaintiff used the term "associate manager training program" in his Amended Complaint and Declaration, but testified that he meant shift leader training program. Alexander Dep. at 125, 158.

Mr. Alexander alleges that in 1999 he informed his store General Manager of his interest in becoming an Associate Manager. Am. Compl. ¶ 69; Alexander Decl. at 3; Alexander Dep. at 162-64. The General Manager responded that he could not afford to lose plaintiff from the dish room. Id. Plaintiff was aware of the Company's Management Internship Program because an unnamed male associate manager told him about it and because of postings about it in the store. However, he never applied for it. Alexander Dep. at 117-21, 129; Alexander Dep. Exs. 9-10.

In his Declaration, Mr. Alexander claimed that African Americans are channeled to the back-of-the-house jobs even when front-of-the-house jobs they sought are available. Alexander Decl. at 3. However, Mr. Alexander admitted in his deposition that he could not identify anyone who sought an available front-of-the-house job who was hired in a back-of-the-house job instead. Alexander Dep. at 165-66. He further admitted that he had no personal knowledge of which jobs were available or which jobs applicants were seeking. Id. at 166.

Mr. Alexander testified that his store has had African-American hosts, servers, and cashiers. Alexander Dep. at 104-105. In fact, Mr. Alexander testified that he did not know of any African Americans who had been denied employment. Id. at 106. He also stated that African Americans obtained front-of-the-house positions at his Cracker Barrel store. Id. at 107.

Mr. Alexander alleges that Cracker Barrel's allegedly discriminatory compensation procedures have adversely affected him. For example, in December 2001, he and other African-American employees requested their evaluations so that they could receive annual pay increases. Am. Compl. ¶ 71; Alexander Decl. at 4. Several White employees had been awarded their raises six months before. Id. However, plaintiff admitted that he was unaware of any White dishwasher who made more money than he did. Alexander Dep. at 115.

Mr. Alexander's PAR testing history is as follows: Date of Test Skill Par Level of Test Result

08/19/91 Dishwasher 1 Pass 01/15/92 Dishwasher 2 Pass 05/19/92 Dishwasher 3 Pass 12/14/92 Dishwasher 4 Pass H/T Report at 74. This plaintiff had the opportunity to take PAR tests.

Mr. Alexander was the third most highly paid PAR IV dishwasher at his store in 1995 and 1996, earning less than one African American and one non-African American. H/T Report at 91. In 1997 through 2001, he earned the second highest year-end hourly rate among the PAR IV dishwashers.Id. The most highly paid dishwasher was a non-African American, but had worked at Cracker Barrel for five months longer than plaintiff. Id. at 92.

C. Plaintiffs' Witnesses Hawk and Ullom

The most inflammatory testimony relied upon by plaintiffs comes from two White former Cracker Barrel managers, Ms. Marilyn Hawk and Mr. Warren Ullom. Pls.' Br. in Supp. of Class Certification [82] at 14-19. However, when the Court considers their depositions and subsequent Declarations together, disregards statements from the Declarations that directly contradict without explanation prior sworn testimony, Van T. Junkins Assocs. v. U.S. Indus., 736 F.2d 656, 657-58 (11th Cir. 1984), and disregards conclusions, hearsay, and statements not based on personal knowledge, Lott v. Westinghouse Savannah River Co., 200 F.R.D. 539, 548 (D.S.C. 2000); Fed.R.Civ.P. 56(e), the remaining, admissible testimony summarized below does not support the inference of a Company-wide policy of discrimination against African Americans.

1. Marilyn Hawk

Cracker Barrel hired Marilyn Hawk as a server at its Melbourne, Florida, store. Hawk Dep. at 61. Ms. Hawk had seen and understood Cracker Barrel's equal employment opportunity policy and knew that it applied to hiring and promotion. Id. at 175-76. In early 1994, she applied for and was selected to work as a trainer in the new-store-opening department.Id. at 289-90. In this position, she traveled to new stores to train new employees. Id. at 398. When she first started this job, she returned to Melbourne between new store openings. Id. at 60. Later, she performed this task full time. Id. In this position, Ms. Hawk worked at or visited sixteen Cracker Barrel stores, scattered throughout the continental United States. Id. at 60-61, 161-66, 173, 176, 274, 332, 357, 361, 363-65, 373, 377, 380, 382. She left the new-store-opening department in the first quarter of 1997 and resumed work as a server in Johnson City, Tennessee. Id. at 411-12. She worked at that location until her termination in April 1999. Id. at 342.

Ms. Hawk asserts that she witnessed managers using a racially biased numbering system on applications to exclude African Americans. Hawk Dep. at 160-62. She also asserts that she witnessed White managers channel African Americans to back-of-the-house jobs, deny them cross-training, and make biased remarks. Hawk Decl. at 2-6.

However, in reality, "only" three Cracker Barrel managers (two of them in the same store) explained to Ms. Hawk that they had a racial basis for numbers they had placed on applications.³⁶ Hawk Dep. at 162, 275-76. "Only" four managers (including the three mentioned in the prior sentence) ever said anything to Ms. Hawk about hiring, promoting, cross-training, firing, or employment opportunities for African Americans. Id. Despite her knowledge that what these four managers allegedly did or said was against Company policy, Ms. Hawk never reported them. Id. at 276, 337, 339, 348, 468-69. Ms. Hawk's testimony only shows bias by four individual managers. It sheds no light on whether such bias was so pervasive throughout the Company as to justify certification of a class.

36 The Court places the word "only" in quotation marks because one manager engaging in that type of conduct is too many. However, the point is that Ms. Hawk witnessed racial numbering of applications in two of the sixteen stores in which she worked.

2. Warren Ullom

Cracker Barrel hired Warren Ullom in November 1989. Ullom Dep. at 40. After attending a three-month training program (which included classroom sessions and on-the-job training in stores), he worked as an Associate Manager at two stores (Cedar Bluff, Iowa, and Knoxville, Tennessee) over a two-year period before being named General Manager at the Company's Cartersville, Georgia, store in 1992. Id. at 52-54. He remained at that location until his termination in 1998. Id. at 54, 237.

He knew of the Company's equal employment opportunity policy and understood that, as a manager, it was his responsibility to enforce it. Ullom Dep. at 54-55, 59. No one ever told him not to follow that policy. Id. at 63-64. He testified that he did not take race into account in making employment decisions. Id. at 65-66.

Mr. Ullom's declaration makes serious allegations about comments he allegedly heard from store manager colleagues and managers to whom he reported about channeling African-American employees to back-of-the-house jobs, the telling of racist jokes, and the closing of a store because of the area's racial mix. However, most of Mr. Ullom's allegations are contradicted by his earlier deposition testimony, are embellished, taken out of context, or lack probative value.³⁷ In any event, even if one were to credit Mr. Ullom's Declaration, it hardly shows a Company-wide bias against African-American employees. Mr. Ullom worked primarily in one store during his career at Cracker Barrel. He was not in a position to know whether the bias he allegedly witnessed was rampant throughout the Company's nationwide operations.

37 For example, Mr. Ullom's declaration states that Regional Manager John Davoli told him, "Cracker Barrel is a 'white company.'" Ullom Decl. at 3. However, in his deposition, Mr. Ullom admitted that Mr. Davoli never made such a comment. Ullom Dep. at 169. Defendant has listed the numerous discrepancies between Mr. Ullom's allegations and his prior deposition testimony, which the Court has reviewed and found to be accurate. See App. of Plaintiffs' Witnesses, Vol. I, Warren Ullom.

IV. STATISTICAL EVIDENCE

The statistical evidence presented through expert witnesses is voluminous. The Court has reviewed excerpts from depositions of the respective experts as well as the following reports:

Plaintiffs' Expert Statistical Report Regarding Class Certification, prepared by Drs. Edwin L. Bradley and Liesl M. Fox, dated January 11, 2002 (copy filed at Tab 1 of Plaintiffs' Evidentiary Materials in Support of Class Certification, Vol. I) [hereafter " B/F Report"];

Supplemental Expert Statistical Report Regarding Class Certification, prepared by Drs. Bradley and Fox, dated February 8, 2002 (copy filed at Tab 2 of Plaintiffs' Evidentiary Materials in Support of Class Certification, Vol. I) [hereafter " Supp. B/F Report"];

Analysis of Employment Decisions Relevant to Putative Class in the Matter of Rhodes et al. v. Cracker Barrel Old Country Store. Inc. , prepared by Drs. Joan G. Haworth and Janet R. Thornton, dated March 22, 2002 (copies filed in Defendant's Statistical Evidence Appendix and at Tab 3 of Plaintiffs' Evidentiary Materials in Support of Class Certification, Vol. I) [previously identified as H/T Report];

Plaintiffs' Rebuttal Expert Report, prepared by Drs. Bradley and Fox, dated April 24, 2002 (copy filed at Tab 4 of Plaintiffs' Evidentiary Materials in Support of Class Certification, Vol. I) [hereafter " B/F Rebuttal Report"]; and

Response to Plaintiffs' Experts' Rebuttal Report in the Matter of Rhodes et al. v. Cracker Barrel Old Country Store, Inc. , prepared by Drs. Haworth and Thomton, dated June 28, 2002 (copy filed in Defendant's Statistical Evidence Appendix) [hereafter " H/T Rebuttal Report"].

Through these various reports, the experts studied and reported upon:

(A) racial distribution in various positions at Cracker Barrel;

(B) a Job Interest Survey of current Cracker Barrel employees;

(C) a sample of Cracker Barrel employment applications;

(D) hourly position assignments at hire;

(E) hourly employee progress through PAR levels;

(F) salaries of hourly PAR employees;

(G) data on promotion to management;

(H) hiring data; and

(I) termination data and management performance evaluations.

Before addressing each of these studies, an explanation of statistical terms is necessary. Statisticians use a criterion of 5% or less or 0.1% or less probability of occurring by chance ("greater than two or three standard deviations")³⁸ to categorize a result as "statistically significant." The Supreme Court adopted this standard in voting rights cases (e.g., Castaneda v. Partida, 430 U.S. 482, 496-97 n. 17 (1977)) and carried it over to civil rights cases (e.g., Hazelwood Sch. Dist. v. United States, 433 U.S. 299, 308-09 n. 14 (1977)). Kilgo, 789 F.2d at 871 n. 18; H/T Report at 18-19 n. 27; B/F Report at 5-6.

38 The Eleventh Circuit has explained that

[s]tandard deviation analysis tests the hypothesis that under representation of a protected minority group in any sample made up of a protected and a nonprotected group (binomial distribution) might be attributable to normal fluctuations of chance rather than to discriminatory design. The standard deviation is the measure of the predictable fluctuation in a random selection process. The difference between actual ("observed") numbers of the protected group in such a sample, and the number that would be "expected" in a perfectly proportional process of selection from the appropriate pool can then be expressed in numbers of standard deviation. In turn, standard deviations can be expressed in terms of the mathematical probability, that chance is the cause of the disparities (differences between "observed" numbers and "expected" numbers) measured.

Kilgo v. Bowman Transp., Inc., 789 F.2d 859, 871 n. 16 (11th Cir. 1986) (citing B. Schlei P. Grossman, Employment Discrimination Law 1372 n. 331 (2d ed. 1983)).

In Maddox v. Claytor, 764 F.2d 1539 (11th Cir. 1985), the court warned against drawing inferences of discrimination from standard deviations in the range of one to three and that the significance of statistical evidence should be evaluated in light of all the surrounding facts and circumstances.

[I]t is important to note that no bright line or two, three or more standard deviations can be drawn to distinguish automatically between fair and discriminatory employment practices. Just as a standard deviation of two or three does not necessarily exclude legitimate causes other than chance, so a [standard deviation] below that range does not necessarily exclude discrimination as the cause. The Supreme Court has repeatedly warned, "We caution only that statistics are not irrefutable; they come in infinite variety and, like any other kind of evidence, may be rebutted. In short, their usefulness depends on all the surrounding facts and circumstances."

Id. at 1552 (citation omitted).

In addition to examining statistical significance as a criterion for assessing the results of analyses, courts also use the four-fifths, or 80%, rule, taken from the Uniform Guidelines on Employee Selection, which provides,

A selection rate for any race, sex, or ethnic group which is less than four-fifths (415) (or eighty percent) of the rate for the group with the highest rate will be generally regarded by the Federal enforcement agencies as evidence of adverse impact, while a greater than four-fifths rate will generally not be regarded by Federal enforcement agencies as evidence of adverse impact.

29 C.F.R. § 1607.4 (D) (2002). When the ratio is greater than 125%, it is outside the acceptable four-fifths range, but is favorable to the protected group. H/T Report at 17-18; B/F Report at 6-7; B/F Rebuttal Report ¶ 14.

A. Employee Racial Distribution

Plaintiffs' experts analyzed the racial distribution of Cracker Barrel's workforce on three "snapshot" dates (March of 1996, 1998, and 2000). African-American hourly employees were more heavily concentrated in lower PAR levels. B/F Report at 7 Exs. 1-2. The percentage of African-American employees in back-of-the-house positions exceeded their workforce percentage, while the percentage of African-American employees in front-of-the-house positions was below their workforce percentage.Id. at 7-8 Exs. 3A-3C. Finally, when one compares the percentage of African Americans in each position on those same snapshot dates to their workforce percentage, African Americans were "over-represented" in back-of-the-house positions and "under-represented" in front-of-the-house positions. Id. at 8 Ex. 4.

Unfortunately, these analyses are not very useful. They fail to account for the many variables that may properly affect job assignment, such as the applicant's or employee's job preference and job experience, job openings, and labor-market availability. For example, African Americans may have sought back-of-the-house positions because they pay, on average, higher wages than front-of-the-house positions. These analyses show only that a racial distribution exists, not why it exists. The job interest survey and the random sample of applications (discussed infra) help explain why this distribution exists.

B. Job Interest Survey

Cracker Barrel conducted a job interest survey in May 1999. B/F Report at 8; H/T Report at 53-54. A total of 11, 159 employees returned survey forms, of whom 1,073 (9.7%) were African Americans. of these, 99.7% (11, 127) were in one of the nine primary hourly jobs. B/F Report at 8; H/T Report at 54-55. The survey elicited employee interest in different positions with a series of questions, such as, "Are you satisfied with your current job?" and "Would you like another position?" The survey then listed each job so that employees could identify which job they desired. H/T Report at 53-54.

A majority indicated interest in another job or satisfaction with employment. However, racial differences were apparent. African Americans were most interested in back-of-the-house jobs. Non-African Americans were most interested in front-of-the-house jobs. H/T Report at 2-3, 57-58 Table 18. An employee's interest in a specific job differed depending upon the job he held. Employees holding back-of-the-house jobs were more interested in other back-of-the-house jobs than those holding front-of-the-house jobs, and vice versa. Id. at 3, 58-61 Table 19.

Using these data to develop a benchmark for determining whether the presence of African-American employees in dining or retail jobs was statistically significantly less than predicted by their interest, defendant's experts found that in 1996 there were statistically significantly fewer African-American employees in the server jobs than one would have expected by their interest, but more than predicted by their interest in 2000, with a neutral outcome in 1998. In no year were there statistically significantly fewer African-American employees in the cashier, host/hostess, and night maintenance jobs as expressed by their interest. There were fewer African-American employees than predicted from expressed interest in retail non-management (i.e., gift shop) and restaurant management jobs all three years. There were fewer African-American employees than predicted by their interest in the dishwasher position in two of the three years. H/T Report at 3, ¶ 1-63 Table 20.

When plaintiffs' experts examined the same data, their conclusions were not markedly different. See B/F Report at 8-10 Exs. 5-7. For example, their analyses of the job interest data also showed that the percentage of African Americans expressing interest in back-of-the-house jobs was greater than the percentage of non-African Americans expressing interest in those jobs. Conversely, the percentage of non-African Americans expressing interest in front-of-the-house jobs was greater than the percentage of African Americans expressing interest in those jobs. B/F Report at Ex. 5.³⁹

39 Even when plaintiffs' experts modified the percentage of African Americans in the overall Cracker Barrel workforce for the three snapshot dates to incorporate the results of the May 1999 job interest survey using Bayes' theorem (B/F Report at 9-10 Ex. 6), they found that African Americans were over-represented in dishwasher only in 1996, but under-represented in that position in 1998 and 2000. For all three years, African Americans were over-represented in night maintenance, back-up cook, prep cook, and grill cook positions. B/F Report at 9-10 Ex. 7. There were just as many African Americans in the host/hostess and cashier positions as one would have expected for all three years on the basis of their interest. However, African Americans were under-represented in server, gift shop, and management positions for all three years on the basis of their interest. Id.

Defendant's experts point out that their analysis of this job interest survey data differs from that conducted by plaintiffs' experts only because Drs. Bradley and Fox did not include people in their benchmark who were already in a particular position and who did not express interest in another job. H/T Report at 3, ¶ 1-62. Defendant's experts contend, and the Court agrees, that all employees who expressed an interest in a job should have been counted, including those employees who indicated that they were satisfied with their current job and did not indicate another job of interest. If one counts those who expressed satisfaction with their current job, then defendant's Table 20 (discussed supra) contains the accurate analysis. Id. at 62.

Plaintiffs' experts assert that the job interest survey is flawed and should be disregarded because (1) the African-American response rate was less than that of non-African Americans; (2) there was no follow-up on those employees who failed to return a survey to test for non-response bias (which occurs if non-respondents hold different opinions than respondents); and (3) the survey was conducted only on current employees. B/F Report at 8-9. Cracker Barrel's experts respond that plaintiffs' experts did not offer any evidence to suggest that the group of African Americans who responded was likely to provide different preferences than those who did not or any evidence to suggest that those who responded were not a representative sample. H/T Report at 54. In fact, plaintiffs' expert, Dr. Bradley, testified that he had no evidence that non-respondents held different opinions than respondents. Bradley Dep. at 131-32. Given the lack of any reason to reject this survey, the Court finds it helpful in understanding Cracker Barrel's racial distribution statistics.

C. Sample of Employee Applications

Cracker Barrel's experts examined a random sample of applications of employees hired between 1995 and 2001 to determine whether those persons were hired into the jobs they had requested. H/T Report at 46.⁴⁰ This review shows that the proportion of employees interested in each job varied and that job interest varied by race. Id. at 47-48 Table 11. For example, 70% of non-African-American hires applied for server jobs, but only 44% of African-American hires applied for server jobs.Id. at 48. In contrast, only 34% of non-African-American hires applied for dishwasher jobs, but 53% of African-American hires applied for dishwasher jobs. Id. In those situations where the number of standard deviations exceeded the two-or three-standard-deviations threshold, the adverse impact ratios under the four-fifths rule were almost all favorable to the protected group. Id.

40 The experts selected a sufficient number of applications in each region to permit them to make estimates with a sampling error that would be no more than 2%. H/T Report at 46-47 Table 10. They recorded jobs applied for, job hired into, and each candidate's prior job experience. Every job for which a new hire applied was counted. If an employee did not specify a job or indicated that he would take "any" job, then he was counted as if he had applied for all jobs. Id. at 47.

Defendant's experts grouped this same data by jobs in the dining/retail areas (front-of-the-house) and cook/dishwasher areas (back-of-the-house) to determine if there were differences by race in applications for jobs. Their Table 12 shows that African Americans were statistically significantly more likely than non-African Americans to apply for cook/dishwasher, while non-African-American employees were statistically significantly more likely to apply for dining/retail area jobs. H/T Report at 49 Table 12.

Defendant's experts also show that 95% of employees were hired into the job that they requested with little variation by race. H/T Report at 49. Their Table 13 shows the percentage of hires who received the job for which they applied, or who applied for "any" job, by race and overall. Ninety-four percent of the persons hired as dishwashers had requested that job or any available job. Id. African-American hires were statistically significantly more likely to be placed in a server job when it was not requested on the application. Id. at 49-50.

Plaintiffs' experts contend that defendant's Table 13 provides no useful information because it includes applications where the applicant wrote that he would take "any" job. B/F Rebuttal Report ¶ 35. In response, defendant's experts prepared a new analysis excluding those applications marked "any," but found that it made no difference. H/T Rebuttal Report at 22-23 Table 4. The vast majority of African-American (92%) and non-African-American (94%) hires was placed into a position for which the hires showed interest. Id. at 22. African Americans were still more likely to be placed in a server job when they did not request it. Id.

Plaintiffs' experts also examined the 1, 483 applications on which applicants had indicated that they would take "any" hourly job. Plaintiffs' experts computed the distribution of job assignments for both African Americans and non-African Americans and compared them, creating their Exhibit 22. B/F Rebuttal Report ¶ 36. According to Exhibit 22, African-American applicants expressing interest in any job were assigned more often than their non-African-American counterparts to back-of-the-house positions and assigned less often than their non-African-American counterparts to front-of-the-house positions. Id. ¶¶ 37-38.⁴¹

41 Defendant's experts level "summary" criticism of plaintiffs' Exhibit 22 because it does not take into account the applicant's prior experience, does not properly use the sample data, and does not adjust for openings available in the store for each placement. H/T Rebuttal Report at 29.

Although plaintiffs' experts did not prepare a chart summarizing their analysis, they also examined a subset of 498 hires who had expressed interest in "any" job but who had no prior experience listed on their applications. B/F Rebuttal Report ¶ 39. African-American applicants were assigned more often than their non-African-American counterparts to back-of-the-house positions and assigned less often than their non-African-American counterparts to front-of-the-house positions. Id. Plaintiffs' experts note that, of the seventy African Americans from this group assigned to back-of-the-house positions, 89% became dishwashers.Id. of the 159 non-African Americans assigned to front-of-the-house positions, 52% became servers. Id.

Defendant's experts respond that plaintiffs' experts neglect to mention that 82% of the inexperienced, non-African Americans assigned to back-of-the-house jobs became dishwashers and that 73% of the inexperienced, African Americans placed in front-of-the-house jobs became servers. H/T Rebuttal Report at 30. The percentage of inexperienced African-American and non-African-American applicants who were placed in dishwasher and server jobs was not statistically different by race. Id.

When defendant's experts examined the jobs into which applicants were hired and those that they requested by area of the store, the data show that the Company hired employees into the job category requested, regardless of race. H/T Report at 50 Table 14. Over 97% of those who requested only cook/dishwasher-area (back-of-the-house) jobs were placed there, while over 99% of those who requested only dining/retail (front-of-the-house) area jobs were placed there. Id.

Defendant's experts also examined whether racial differences in the prior experience of hires were evident with respect to the job for which they were selected. H/T Report at 51. Their Table 15 compares the percentage of hired applicants with previous experience in each job, for all jobs and by race. Regardless of race, dishwashers and hosts/hostesses were less likely to have had relevant experience before hire. In four of the jobs (cashier, host/hostess, dishwasher, cook), a greater percentage of the African-American hires had prior experience in the hire job (but those differences were not statistically significant). See id. at Table 15. African Americans hired as servers were significantly less likely to have had prior experience in that job. Id. at 52.

When experience is compared to job placement, over 97% of those who applied for jobs in only one area of the store (dining/retail or cook/dishwasher) were placed in that area. H/T Report at 52. There was no pattern adverse to African Americans when persons sought jobs in both areas when experience was considered. Id. at 52 Table 16. For example, 36% of the non-African-American employees and 34% of African-American employees who were hired into the dining/retail area of the store had prior experience in those types of jobs. Id. at 52. Similarly, 35% of non-African-American employees and 34% of African-American employees were hired into jobs in the dining/retail area when they had experience in both the dining/retail and cook/dishwasher areas. Id. at 52-53.

Plaintiffs' experts responded to the above analysis by taking the application rate benchmarks from Tables 10-12 of the H/T Report to calculate whether differences in the selection rates of African-American employees for front-of-the-house positions were significantly correlated to race. See B/F Rebuttal Report ¶¶ 28-31.⁴² The analysis conducted by plaintiffs' experts is found in their Exhibit 21, which shows that, for the period January 1995 through November 2000, using defendant's job interest benchmarks and modifying them using Bayes' theorem,⁴³ there was statistical significance both in the shortfall of African Americans selected for server and in the over-selection of African Americans for dishwasher. Id.

42 Plaintiffs' experts assert that using application data to determine whether African Americans received the jobs they sought is flawed. B/F Rebuttal Report ¶ 34. They contend that, unless it is known whether the applications were filled out before or after an applicant's interaction with management, it is impossible to draw any conclusion about whether the applications were voluntary or themselves the product of efforts to steer African-American applicants to the back of the house. Id. For example, if an African American were told that the only job available was dishwasher before he completed the application, he might have listed dishwasher as the job sought. Thus, it cannot be assumed that he sought that job voluntarily. Id. They also assert that this analysis does not take into account that African Americans may have sought back-of-the-house jobs because of Cracker Barrel's reputation or if they were discouraged from front-of-the-house jobs in subtle ways.Id. However, Dr. Bradley conceded that he could not identify a single situation where (1) an African American was "chilled" from applying for a particular job; (2) an African American was "steered" into a job he did not want; or (3) an African American was falsely told that the only job available was dishwasher. Bradley Dep. Vol. 1 (3/6/02) at 113, 312-13; Vol. 11 (5/29/02) at 572. Therefore, these speculative assertions do not undermine the usefulness of reviewing actual applications.

43 Bayes' theorem is a theorem about conditional probabilities and states that the probability that an event A occurs given that another event B has already occurred is equal to the probability that the event B occurs given that A has already occurred multiplied by the probability of occurrence of event A and divided by the probability of event B.Merriam-Webster's Collegiate Dictionary 98 (10th ed. 2001).

Defendant's experts point out critical deficiencies in the analysis by plaintiffs' experts in their Exhibit 21.⁴⁴ Because of these errors, the Court cannot use plaintiffs' Exhibit 21 to assist in making a class certification determination. Overall, the data discussed above support the conclusion that employees were placed into jobs for which they had expressed interest. Thus, applicant preference, not discrimination, apparently accounts for the racial distribution that exists at Cracker Barrel.

44 For example, Cracker Barrel points out that plaintiffs' experts constructed the proportion requesting each job from a sample of hires between 1995 and 2001, but applied it to a smaller time range that excluded hires after November 2000. Cracker Barrel created two new regions in 2001 that were included in the application sample examined by defendant's experts, but they were not in the hires analyzed by plaintiffs' experts. H/T Rebuttal Report at 24. Second, plaintiffs' experts did not correctly weight the regional representation of applicants, so the comparisons they made are inaccurate. Id. at 24 n. 29. Third, the sample was not drawn by defendant's experts to be able to determine, within a 2% error rate, whether or not there were differences in job interest and placement at the store level. Id. at 24-25. Any comparison of racial composition using the sample should have incorporated the expected error rate for each region in the significance tests, but plaintiffs' experts' analysis did not do so. Id. at 25.

D. Position Assignments at Hire

Plaintiffs' experts examined the distribution of position assignments at time of hire into PAR level 0 between 1995 and 2000. B/F Report at 10-11 Ex. 8. Their Exhibit 8 shows that, for each year from 1995 through 2000, the proportion of African-American hires assigned to back-of-the-house positions was higher than the proportion of non-African-American hires assigned to back-of-the-house positions. Id. at 11. Conversely, the proportion of African-American hires assigned to front-of-the-house positions was lower than the proportion of non-African-American hires assigned to front-of-the-house positions. Id. The differences were statistically significant for each year. Id.

Although plaintiffs' experts did not create an exhibit, they also analyzed employees hired into PAR level 0 positions as servers and dishwashers compared to their representation in the county where the store in which they were hired was located (using 1990 Census data). B/F Report at 11. For the years 1995 through 2000, the number of African Americans hired into the dishwasher position was compared to their representation in census category 439 (kitchen workers food preparation), while the number of African Americans hired into the server position was compared to their representation in census categories 435 (waiters and waitresses) and 433 (waiter and waitress assistant). Id. at 11-12. Using a multiple-pools analysis, these experts found that 9,221 more African Americans were hired into the dishwasher position than one would have expected on the basis of their representation in the counties where their stores were located. Id. at 12. Conversely, between 1995 and 2000, there were 5,810 fewer African Americans hired into the server position than one would have expected on the basis of their representation in the counties where their store were located. Id. Both numbers are statistically significant. Id.

Defendant's experts respond that the analyses of position assignments at hire by plaintiffs' experts do not account for the fact that initial placement is generally determined by the applicant's expressed interest in specific jobs and schedule preferences, as well as the positions available at the time of application. H/T Report at 14-15, 43. The job interest survey and application sample, discussed supra at Part IV.B-C, support this assertion.

If these factors are ignored, however, defendant's experts contest the use by plaintiffs' experts in their comparison of census code 433 (waiter and waitress assistant) when that job is not part of a server's duties. H/T Report at 44.⁴⁵ Moreover, defendant's experts criticize plaintiffs' expert for using the wrong census code for dishwasher. Id. Plaintiffs' experts admit that using census code 439 (kitchen workers food preparation) for the dishwasher comparison was a mistake, but they reply that using the proper code (444, miscellaneous food preparation occupations) still results in statistical significance. B/F Rebuttal Report ¶ 19; Bradley Dep. Vol. 11 (5/29/02) at 527, 542.

45 Plaintiffs' experts respond that servers perform work listed in census code 433 (waiter and waitress assistant). B/F Rebuttal Report ¶ 22. Even if this census code 433 is not used, however, they assert that statistical significant exists using only code 435 (waiters and waitresses). Id. ¶ 18. Defendant's experts respond that servers spend very little (less than 2% of weekly hours) performing work as a server assistant (which is comparable to waiter and waitress assistant). H/T Rebuttal Report at 10-13, Figs. 4-5. The Court need not resolve this dispute, given the inappropriate use of the county in which the store was located (discussed supra) and errors by plaintiffs' experts in use of census data, discussed infra note 46, that have not been shown to be harmless.

However, defendant's experts validly criticize the plaintiffs' experts for using the county in which the store was located to perform their comparisons. (Plaintiffs' experts do not refute this criticism.) In some counties, the number of observations from the census data is very small, so that the estimated racial composition of the selected occupation is likely to be unreliable. H/T Report at 45. For example, in Gordon County, Georgia (store 320, Calhoun), the number of waiters and waitresses in the census is two. Both of those waiters/waitresses were African Americans. Id. Thus, plaintiffs' experts would have assumed an inaccurate 100% Black benchmark for this county. Id.; see also H/T Rebuttal Report at 3-10.⁴⁶

46 The deposition of plaintiffs' expert revealed that he also made serious errors in using census data. For example, the expert reported the availability of African-American servers in Autauga County, Alabama (store 273), was 52.3%; the availability of African-American servers in Van Zandt County, Texas (store 266), was 10.1 %; the availability of African-American servers in the county containing store 366 was 12.3%; the availability of African-American servers in Hunterdon County, New Jersey (store 386), was 3.9%; and the availability of African-American servers in Harrisonburg City County, Virginia, was 6.2%. However, the census data actually reflect that there were no African-American servers in those counties. Bradley Dep. Vol. 11 (5/29/02) at 517-20 and Bradley Dep. Ex. 12. Such errors, which could create the appearance of statistical significance where none exists, undermine the reliability of the analyses conducted by plaintiffs' experts.

By using the larger Metropolitan Statistical Areas where available and the correct census categories, defendant's experts report that Cracker Barrel hired fewer African-American servers than one would predict, but the same number of African-American dishwashers that one would predict. H/T Report at 45; H/T Rebuttal Report at 15 n. 19.

E. Progress Through PAR Levels

Plaintiffs' experts conducted analyses of the progress of African Americans through PAR levels. B/F Report at 13. They first examined the length of time that employees remain in each PAR level. Id. Their Exhibit 9 compared the average length of time that African-American and non-African-American employees were in each PAR level from 1995 to 2000, adjusting for position held using multiple-regression analysis. Id. Except for 2000, no statistically significant differences were found between African-American and non-African-American average length of time in PAR level 0. Id. at Ex. 9. However, the differences were statistically significant for every year for PAR levels 1-4, except for PAR level 3 in 1999. Id. at 13.

Plaintiffs' experts next compared the percentage of African-American employees who moved to higher PAR levels from January 1995 through November 2000 to their representation in lower PAR levels, controlling for year of promotion, position, and store. B/F Report at 14 Ex. 10.⁴⁷ Their Exhibit 10 shows that during this time frame there were 225,254 movements from PAR levels I through 4, of which 24,802 were by African-American employees. Id. at Ex. 10. There were 3,532 fewer African-American movements than one would have expected for each PAR level. Id. at 14. The differences were statistically significant at each PAR level, although the adverse impact ratios were greater than 80% for each level and not outside the acceptable range. Id. at Ex. 10. In this analysis, as they did with Exhibit 9, plaintiffs' experts created a comparison benchmark by combining lower PAR levels. See supra note 46. From this analysis, plaintiffs' experts opine that the movement of hourly employees from lower to higher PAR levels is related to race because African Americans spend more time in each PAR level and do not advance as one would expect. B/F Report at 14-15.

47 The percentage of African Americans who moved to PAR level 2 was compared to their representation in PAR levels 0 and 1 combined, while the percentage of African Americans who moved to PAR level 4 was compared to their representation in PAR levels 0, 1, 2, and 3 combined. B/F Report at 13. Plaintiffs' experts claim that, because the distribution of African Americans is concentrated in lower PAR levels, it is reasonable to assume that African Americans are not advancing through the PAR system and, therefore, that each PAR level would be a tainted pool if used as a comparison benchmark. They argue that combining lower PAR levels removes the racial taint. Id. at 13-14; B/F Rebuttal Report ¶ 48. As defendant's experts correctly point out, however, this combination of lower PAR levels does not reflect how persons move from lower to higher PAR levels. H/T Report at 67-68. For example, persons move only one PAR level at the time; thus, persons in PAR level 1 cannot move to PAR level 4. See id. Because the percentage of African Americans in lower PAR levels is higher, combining them inflates the percentage of African Americans in the pool and increases the likelihood of finding statistical significance. Id. at 68.

Defendant's experts counter that the above analysis misunderstands how movement through PAR levels occurs. An employee must have been in his current job and PAR level for a specified period of time, must have had a successful performance evaluation, and must demonstrate his knowledge of the job level he seeks by passing the relevant PAR test. H/T Report at 15-16. The decision to take a PAR tests rests with the employee. Id. at 64.⁴⁸ Any test that is not passed can be retaken at another testing time without penalty or limiting the number of retests one can take. Id. at 64-65. Each store has a defined testing schedule, although the frequency has varied over the years. Id. at 65. Employees have the opportunity to use written "PAR Manuals" to study for PAR tests. Some stores use study groups, while others leave that to individual employees. Id. Defendant's experts state that the analysis by plaintiffs' experts in their Exhibit 10 ignores whether an employee expressed an interest in moving to the next higher PAR level by choosing to take the test and studying sufficiently to pass it. Id. at 68.⁴⁹

48 The Cracker Barrel employee handbook provides: "'The Personal Achievement Responsibility, or PAR, program has been developed by Cracker Barrel Old Country Store to open up opportunity for you, but you choose whether to participate.'" H/T Report at 68 (quoting Employee Handbook 21 (1997), Bates No. 4405). Plaintiffs' experts assert that an employee must "have Cracker Barrel's approval in order to take the test," but they cite no record evidence to support that assertion. B/F Rebuttal Report ¶ 46.

49 This point is well illustrated by the testimony of class representative Jerry Barbee. When asked why he did much better the second time he took a PAR test, Mr. Barbee responded: "I studied harder. I just tried to do the first test too fast." Barbee Dep. at 158. Obviously, individual effort is important to passing and is not measured in any statistical report.

Moreover, defendant's experts point out that plaintiffs' experts prepared an analysis of movements that controlled for year of move, job, current PAR level, and store, but report only aggregated results, which combined all 225, 254 moves across almost six years by PAR level.⁵⁰ H/T Report at 65. With this number of events, even a shortfall of 1% would be statistically significant. Id. at 65-66. The statistically significant shortfall reported by plaintiffs' experts is 1.6% of all PAR level moves. Id. at 66.

50 The danger of aggregated data is well illustrated in the following example:

If Microsoft-founder Bill Gates and nine monks are together in a room, it is accurate to say that on average the people in the room are extremely well-to-do, but this kind of aggregate analysis obscures the fact that 90% of the people in the room have taken a vow of poverty.

Abram v. United Parcel Serv. of Am., Inc., 200 F.R.D. 424, 431 (E.D. Wis. 2001).

Defendant's experts contend that statistics aggregated to a large number like this do not properly reflect the wide variation among the over 440 Cracker Barrel stores in movement to higher PAR levels by African-American employees. H/T Report at 67. They show that many stores have no differences between the actual number of African-American PAR moves and the number predicted. Id. Other stores have more than predicted (such as the stores in which three of the representative plaintiffs worked), and other stores have less. Id. The largest shortfall in any store where a plaintiff worked was 4.1 PAR moves per year. Id. This variation in advancement to higher PAR levels from one store to another, they contend, shows that there is no consistent pattern across the Company of denying PAR moves to African-American employees. Id.

Defendant's experts further assert that, because an employee's move to a higher PAR level depends first on his initiative, job, PAR level, tenure, and evaluation scores, PAR level increases reflect self-selection and test results. H/T Report at 68. Thus, differences in the composition of those who actually choose to take a PAR test could explain the differences that plaintiffs' experts found between the actual and expected number of African-American PAR level moves. Id.

Therefore, defendant's experts propose that analyzing the PAR testing results would be more relevant. They first compared the percentage of African-American employees taking PAR tests in each store to their presence in the store's workforce. H/T Report at 69. Fifty-seven percent of the stores have a higher percentage of African Americans in the workforce than the percentage of African-American employees who took PAR tests. Id. at 69 Fig. 3. Defendant's experts then examined these data to determine whether there was a statistically significant racial difference in the number of PAR tests taken, after controlling for each year and month, store, job, and PAR level. Id. at 69. The percentage African-American in each of these pools of employees was used to estimate the number of African-American employees who could have taken a PAR test. Id. at 69-70. The predicted number was compared to the number of African Americans who actually took a test. Id. at 70.

This analysis produced an estimated average shortfall in the number of PAR tests taken over the seven-year period (1995-2001) of 9.1 African-American employee tests per store (or slightly over one test per store per year). H/T Report at 70. This shortfall, as a percentage of tests taken (338,000), is 1.2% over the entire period. Id. This is statistically significant, but the adverse impact ratio is over 80% and not outside the acceptable range for the four-fifths rule. Id. Defendants' experts thus contend that this result suggests that the shortfall in the number of African-American PAR level moves is related to the difference in interest in taking the test.⁵¹ Id.

51 Plaintiffs' expert, Dr. Bradley, agreed that there were two components to testing: (1) taking the test and (2) passing or failing it. Bradley Dep. Vol. 11 (5/29/02) at 609. He admits that failing to take the test is a reason that persons do not advance in PAR level. Id. at 608.

Plaintiffs' experts disagree, however, that lack of interest explains the shortfall, citing the requirement that African Americans satisfy eligibility criteria and the supposed effect of discouragement in taking or retaking tests. B/F Rebuttal Report ¶ 46. However, all employees, regardless of race, must satisfy eligibility criteria, and plaintiffs' experts do not quantify this "discouragement" factor. In the Court's view, the most probative evidence discounting this alleged discouragement factor is the often rapid rise of almost all the putative class representatives through PAR to level 4 and their frequent taking and retaking of tests. See supra at III.B.

Plaintiffs' experts in their Supplemental Report reviewed individual PAR testing results. Supp. B/F Report at 1. Their Exhibit 19 compares the percentage of African-American employees taking each level of the PAR test to their representation in the PAR level immediately below the level of the test taken, controlling for PAR level, year, store, and position, using a multiple-pools analysis. Id. at 2 Ex. 19. For each year, the percentage of African Americans who took a PAR test was less than their representation in the pool of eligible individuals. Id. The differences were statistically significant. Id.

In their Exhibit 20, plaintiffs' experts analyzed PAR test failure rates by race, controlling for year, position, and PAR level (but not store). Supp. B/F Report at 3 Ex. 20. For each PAR level for each year (1995-2001), the percentage of African-American hourly PAR employees who failed PAR tests was greater than the percentage of non-African Americans who failed the PAR tests. Id. The differences were statistically significant. Id.

Defendant's experts do not question plaintiffs' Exhibit 19, but focus on Exhibit 20. They assert that the analysis used to create Exhibit 20 did not control for store, although other analyses conducted by plaintiffs' experts did. H/T Report at 71. They assert that failure to control for store ignores the different pass rates among them. Also, defendant's experts criticize Exhibit 20 because its results are aggregated nationally by PAR levels across all years. Id.

Defendant's experts prepared a different analysis. They examined the passing rates for employees taking PAR tests to determine if there are differences among the test-takers, regardless of race, by PAR levels, years, jobs, and stores. H/T Report at 71-72. They found that passing rates are statistically significantly different for each of these factors. Id. at 72. For example, as PAR level increases, the percentage failing the test also increases. Id. While over 99.2% of the test-takers passed the PAR level 1 test, only 86.5% passed the PAR level 4 test. Id. As for job type, the highest pass rates are among cashiers (96%) and lowest among grill and back-up cooks (89%). Id. The pass rates are statistically significantly different by year, ranging from 91.7% to almost 96%, and by store, ranging from 85% to 100%. Id. They report similar patterns for both African-American and non-African-American test-takers. Id.

Because store pass rates were statistically significantly different, defendant's expert repeated the analysis conducted by plaintiffs' expert in their Exhibit 20, but controlled for the test-taker's store. H/T Report at 72. In most years, jobs, PAR levels, and stores, the pass rates of African Americans and non-African Americans were identical.⁵² Id. Defendant's experts also measured the difference in the number of African-American test-passers compared to the number predicted from their presence among test-takers in the same year, store, PAR level, and job. This analysis had more comparisons because the only requirement was that there be at least one of each race in a pool.⁵³

52 During the seven-year period, there were 878 combinations of years, stores, jobs, and PAR levels with at least five African-American and five non-African-American test-takers. (Five of each provided sufficient numbers to test the proportion passing with reasonable accuracy, using a t-test.) H/T Report at 72. In 718 instances, the pass rates of African Americans and non-African Americans were identical. Id. In 61 instances (but only two statistically significant), African Americans had higher passing rates, and in 99 instances (only one statistically significant), non-African Americans had higher passing rates. Id. at 72-73. This does not show a pattern of passing rates adverse to African Americans who elected to take tests.

53 Overall, the difference between the actual and predicted number of passes amounted to 0.08% of the 334,783 tests taken, with an adverse impact ratio close to 100%. The results of the analysis are summarized in Table 22. H/T Report at 73 Table 22.

The most probative part of this analysis for class certification purposes is the results by year and store found in Appendix C, Table C2 of the H/T Report. This appendix shows graphically that, in most stores over most years, there was no difference in pass rates between the races. In only eleven of the stores over a seven-year period was the difference between the actual and predicted pass rate of African Americans and non-African Americans statistically significantly adverse to African Americans. H/T Report at App. C, Table C2. This is powerful evidence that there was no pattern of adverse pass rates against African-American test-takers across the entire Company.

Plaintiffs' experts respond that, even if they control for the test-taker's store, there is still statistical significance. B/F Rebuttal Report ¶ 54. As shown in their Exhibit 23, for the years 1995 to 2000, the percentage of African-American PAR test failures was greater than the percentage of non-African-American PAR test failures. Id. ¶ 55 Ex. 23. The difference was statistically significant. Id. ¶ 55. The adverse impact ratios are less than 80% for each PAR level. Id. To. focus on the failure rate of first-time PAR test-takers, plaintiffs' experts created Exhibit 24, which shows that, for each PAR level for the years 1995 to 2000, the percentage of African-American PAR test failures among first-time test-takers was greater than the percentage of non-African-American PAR test failures among first-time test-takers. Id. ¶¶ 56-57 Ex. 24. The difference was statistically significant.Id. ¶ 57. The adverse impact ratios are less than 80% for each PAR level. Id.

Defendant's experts state that the above analysis of PAR testing results prepared by plaintiffs' experts focused on failures rather thanpasses. The plaintiffs' rebuttal report did not disclose the number of tests that make up the difference in pass rates between African-American and non-African-American test-takers. However, during his deposition, Dr. Bradley calculated the adverse impact ratios of the "passing" rates of African-American and non-African-American test-takers on the basis of Exhibit 23 of the B/F Rebuttal Report. Bradley Dep. Vol. 11 (5/29/02) at ¶ 19-22. These ratios, listed in Table ¶ of the H/T Rebuttal Report (at 32), show that pass rates for African-American test-takers are very close to that of non-African-American test-takers. H/T Rebuttal Report at 32. They range between 92% and 99% and, overall, have an acceptable adverse impact ratio of 98%. Id.

Defendant's experts determined from plaintiffs' experts' analyses that the difference between the actual and predicted number of African-American "passes" and "fails" was 234 fewer tests passed by African-American employees out of 292,000 tests taken over a six-year period. H/T Rebuttal Report at 33. This is a difference of less than one-tenth of one percent. Id. Even using an estimated shortfall number (690) that turned out to be over twice the size of the actual passed test shortfall (234), Dr. Bradley calculated that the shortfall in African-American passes per store per year was twenty-nine hundredths of a person (0.29), or, if rounded to the nearest whole number, zero. Id.;see also Bradley Dep. Vol. II (5/29/02) at 635-36. An alternative comparison is the difference between the percentage African-American among those who took the tests across all years and store (11%) and the percentage African-American among those who passed the test (10.8%). H/T Rebuttal Report at 33.

Dr. Bradley also calculated at his deposition the adverse impact ratios for his Exhibit 24 (which examined the first PAR test of an individual for each job). Bradley Dep. Vol. 11 (5/29/02) at 636-37. These ratios ranged from 94% to 99%, with an overall adverse impact ratio of 98%. H/T Rebuttal Report at 33-34. As with the other Exhibit (no. 23), the difference in tests passed by African-American test-takers and the number predicted is small. Even using the estimated shortfall that turned out to be twice as large as the actual, Dr. Bradley calculated the average shortfall for each store/year combination to be so small as to round to zero. Id. In addition, the percentage African-American among first-time test-takers across all years and stores is 10.9%, while the percentage African-American among those who passed the tests is 10.6%. Id. at 34. The test analyses on which plaintiffs' experts focused their attention (Exs. 23-24) had adverse impact ratios that were very close to 100%, with differences in the number of tests per store per year that were near zero despite the large number of tests taken. Id.

Continuing with the analysis of test-passers, defendant's experts point out that Dr. Bradley had not examined and did not know what percentage of the stores were statistically significantly the same or favorable to African Americans. H/T Rebuttal Report at 34. Defendant's experts asserts, and the Court concurs, that this is a relevant question in a class certification proceeding because a store-level analysis helps answer the question of whether the difference that plaintiffs' experts found when evaluating test failures was (1) a common pattern across all stores or (2) the result of aggregating statistics across a large number of tests over a long period of time. Id.

Defendant's experts report that the analyses by plaintiffs' experts reported in their Exhibits 23 and 24 do not present results by store.⁵⁴ H/T Rebuttal Report at 34. Had they done so, the Court would have seen that, in most years and in most stores, the difference between the actual and predicted number of African-American test passes is within the neutral range. See id. Over 40% of the store/year combinations had a zero difference between the predicted and actual passes for African Americans. Id. In 70% of the store/year combinations, the results were perfectly neutral or more African-American test-takers passed than predicted. Id. at 34-35.

54 On average, stores administered 141 tests each year, ranging from an average of 128 in 2000 to 161 in 1997. H/T Rebuttal Report at 37. Across the six years analyzed by plaintiffs' experts, stores administered an average of 658 tests. Id. Table 7 shows the distribution of the number of tests administered by stores each year. Id. at 37-38 Table 7.

This analysis is graphically presented in Figure 8, which reflects the number of store/year combinations that show no difference (zero standard deviations) as "perfect parity." H/T Rebuttal Report at 35 Fig. 8. An overwhelming number (96%) of the store/year combinations had results that were neutral (between 0 and ± 1.96 standard deviations) or statistically significantly favorable (more than 1.96 standard deviations) to African Americans. Id. In only 4% of the store/year combinations were the results statistically significantly adverse to African Americans. Id. at 36. In 93.3% of the store/year combinations, the adverse impact ratios associated with the pass rate were greater than 80%. Id. These results are not consistent with a pattern adverse to African Americans that is common to all stores. Id.

Defendant's experts also examined the actual differences between the number of tests passed by African-American employees compared to the number predicted. H/T Rebuttal Report at 36-37 Fig. 9. As shown in Figure 9, there were 891 store/year combinations in which the difference between the actual and predicted number of tests passed (or failed) by African-American employees was zero. Id. at 36 Fig. 9. In 527 store/year combinations, there were between zero and one more test passed by African-American employees than predicted. Id. In twenty-six store/year combinations, there were between one and three more tests passed by African-American employees than predicted. Id. There were 492 store/year combinations in which the number of test passed by African-American test-takers was less than one test different than predicted, and in 103 store/year combinations, the shortfall in passes was between one and two tests. Id. There were twenty-four store/year combinations with between two and five fewer tests passed by African-American test-takers. Id.

Defendant's experts also report that, after controlling for store, job, and PAR level of test, African-American employees take an average of 1.05 tests and that non-African Americans take an average of 1.04 tests. H/T Rebuttal Report at 38. In most store, job, and PAR levels (67.6%), there was no difference in the average number of tests taken by race.Id. In some store, job, and PAR levels (15.7%), African-American employees took the tests fewer times than non-African Americans. Id. In most of these comparisons (98.9%), there was no statistically significant difference in the number of times employees took tests by race. Id. at 38-39.⁵⁵

55 Defendant's experts also examined whether employees "ever passed" a particular PAR test. H/T Rebuttal Report at 39. Over the six-year period examined by plaintiffs' experts, the overall shortfall in the number of African Americans ever passing was 133. Id. Figure 10 reflects that analysis by store. In the majority of stores (52.5%), African Americans passed the tests at the same rate as non-African Americans or passed at a higher rate. Id. at 39-40 Fig. 10. Figure 11 shows these differences in "ever passed" in terms of standard deviations. In 91% of the stores, the results are neutral with respect to race. Id. at 40-41 Fig. 11. In 52.5% of the stores, the number of standard deviations is either equal to zero, or the number is positive, favoring African-American test-takers. Id. In 8.3% of the stores, the results are statistically significant adverse to African-American test-takers. Id. In 99% of the stores over the six-year period, the adverse impact ratios associated with "ever passed" rates are within the acceptable rage of 80% to 125%. Id.

As noted above, plaintiffs' experts assert that the failure to pass the PAR tests among African-American employees prohibits them from moving upward in PAR level. B/F Report at 14-15; B/F Rebuttal Report ¶ 55. To test this assertion, defendant's experts examined the upward movement among those who took each test. H/T Rebuttal Report at 42. Using plaintiffs' experts analysis, they identified test-takers, regardless of outcome (pass or fail) by store, year, job, and PAR level of test, and considered whether or not these test-takers had an upward movement in PAR level around the time they took the test. Id. Overall and by year, more African-American employees (+253) moved to a higher PAR level among test-takers than predicted. Id. Overall and by year, adverse impact ratios were close to 100%. Id. Figure 12 shows the results of this analysis by store. Id. at 42-43 Fig. 12. In the majority of stores (59.9%), African-American employees moved to a higher PAR level among test-takers at the same rate (difference of zero) as non-African-American test-takers (15 stores) or moved to a higher PAR level at a higher rate than non-African-American test-takers (251 stores). Id. In 178 stores, fewer African-American test-takers moved upward than predicted. Id.

Figure 13 reflects the results of this analysis in standard deviations. H/T Rebuttal Report at 43-44 Fig. 13. Only 3.4% (15 stores) of all stores had statistically significantly fewer African-American test-takers moving upward than predicted. Id. In the vast majority of stores (96.6%), the results are neutral or favorable to African Americans. Id. The majority of the adverse impact ratios are either within the acceptable range or above (i.e., favoring African Americans). Id. Overall, the data do not suggest that PAR tests on a store-level analysis are a barrier to movement by African Americans.

F. Salaries of Hourly PAR Employees

Plaintiffs' experts examined the pay of African-American employees versus that of non-African Americans (excluding tips), and claimed that, for every year, African Americans were paid less. B/F Report at 17-19 Exs. 12-13. Defendant's experts also performed analyses that excluded tips and came to the opposite conclusion, but validly criticized the plaintiffs' experts. H/T Report at 16-17, 22-23, 30-31 Tables 4-9.⁵⁶

56 Even with the faulty analysis, one of plaintiffs' experts still found that, in some years, African Americans were paid more than non-African Americans for certain jobs. Bradley Dep. Vol I. (3/6/02) at 204-05 (discussing Exs. 12-13).

In response, plaintiffs' experts abandoned that analysis and devised a new one, although it is questionable as well. On this second effort, Drs. Bradley and Fox computed the total compensation of Cracker Barrel's employees as the sum of regular income, overtime income, and "adjusted" tip income. B/F Rebuttal Report ¶¶ 64-69. They claimed that tip income is regularly under-reported, and searched the Internet to learn that estimates of under-reporting of tips vary between 16% and 84% of reported tips. Id. ¶ 67. They, therefore, assumed under-reporting of tips by Cracker Barrel's servers, doubled each server's reported tip income, and then ran their calculations performing a multiple-regression analysis that modeled total compensation for a year, adjusting for years of service, total number of hours worked, and store. Id. ¶¶ 66-68.

They assert that, for the years 1997 to 2000, the average compensation of African-American employees was from $425 to $527 less per year than their non-African-American counterparts. B/F Rebuttal Report ¶ 69 Ex. 27.⁵⁷ Dr. Bradley admitted, however, that he had no factual basis to assume that Cracker Barrel's servers under-reported tips and that he had never made this assumption in any other case, report, or consulting project. Id.; see also Bradley Dep. Vol. 11 (5/29/02) at 679, 694.

57 Although these results were statistically significant, significance was created by doubling server's tip income. H/T Rebuttal Report at 46-48 Fig. 14. Plaintiffs' experts also made a significant error in their calculations by doubling overtime hours. Id. at 47. This error resulted in too many hours for those persons who had overtime. Id. Because plaintiffs' experts adjusted an individual's compensation for hours worked, this resulted in a bias against persons who worked overtime. Id.

Defendant's experts recreated the analysis conducted by plaintiffs' experts, but included reported tips rather than doubling them. (Defendant's experts also corrected the error plaintiffs' experts made of counting overtime hours twice.) H/T Rebuttal Report at 47. When actual tips were analyzed using plaintiffs' experts' program, African-American employees had statistically significantly higher compensation every year. Id. Moreover, when actual tips were included (not doubled), the data show that, overall, jobs in the cook/dishwashing area of the store in 1996 earned a higher rate of pay than those in the dining/retail area of the store. Id. at 49 Fig. 15 (similar patterns existed for 1998 and 2000).⁵⁸ Thus, a correct analysis shows no pay discrimination against African Americans.

58 The average hourly rates of pay in 1996 were (1) $6.02 for all dining/retail area employees; (2) $6.13 for dining/retail employees (excluding servers); and (3) $6.64 for all cook/dishwasher area employees. H/T Rebuttal Report at 49-50 Fig. 16.

G. Promotion Into Management

Plaintiffs' experts analyzed the promotion of hourly PAR employees into management positions using the percentage of African-American employees at PAR IV level as the comparison benchmark. B/F Report at 15. Their analysis, found in Exhibit 11, controlled for year of promotion, management position, and store. Id. at 16 Ex. 11. Overall, from January 1995 through November 2000, there were 593 promotions to management, of which twenty-one went to African Americans. Id. One would have expected that 8.03% of those promotions, or about 48, would go to African Americans. Id. This discrepancy (twenty-seven fewer African-American promotions than one would expect) represents a statistically significant difference of -4.39 standard deviations, with an adverse impact ratio of 42%, well below the 80% threshold of the four-fifths rule. Id.

Defendant's experts fault this analysis because it combines management and trainee positions that have different criteria for entry. H/T Report at 16, 76. The plaintiffs' experts assumed that all PAR IV employees have an equal chance of entry into management, but ignored the fact that most of the MIT participants are from external labor markets. Id. at 76. (Plaintiffs' experts replied that they were provided no data on entry into the MIT program. B/F Rebuttal Report ¶ 63.) Defendant's experts argue that the analysis by plaintiffs experts also ignores the differences in the opportunities for moving to each of the different management positions analyzed and does not incorporate available data on employee preference. H/T Report at 80-81. Defendant's experts assert that an analysis of employees who expressed an interest in a management job and who met the basic qualifications necessary to enter the management program shows that African Americans and non-African Americans are equally likely to be part of the Management Internship Program. Id. at 81-82.

Defendant's experts show that plaintiffs' experts examined 593 promotions of hourly employees into various management positions. H/T Report at 79. Three hundred thirty-two of these promotions were of employees who met the entry requirements for the Internship Program from a variety of PAR levels (not just PAR level IV). Id. at 79 Table 24. Plaintiffs' experts also included in their analysis twelve Associate Manager entrants into the MIT program, as well as promotions into non-management positions such as Gift Shop Management Trainee, Employee Training Coordinator, and Restaurant Management Trainee. Id. at 80 Table 24.

Defendant's experts showed that all the jobs held by PAR IV employees do not have the same likelihood of entering the various "management" positions analyzed by plaintiffs' experts. H/T Report at 80. of the 332 moves into the Internship Program, 204 came from the server position and 64 from the grill cook position. Id. Almost all of the 124 moves into the Gift Shop Management Trainee came from gifts (73) and cashier (30). Id. Most of the fifty-five employees who entered the Employee Training Coordinator job came from the server position (42).⁵⁹ Id.

59 According to defendant's experts, the assumption by plaintiffs' experts that all PAR IV level employees have an equal chance of promotion is unfounded. H/T Rebuttal Report at 51. The Management Internship Program guidelines do not require cross-training in all of the store's hourly jobs. Id. Given the nature of the Gift Shop Manager job, it is more likely that employees who have worked in the jobs managed by the Gift Shop Manager will have the background for this job. Id. Defendant's experts used those in PAR IV in the jobs from which the promotion occurred to adjust for the fact that employees with a PAR IV in jobs that are not required as part of the cross-training requirement are less likely to move into the Management Internship Program. Id.

Therefore, defendant's experts conducted an analysis that takes these factors into account by defining the candidate "pool" to include employees in the same job as the mover. H/T Report at 80. Table 25 summarizes the results of this modification to the plaintiffs' experts' benchmark for the three positions that had statistically significant differences that were adverse to African Americans under plaintiffs' experts' analysis: Employee Training Coordinator, Gift Shop Management Trainee, and Intern. Id. at 80-81 Table 25. None of the three had statistical significance. Id. at 81. A comparison of the benchmark (percent African-American) in Table 24 to that in Table 25 shows that the effect of plaintiffs' experts' assumption that all PAR IV employees have an equal chance of movement into management, regardless of the job in which they worked, was to inflate the percentage of African Americans in the pool. Id.

Defendant's experts state that an analysis of hourly employees movements into management should consider only those who have applied and who meet the minimum qualifications. H/T Report at 81. An analysis of these data shows no statistically significant difference between the number of African-American employees who moved into the Internship program and the number predicted by their presence among those qualified who applied. Id. at 81-82. Defendant's experts were unaware of any evidence that African Americans were deterred from applying for the Internship program. Id. at 81 n. 94. Plaintiffs' expert also had no evidence of deterrence. Bradley Dep. Vol. I (3/6/02) at 113.

Plaintiffs' experts assert that they conducted the analyses urged by defendant's experts (i.e., controlling for "job of the mover") in all 593 promotions, but still found a difference of -2.27 standard deviations with an adverse impact ratio of 67.9%. B/F Rebuttal Report ¶¶ 58-60 Ex. 25. Plaintiffs' experts also asserted that they disagreed that one should control for "job of the mover." Id. ¶¶ 58, 61. Restricting promotion opportunities to any one group (i.e., promotions to Internship Program limited to server classification) is at odds with Cracker Barrel's actual selection process, which does not give a preference to any hourly group. Id. ¶ 61.

Plaintiffs' experts also analyzed the movement of hourly employees into management positions who applied for such positions and met the minimum qualifications. B/F Rebuttal Report ¶ 62 Ex. 26. In contrast to defendant's experts, who found no statistical significance adverse to African Americans, plaintiffs' experts report that the difference between the number of qualified African Americans who moved into the Internship program compared to the number predicted was -2.11 standard deviations, with an adverse impact ratio of 68%. Id. The shortfall was 5.7 African-American selections to the Internship program over a four-year period. Id.

Defendant's experts respond that they calculated the exact number of standard deviations as -1.915, which is not statistically significant, and that plaintiffs' experts used a hypergeometric approximation⁶⁰ to calculate the number of standard deviations. H/T Rebuttal Report at 52. Even if the Court credits plaintiffs' experts' approximation over an exact calculation, -2.11 is barely within the statistically significant range of two to three standard deviations. See Maddox, 764 F.2d at 1552.

60 A hypergeometric distribution is a probability function that gives the probability of obtaining exactly x elements of one kind and n — x elements of another if n elements are chosen at random without replacement from a finite population containing N elements of which M are of the first kind and N — M are of the second kind.Merriam-Webster's Collegiate Dictionary 569 (10th ed. 2001).

H. Hiring Data

Plaintiffs' experts used the broad service workers category (EEO-9), used in government-mandated, annual EEO-1 filings, to evaluate the racial composition of hourly hires at Cracker Barrel from 1995 through 2000. B/F Rebuttal Report ¶ 25. On the basis of the percentage of African-American service workers in that category in the county where each store was located in the 1990 Census, plaintiffs' experts opined that Cracker Barrel hired over 18,000 fewer African Americans than one would have expected (-79.36 standard deviations). Id. Plaintiffs' expert, Dr. Bradley, testified that he used the EEO-9 service worker category to make his comparison because the jobs Cracker Barrel fills are included among the long list of service occupations that make up the EEO-9 category and because defendant hires unskilled, entry-level people. Bradley Dep. Vol. 11 (5/29/02) at 559-60.

Using EEO-9 data for this comparison is inappropriate, however. The EEO-9 service worker category includes over fifty jobs, most of which involve no restaurant or gift shop work. For example, included in the EEO-9 category are jobs such as elevator operator, hairdresser, child care worker, guard, nursing aide, barber, and bellhop. H/T Rebuttal Report at 18-19 Table 3. Dr. Bradley agreed that the large majority of EEO-9 occupations were ones for which Cracker Barrel did not hire. Bradley Dep. Vol II (5/29/02) at 561.

Moreover, the sample of employment applications received shows that most hires had prior experience in the job sought. H/T Rebuttal Report at 19. Thus, while Cracker Barrel may not require prior experience, hiring managers obviously chose persons with experience to reduce training costs. Id. at 19-21 Fig. 6. Because it was inappropriate to use this broad EEO-9 benchmark, the comparison made by plaintiffs' experts should be disregarded.

I. Other Analyses

The experts also examined termination data and management performance evaluation data by race. B/F Report at 19-20 Exs. 14-16; B/F Rebuttal Report ¶¶ 70-71; H/T Report at 82-90 Tables 26-28; H/T Rebuttal Report at 52-53. However, because neither terminations nor disparate evaluations are at issue here, the Court does not address these analyses. The Court also notes that the respective experts did not analyze cross-training data because Cracker Barrel does not maintain it.

V. LEGAL ANALYSIS

The Court's analysis begins with the following directive from the Eleventh Circuit:

[T]he legitimacy of a private Title VII suit brought on behalf of a class depends upon the satisfaction of two distinct prerequisites. First, there must be an individual plaintiff with a cognizable claim, that is, an individual who has constitutional standing to raise the claim (or claims) and who has satisfied the procedural requirements of Title VII. Second, the requirements of Rule 23 of the Federal Rules of Civil Procedure must be fulfilled; in other words, the individual plaintiff must be qualified to represent the members of the class in accordance with the four prerequisites of Rule 23(a), and the action must be one of the three types Rule 23(b) identifies.

Griffin v. Dugger, 823 F.2d 1476, 1482 (11th Cir. 1987) (footnotes and citations omitted). Therefore, the Court will address (a) standing, (b) Title VII's procedural requirements, and (c) the requirements of Rules 23(a) and (b)(2).

A. Standing

"'A named plaintiff in a class action who cannot establish the requisite case or controversy between himself and the defendants simply cannot seek relief for anyone — not for himself, and not for any other member of the class.'" Cooper v. Southern Co., 205 F.R.D. 596, 605 (N.D. Ga. 2001) (quoting Griffin, 823 F.2d at 1483). "Under elementary principles of standing, a plaintiff must allege and show that he personally suffered injury." Griffin, 823 F.2d at 1482. Each claim asserted by a named plaintiff "'must be analyzed separately, and a claim cannot be asserted on behalf of a class unless at least one named plaintiff has suffered the injury that gives rise to that claim.'"Prado-Steiman ex rel. Prado v. Bush, 221 F.3d 1266, 1280 (11th Cir. 2000) (quoting Griffin, 823 F.2d at 1483).

Despite the Eleventh Circuit's emphasis on standing, plaintiffs' Brief fails to address it. Defendant's Brief addresses the issue and asserts that plaintiffs do not have standing to pursue several of the claims included in their proposed class definition, and that no plaintiff has standing to assert denial of promotion to management. Def.'s Br. in Supp. of Resp. in Opp. to Pls.' Mot. for Class Certification [96] at 7-9.

Although some plaintiffs make individual claims not asserted on behalf of the class,⁶¹ plaintiffs assert in the Amended Complaint (see supra Part II) that Cracker Barrel engages in systemic, class-wide discrimination by "channeling" African-American applicants to back-of-the-house positions, and then keeping them there through other discriminatory selection policies and practices. Plaintiffs have elaborated upon their channeling claim in their Brief as follows:

61 For example, plaintiff Wilson alleges constructive discharge, individual disparate treatment, and hostile work environment; plaintiff Regan alleges hostile work environment; and plaintiffs Barbee and Keel allege individual disparate treatment.

Cracker Barrel's selection procedures have disproportionately steered or channeled African-Americans away from front-of-the-house and managerial positions, leaving them to work primarily in the back-of-the-house cooking and cleaning positions traditionally reserved for their race.

. . . .

Cracker Barrel has obviously structured its employment practices to maximize the cumulative adverse effect of its channeling practices, first reducing the number of African-Americans in the front-of-the-house positions that interact with customers at the restaurants and then compounding the effect of such channeling by reducing the number of African-Americans who are eligible to compete for managerial, PAR Level and cross-training opportunities through the requirements of prior experience or training in front-of-the-house positions.

. . . .

[T]he class representatives in this case claim that Cracker Barrel's selection process "steers" or "channels" African-Americans to back-of-the-house positions such as Dishwasher at hire, that this practice caused them not to be selected for the front-of-the-house, and that this in turn caused them to lack the requisite service and PAR level which would have led to their advancement to managerial training and positions. In order to prove their individual claims, each of these class representatives will necessarily have to prove the class claim that African-Americans are disproportionately "channeled" into back-of-the-house positions because such "channeling" is inherently dependent upon proof of a pattern of selection decisions that go beyond just one individual situation.

Pls.' Br. in Supp. of Class Certification [82] at 2, 22, 29-30.

As is clear from the above-quoted excerpts, the foundation of plaintiffs' class-wide claims is the contention that Cracker Barrel channels African-American applicants at hire and initial job assignment to undesirable, back-of-the-house positions traditionally reserved for their race. Then, through application of other discriminatory policies, Cracker Barrel allegedly prevents its African-American employees from obtaining cross-training, PAR levels, and experience necessary to obtain front-of-the-house and/or management positions. Thus, plaintiffs identify the following four areas of alleged discrimination:

• Denial of front-of-the-house positions to African-American applicants at hire or channeling them to back-of-the-house jobs;

• Denial of cross-training or transfer that would allow African Americans to move from back-of-the-house jobs to front-of-the-house jobs and obtain the experience necessary to obtain management positions;

• Denial or delay in obtaining PAR levels or PAR tests that would allow African Americans to move to from back-of-the-house jobs to front-of-the-house jobs and obtain the experience necessary to obtain management positions; and

• Denial or delay in obtaining management training or positions.

Pls.' Br. in Supp. of Class Certification [82] at 40.

As noted above, a claim cannot be asserted on behalf of the class unless at least one named plaintiff has suffered the injury that gives rise to that claim. When the Court applies this rule of law to the facts here, it is clear that no representative plaintiff has standing to assert many of the claims alleged as part of the expansive class definition (quoted supra). For example, no putative class representative was deterred from applying for employment at Cracker Barrel. Therefore, no putative class representative can represent anyone who allegedly was so deterred. Cf. Mauldin v. Wal-Mart Stores, Inc., No. 1 :01-CV-2755-JEC at 16-17, 25 (N.D. Ga. Aug. 23, 2002) (excluding from class definition persons who "wished to use" prescription contraceptives).

Additionally, class members must be back-of-the-house, non-salaried, nonmanagement employees to have suffered the channeling injuries claimed to be represented by the putative plaintiffs. According to basic standing analysis, only plaintiffs who personally experienced the injuries can represent or be represented by a class in a suit; "'[f]eelings of solidarity do not confer standing to sue.'" Griffin, 823 F.2d at 1483-84 (quoting Minority Police Officers Ass'n v. City of South Bend, 721 F.2d 197, 202 (7th Cir. 1983)). Here, persons employed in salaried or management positions or in another part of the Cracker Barrel system (such as its corporate headquarters or distribution center) could not have been injured in the same way as members of a class whose injuries are based on denial of those types of positions and channeling into jobs only available at the restaurants. Therefore, the class representatives can only represent hourly restaurant employees.

With regard to the four primary claims made on behalf of the class, upon review of the facts regarding each putative representative (see supra Part III.B.), the Court concludes that the following plaintiffs have standing to assert the following claims on behalf of the proposed class:

Denial of front-of-the-house positions to African-American applicants at hire or channeling them to back-of-the-house jobs:

Plaintiff Workue.

Denial of cross-training or transfer that would allow African Americans to move from back-of-the-house jobs to front-of-the-house jobs and obtain the experience necessary to obtain management positions:

Plaintiffs Wilson, Keel and Alexander.

Denial or delay in obtaining PAR Levels or PAR tests that would allow African Americans to move from back-of-the-house jobs to front-of-the-house jobs and obtain the experience necessary to obtain management positions:

Plaintiffs Rhodes, Wilson, Regan, Barbee, and Keel.

Denial or delay in obtaining management training or positions:

Plaintiffs Regan and Alexander.

Therefore, the Court concludes that at least one putative representative has standing to raise each class-wide claim asserted. Additionally, because several plaintiffs remain employed by Cracker Barrel, they have standing to seek injunctive relief. Reid v. Lockheed Martin Aeronautics Co., 205 F.R.D. 655, 665 (N.D. Ga. 2001) (citingWooden v. Bd. of Regents of Univ. Sys. of Ga., 247 F.3d 1262, 1283 (11th Cir. 2001)).

The Court notes, however, that no putative representative has standing to bring the very claim that underpins their cause of action — channeling at hire and initial job assignment. With the exception of Ms. Workue, all of the putative representatives either sought only back-of-the-house jobs (so they could not have been channeled there) or failed to present any evidence that they only accepted back-of-the-house jobs because a Cracker Barrel manager lied about the availability of front-of-the-house jobs.

Plaintiffs assert that Cracker Barrel's alleged failure to hire African-American applicants like Ms. Workue for front-of-the-house positions is part of this alleged channeling practice. Assuming that channeling encompasses a failure-to-hire claim, the Court found above that Ms. Workue has standing to assert a channeling claim. However, the Court has serious concerns about recommending certification of a class premised on a channeling claim when only one putatively representative plaintiff (Ms. Workue) has standing to assert it (and even that is tenuous).

The Court also has serious concerns about the strength of this putative representative's claim. As shown in Ms. Workue's factual summary (see supra Part III.B.6.), she falsified her employment applications or failed to complete them, and has no proof that White applicants were even hired for the server jobs she sought, or that if Whites were hired, she had better qualifications. Moreover, a legitimate, non-discriminatory reason not to hire her would have been the difficulty of getting to work at Cracker Barrel stores in northern and/or coastal Virginia when Ms. Workue lived dozens or hundreds of miles away in the District of Columbia without her own transportation.

Although case law cited below holds that it is premature to judge the merits of a plaintiffs claim at the class certification stage, the weakness of Ms. Workue's claim cannot simply be ignored. Although Ms. Workue may technically have standing, the Court has a responsibility well articulated by Chief Judge Evans:

If a class representative with a weak case loses and individual class members with strong cases are bound by the negative outcome, an injustice will have occurred. . . . The Court's responsibility is to make sure that the common bond between the class representatives' claims and those of the class is strong enough so that it is fair for the fortunes of the class members to rise or fall with the fortunes of the class representatives. That is the very purpose of Rule 23(a).

Cooper, 205 F.R.D. at 608-09. The fortunes of the absent class members should not have to rest with so unworthy a putative class representative as Ms. Workue.

B. Procedural Requirements 1. Title VII

The procedural prerequisites to a Title VII claim include the filing of a timely charge of discrimination with the Equal Employment Opportunity Commission ("EEOC") and the filing of suit within ninety days of receipt of a notice of right to sue from the EEOC. 42 U.S.C.A. § 2000e-5(e)(1), (f)(1) (West 1994).

a. Timely Charge

The timeliness of a charge depends upon the state in which an aggrieved person lives. A person who lives in a non-deferral state must file his charge of discrimination with the EEOC within 180 days of an allegedly unlawful act. A person who lives in a deferral state must file his charge within 300 days of an allegedly unlawful act. 42 U.S.C.A. § 2000e-5(e)(1); Hipp v. Liberty Nat'l Life Ins. Co., 252 F.3d 1208, 1214 n. 2 (11th Cir. 2001) (per curiam), cert. denied, 534 U.S. 1127 (2002).⁶²

62 "Non-deferral states are those states without a state law banning . . . discrimination in employment and without a state entity authorized to grant or to seek relief for victims of such discrimination."Grayson v. K Mart Corp. 79 F.3d 1086, 1100 n. 20 (11th Cir. 1996). "Deferral states are those states having a state law banning . . . discrimination in employment and a state entity authorized to grant or to seek relief for victims of such discrimination."Id. at 1100 n. 21.

However, in the class action context, "it is not necessary for all class members to have filed EEOC charges or received notices of the right to sue in order to be represented by the class." Alexander v. Fulton County, Ga., 207 F.3d 1303, 1333 (11th Cir. 2000). "As long as at least one named plaintiff timely filed an EEOC charge, the precondition to a Title VII action is met for all other named plaintiffs and class members." Griffin, 823 F.2d at 1492. The reason for this rule is that the charge-filing requirement is designed to put the EEOC and the employer on notice of allegations of discrimination. Hipp, 252 F.3d at 1225.⁶³ If others have already filed charges, the notice requirement is satisfied, and requiring all aggrieved persons to file charges would be a "useless act." Id. This principle has come to be known as the "piggybacking" or "single filing" rule. Thus, someone "who has not filed his own EEOC charge may 'piggyback' his claim onto the claim of a plaintiff who has filed a timely charge." Grayson, 79 F.3d at 1101; see also City of Hialeah, Fla. v. Rojas, No. 00-14736, 2002 WL 31487602, at *5 (11th Cir. Nov. 8, 2002) (pursuant to the single-filing rule, as long as at least one named plaintiff timely filed an EEOC charge, the precondition to a Title VII action is met for all other named plaintiffs and class members).

63 Although cases like Grayson and Hipp arose under the Age Discrimination in Employment Act ("ADEA"), "the purposes underlying ADEA and Title VII, specifically their respective requirements that employees file charges of discrimination with the EEOC so that the employer can attempt to resolve the issue through conciliation, are similar." Hipp, 252 F.3d at 1221 n. 10. Therefore, courts look to Title VII and ADEA cases in examining these issues. Id.

The piggybacking rule has the following limitations. One may piggyback on another's charge provided that (1) the charge upon which he is piggybacking is valid, and (2) the individual claims of the filing and non-filing plaintiff arise out of similar discriminatory treatment in the same time frame. Grayson, 79 F.3d at 1101-02; Calloway v. Partners Nat'l Health Plans, 986 F.2d 446, 450 (11th Cir. 1993). With regard to the first requirement, in the class action context, the valid charge must also contain allegations of class-wide discrimination. Such allegations provide notice both to the EEOC and the employer that the discrimination alleged is more than an isolated act. Grayson, 79 F.3d at 1107.

With regard to the second requirement, the temporal scope of the class action is further refined. The rearward temporal scope of a Title VII class is limited to those persons who allege discriminatory treatment within 180 or 300 days (depending on their state of residence) before the representative charge was filed. Hipp, 252 F.3d at 1220. The forward temporal scope of a Title VII class ends on the date the representative charge is filed. Id. at 1225.

Plaintiffs argue that the rearward temporal scope of any potential class extends two years before the earliest charge, or back to April 9, 1994, citing 42 U.S.C. § 2000e-5 (g), which provides: "Back pay liability shall not accrue from a date more than two years prior to the filing of a charge with the Commission." Pls.' Mot. for Class Certification [82] at 2-3. (They do not address the forward temporal scope.) However, "the two year cap on back pay contained in 42 U.S.C. § 2000e-5 (g) is not a statute of limitations. Rather, that provision was inserted by Congress in an attempt to limit the back pay which could be recovered from employers who have been engaged in discrimination for many years." Miller v. Miami Prefabricators, 438 F. Supp. 176, 181 (S.D. Fla. 1977). The temporal scope of any Title VII class is, therefore, governed by the standards discussed above.⁶⁴

64 Accordingly, any class certified under Title VII would extend back to June 14, 1995. Any class certified under § 1981 would extend back to July 30, 1997. See infra Part V.B.2.

An exception to the above-mentioned temporal scope rules occurs when a representative charge alleges a continuing illegal policy. Hipp, 252 F.3d at 1226 (citing Bush v. Liberty Nat'l Life Ins. Co., 112 F. Supp.2d 1251, 1258-59 n. 11 (M.D. Ala. 1998), aff'd mem., 196 F.3d 1261 (11th Cir. 1999)). The Eleventh Circuit did not, however, decide in Hipp how the temporal scope of a class action is affected by a continuing-violation allegation. 252 F.3d at 1208.

The two representative charges here allege that the discrimination "continues." See Attach. to Pls.' Mot. for Class Certification [82] (Workue Charge at 2; Morrow Charge at 2). Moreover, the Amended Complaint alleges that Cracker Barrel's discriminatory practices are continuing and that equitable relief is necessary to stop them and to undo their effects. See Am. Compl. [92] ¶¶ 3, 26, 31, 34, 145-46. However, the parties have neither briefed the complex continuing violation doctrine nor addressed the effect of the Supreme Court's recent decision in National Railroad Passenger Corp. v. Morgan, 536 U.S. 101 (2002), on that doctrine. Like the Hipp court, this Court need not decide this issue given the recommendation that class certification be denied. 252 F.3d at 1208. Should the undersigned's recommendation be rejected on review, however, the parties will need to address the temporal scope of any class and the impact of the continuing violation doctrine.

b. Filing Suit Within Ninety Days

The person on whose charge others seek to piggyback must not only file a timely charge, but must also file a lawsuit within ninety days of receipt of a notice of right to sue from the EEOC. 42 U.S.C.A. § 2000e-5(f)(1); Alexander, 207 F.3d at 1333; Oatis, 398 F.2d at 499 (plaintiff who has satisfied Title VII's jurisdictional requirements may represent class members who have not); Inda v. United Air Lines, 565 F.2d 554, 559 (9th Cir. 1976) ("[T]imely filing of an EEOC charge is not a necessary condition to the obtaining of relief by one as a member of a class . . . However, if one brings suit on his own behalf, or as a named plaintiff on behalf of a class, he must have secured a right to sue by timely following the procedures set forth in Title VII.").

c. The EEOC Charges

One putative representative, Ms. Kebere Workue, and one named plaintiff, Ms. Bessie Morrow, submitted charges alleging class-wide discrimination. Ms. Workue, an unsuccessful applicant, filed her charge with the EEOC on April 9, 1996, in Washington, D.C. (a deferral "state"). The charge arose out of her failure to be hired at Cracker Barrel locations in neighboring Virginia (also a deferral state) over the period from July through October 1995. She alleges that Cracker Barrel discriminated and continues to discriminate against potential applicants and applicants for employment with respect to hiring, job assignments, and job classification, and maintains racially segregated job classifications. Attach. to Pls.' Mot. for Class Certification [82]. Because this charge was filed within 300 days after the last alleged discriminatory act in October 1995, it is timely.

The EEOC mailed its Notice of Right to Sue to Ms. Workue on March 31, 1999, at her Washington, D.C., address. Workue Dep. at 293-95 and Dep. Ex. 8 (right-to-sue notice). The record does not indicate when Ms. Workue received it. However, the parties stipulated that the ninety-day period for filing suit was tolled through July 30, 1999 [120]. Therefore, suit was timely instituted on the Workue charge.

Ms. Morrow signed her charge on May 12, 1996,⁶⁵ with the EEOC in Alabama (a non-deferral state). She alleges that she was denied a promotion in October 1994, that she observed other qualified African-Americans who were overlooked for promotion to better paying management and non-management positions, and that she became discouraged from applying for promotion. She claims that Cracker Barrel's treatment of her was part of a continuing practice of race discrimination with respect to hiring, job assignment, job classification, compensation, transfer, training, promotion, discharge, and retention. She claims that the date of the most recent discriminatory act against her was April 15, 1996 (although she does not specify the act). Attach. to Pls.' Mot. for Class Certification [82]. Assuming that an allegedly discriminatory act occurred on April 15, 1996, Ms. Morrow's charge was timely filed within 180 days of the most recent allegedly discriminatory act.

65 Unlike the Workue charge, the Morrow charge reflects no "file stamp" date from the EEOC. The parties have, however, apparently agreed on the May 12, 1996, date. The Court will accept this date, but notes that the Eleventh Circuit holds that the proper date is "either proof of filing with the state agency in a deferral state or proof of filing with the EEOC in a non-deferral state." Grayson, 79 F.3d at 1102 n. 30.

The EEOC mailed its Notice of Right to Sue to Ms. Morrow on March 31, 1999. See Morrow Notice of Right to Sue [120 at Appendix A]. Again, there is no indication in the record as to when Ms. Morrow received it. However, the parties stipulated that the ninety-day period for filing suit was tolled through July 30, 1999 [120]. Therefore, suit was timely instituted on the Morrow charge.

Although the two class-wide charges relied upon by plaintiffs meet the procedural requirements of Title VII, the more problematic question is whether each representative plaintiff may piggyback on either of them.⁶⁶ in other words, does each putatively representative plaintiff allege discrimination against him/her during the 180/300 day period (depending on the state of residence) before the filing of either charge? The rearward scope of Ms. Workue's April 9, 1996, charge, is as follows:

66 As for potential class members who did not file charges, Cracker Barrel asserts that only unsuccessful applicants should be allowed to piggyback on Ms. Workue's charge, but does not object to employees and former employees piggybacking on Ms. Morrow's charge. Def.'s Resp. in Opp. to Pls.' Mot. for Class Certification [96] at 2. Given the Court's recommendation that class certification be denied, it is unnecessary to resolve this dispute.

October 12, 1995 (180 days before charge filed/non-deferral state) or June 14, 1995 (300 days before charge filed/deferral state)

The rearward scope of Ms. Morrow's May 12, 1996, charge is as follows:

November 14, 1995 (180 days before charge filed/non-deferral state) or July 17, 1995 (300 days before charge filed/deferral state)

If the putatively representative plaintiffs allege that they were victims of discrimination between these relevant rearward dates (depending on whether they live in deferral or non-deferral states) and the date of the filing of either charge, then they can piggyback.

Unfortunately, the six, non-charge-filing putative representatives have not been precise in alleging when they were victimized by alleged discrimination. Indeed, the Amended Complaint and the various Declarations signed by plaintiffs are ambiguous about when most alleged events occurred. Plaintiffs' Motion for Class Certification and supporting Briefs fail to address this critical point.⁶⁷ The Court has reviewed the massive file in an effort to associate dates with the numerous allegations where possible, but judges "are not like pigs, hunting for truffles buried [in the record]." United States v. Dunkel, 927 F.2d 955, 956 (7th Cir. 1991).

67 Plaintiffs likely did not address the issues of timeliness of the charges and the statute of limitations because defendant did not raise them. Although defendant asserted these issues as affirmative defenses,see Answer [96] ¶¶ 2-7, plaintiffs could argue that defendant waived them. In any event, regardless of whether the parties addressed these issues, Griffin v. Dugger requires this Court to do so. 823 F.2d at 1482.

Because the undersigned recommends that class certification be denied, it is not necessary to decide this issue. The undersigned assumes that the putatively representative plaintiffs allege discrimination against them during the 180/300 day period (depending on their state of residence) before the filing of the Workue or Morrow charges. Should the undersigned's recommendation be rejected on review, however, the putative representative plaintiffs will need to show when they were victimized by alleged race discrimination to determine whether they may piggyback on the Workue or Morrow charges. Furthermore, as previously noted, if any representative plaintiffs were allegedly victimized by race discrimination before the rearward or after the forward temporal scope of a class piggybacking on these charges, then the parties may need to address the impact of the continuing-violation doctrine.

2. § 1981 Statute of Limitations

The procedural prerequisites to filing suit under Title VII do not apply to claims under § 1981. Johnson v. Railway Express Agency, Inc., 421 U.S. 454, 461 (1975). Section 1981 does not contain its own limitations period. However, in Wilson v. Garcia, the Supreme Court held that claims under 42 U.S.C. § 1983 are best characterized as personal injury actions; therefore, the forum state's personal injury statute of limitations applies. 471 U.S. 261, 276 (1985). See Rojas, 2002 WL 31487602, at *7 n. 2 ("Section 1983 claims are governed by the forum state's residual personal injury statute of limitations."). In Goodman v. Lukens Steel Co., the Supreme Court extended that reasoning to § 1981 claims. 482 U.S. 656, 661 (1987). For § 1981 claims filed in Georgia, the two-year statute of limitations governing personal injury actions applies. Brown v. E. Cent. Health Dist., 752 F.2d 615, 618-19 (11th Cir. 1985); Ga. Code Ann. § 9-3-33 (2002).

Plaintiffs did not address the statute of limitations issue in their Motion or supporting Briefs. See supra note 67. Had plaintiffs argued for application of the four-year statute of limitations found in 28 U.S.C. § 1658, it would have been rejected. See NAACP v. Cracker Barrel Old Country Store, Inc., No. 4:01-CV-325-HLM (N.D. Ga. filed Aug. 19, 2002) at 38-45. In any event, plaintiffs filed this action on July 30, 1999. For their claims to be timely under § 1981, plaintiffs had to suffer race discrimination during the preceding two years (i.e., on or after July 30, 1997). As previously noted, the Amended Complaint and Declarations of the plaintiffs fail to make specific allegations concerning when discrimination occurred. Because the undersigned recommends that class certification be denied, it is unnecessary to decide this issue. The Court assumes that plaintiffs have satisfied the § 1981 statute of limitations. Should the undersigned's recommendation be rejected on review, however, the limitations issue may need to be revisited.

C. Class Certification Under Rule 23

Class certification is appropriate where maintenance as a class action constitutes the most efficient means to remedy employment discrimination. Gen. Tel. Co. v. Falcon, 457 U.S. 147, 159 (1982) (efficiency and economy are principal purposes of class action procedure). The burden of establishing the specific prerequisites to a Rule 23 action falls on those seeking certification. Hudson v. Delta Air Lines, Inc., 90 F.3d 451, 456 (11th Cir. 1996).

"When presented with a class certification motion, the Court first determines whether Plaintiffs have satisfied the requirements of Rule 23(a), and then proceeds to verify that the proposed class falls within one of the categories described in Rule 23(b)." Anderson v. Garner, 22 F. Supp.2d 1379, 1383 (N.D. Ga. 1997) (Murphy, J.). The Court does not inquire whether plaintiffs have sufficient evidence to prevail on the merits of their claim. Id. at 1382-83. Nevertheless, the Court must conduct a "rigorous analysis" of the arguments supporting class certification. Falcon, 457 U.S. at 161 (noting significant concern for "the potential unfairness to the class members bound by the judgment if the framing of the class is overbroad" and/or the representation of the class inadequate). The Court is not limited to the pleadings themselves, but may probe behind them and review evidence relevant to the class certification issue. Id. at 160. "'Indeed, to protect absent class members and defendants, this Court has an affirmative obligation to 'resolve factual and legal disputes that strongly influence the wisdom of class treatment.'" Abram, 200 F.R.D. at 427 (quoting Szabo v. Bridgeport Machs., Inc., 249 F.3d 672, 675 (7th Cir. 2001)).

Employment discrimination class actions have been a significant aid in moving toward Title VII's goal of a color-blind society, especially in the early days of the statute "where employers completely excluded members of minority groups from positions and other job benefits at any but the lowest levels of the company." Reid, 205 F.R.D. at 660. "As the culture has changed, however, and new statutes have been enacted, it has become apparent that class certification is no longer a nearly automatic outcome." Id. (citing Falcon, 457 U.S. at 161). The Court must ensure that representative plaintiffs carry their burden of proving that they have standing, that they have complied with the procedural prerequisites of Title VII, and that they meet the elements of Rule 23. Certifying a class that fails to meet these multiple requirements gains nothing "except the blackmail value of a class certification that can aid the plaintiffs in coercing the defendant into a settlement." Rutstein v. Avis Rent-A-Car Sys., 211 F.3d 1228, 1241 n. 21 (11th Cir. 2000), cert. denied, 532 U.S. 919 (2001).

1. Rule 23(a)

Rule 23(a) requires plaintiffs to show the following:

(1) the class is so numerous that joinder of all members is impracticable;

(2) there are questions of law or fact common to the class;

(3) the claims or defenses of the representative parties are typical of the claims or defenses of the class; and

(4) the representative parties will fairly and adequately protect the interests of the class.

Fed.R.Civ.P. 23(a). The Court considers each criterion in turn.

a. Numerosity

The first requirement of Rule 23(a) is that the class be so large that joinder of all members is impracticable. Fed.R.Civ.P. 23(a)(1). "Practicability of joinder depends on many factors, including, for example, the size of the class, ease of identifying its numbers and determining their addresses, facility of making service on them if joined and their geographic dispersion." Kilgo, 789 F.2d at 878. When a class is extremely large, the numbers alone may presume that joinder is not feasible. Buford v. H R Block. Inc., 168 F.R.D. 340, 348 (S.D. Ga. 1996), aff'd sub nom. Jones v. H R Block Tax Servs., 117 F.3d 1433 (11th Cir. 1997). Plaintiffs contend that the size of the class could be several thousand people. Am. Compl. [92] ¶ 32. Defendant does not contest this point. Therefore, the numerosity requirement is met.

b. Commonality and Typicality

Commonality does not require that all questions of fact and law raised in the action be common. However, "[t]he claims actually litigated in the suit must simply be those fairly represented by the named plaintiffs."Cox v. Am. Cast Iron Pipe Co., 784 F.2d 1546, 1557 (11th Cir. 1986). Analysis of the typicality requirement is similar to analysis of the commonality requirement, but there is a stronger focus on the representative plaintiffs. Prado-Steiman, 221 F.3d at 1279 ("commonality refers to the group characteristics of the class as a whole and typicality refers to the individual characteristics of the named plaintiff in relation to the class"). There is an interrelationship between the commonality and typicality requirements:

The commonality and typicality requirements of Rule 23(a) tend to merge. Both serve as guideposts for determining whether under the particular circumstances maintenance of a class action is economical and whether the named plaintiffs claim and the class claims are so interrelated that the interests of the class members will be fairly and adequately protected in their absence.

Falcon, 457 U.S. at 157 n. 13.

Judge Forrester's comments about the District Court's task in reviewing commonality and typicality are pertinent here:

Although the Court does not at this stage examine the merits of Plaintiffs' case, it may properly consider evidence relating to the discrimination allegedly suffered by the class, well as the manner in which the evidence fits into the legal framework governing the claims pled in Plaintiffs' complaints. Nelson v. United States Steel Corp., 709 F.2d 675, 679 (11th Cir. 1983). Commonality requires that "a class action . . . involve issues that are susceptible to class-wide proof," Murray v. Auslander, 244 F.3d 807, 811 (11th Cir. 2001), and the court must review the evidence to determine whether this requirement is satisfied. Moreover, "evidence relevant to the commonality requirement is often intertwined with the merits." Nelson, 709 F.2d at 679. In addition, with regard to typicality, the court must be able to determine whether the named Plaintiffs "possess the same interest and suffer the same injury as the class members." Murray, 244 F.3d at 811. In short, Plaintiffs' burden at this stage "entails more than the simple assertion of [commonality and typicality] but less than a prima facie showing of liability." Nelson, 709 F.2d at 680.

Reid, 205 F.R.D. at 666. For the reasons explained below, the Court finds both of these necessary elements lacking here.

(1) The Representative Plaintiffs Have Little In Common With Each Other or the Proposed Class

As noted above, the class representatives must "'possess the same interest and suffer the same injury as the class members.'" Reid, 205 F.R.D. at 666 (quoting Murray, 244 F.3d at 811). Moreover, the Court must focus on whether the named representatives' claims have the same essential elements as the claims of the class at large. Abram, 200 F.R.D. at 428.

However, significant differences exist between the putative class representatives and the potential class members. According to the proposed class definition and asserted claims, the potential class member would be an African American who experienced "channeling" into a back-of-the-house position, was involuntarily kept there through lack of opportunity to cross-train and/or progress through PAR levels to management positions, and was subject to discriminatory pay. Plaintiffs allege nothing about tenure, but the workforce description provided by defendant's experts shows that the average tenure of a Cracker Barrel employee is seventeen weeks. H/T Report at 22-23.

A review of each representative plaintiffs individual claims (see supraPart III.B.) shows that none of them experienced the channeling at hire and initial job assignment, which they allege is the cornerstone of the claims of the class, because all who were hired applied for back-of-the-house jobs. Ms. Workue is the only plaintiff who could be said to have been denied a front-of-the-house position, because she was never hired at all. However, as discussed previously, in addition to significant omissions and errors in her applications, Ms. Workue could not state whether anyone else was hired instead of her nor what the qualifications were of anyone who might have been hired. Therefore, Ms. Workue is tenuously the sole representative of a class of people who, but for racial discrimination, would or should have been hired into front-of-the-house jobs on the basis of their qualifications.

Although three of the plaintiffs allege denial of cross-training for front-of-the-house jobs,⁶⁸ the other three never sought cross-training in front-of-the-house jobs.⁶⁹ While five of them complain of issues with PAR testing,⁷⁰ their complaints relate to tests for back-of-the-house jobs, not problems with obtaining PAR tests for front-of-the-house positions. Moreover, when one examines each plaintiff's situation, it appears that they all progressed through PAR rapidly in the back-of-the-house positions they selected. Five of the six plaintiffs who actually worked for Cracker Barrel are at PAR IV.⁷¹ Their pay was either the best or among the best in the store in which they worked. Those five are still employed and have worked there, on average, for eleven years. Finally, although two complain about denial of management opportunities,⁷² most indicated no desire for this career path.⁷³

68 Wilson, Keel, and Alexander.

69 Rhodes, Regan, and Barbee.

70 Rhodes, Wilson, Regan, Barbee, and Keel.

71 Rhodes, Regan, Barbee, Keel, and Alexander.

72 Regan and Alexander.

73 E.g., Barbee, Keel, and Rhodes.

As noted above, the class plaintiffs describe is one where African Americans are involuntarily placed in back-of-the-house jobs, who cannot cross-train or progress through the PAR levels, and who are poorly paid. Thus, plaintiffs do not appear to possess the same interest or to have suffered the same injuries as the class members, and their claims lack the essential elements of the claims of the class.

In fact, the representative plaintiffs assert largely individualized claims, rather than claims that are common among themselves. One asserts channeling only from one back-of-the-house position to another (Alexander); another that he was denied the opportunity to cross-train for back-of-the-house positions (only), though he was hired as a dishwasher and is now a grill cook (Barbee); another that she was discriminatorily disciplined and given an unequal work schedule (Keel); another that he was denied management training (Regan); another that he was channeled from grill cook to dishwasher (both back-of-the-house) (Rhodes); another that he was denied a night maintenance position (Wilson); and another that she was not hired because of her race (Workue). The extent of commonality among the representatives lies solely in two plaintiffs who were both channeled between back-of-the-house positions. This lone mutual claim neither supports the class definition presented by plaintiffs nor suggests that any class certification would be warranted when only two members emerge with a common claim from the group of supposed exemplars.

(2) Decentralized Decision-Making At Cracker Barrel Undermines Commonality and Typicality

Cracker Barrel publishes extensive guidelines and policies on, inter alia, equal employment, recruitment, and hiring, and utilizes persons employed by the corporate office to assist with new store openings and to conduct training of local store management personnel in those guidelines and policies. There are also certain company-wide minimum standards that employees must meet to be eligible for PAR testing or promotion to management.

However, there is no central person or group at Cracker Barrel's corporate office that applies its human resource guidelines and policies to every store by deciding whom to interview, whom to hire, or whom to promote through PAR, cross-train, and promote to management. Instead, it is undisputed that Cracker Barrel's centrally-created employment policies are applied by hundreds or perhaps thousands of relatively autonomous decision-makers scattered over 450 stores in forty-one states. These local store managers (general, assistant and/or associate) decide who is interviewed, who is hired and in what job, who is allowed to transfer, and who is allowed to cross-train. They set starting hourly rates, conduct evaluations (which determine the right to take PAR tests), and award hourly rate increases. No more decentralized system of personnel decision-making could be imagined. If these local managers engaged in race discrimination in making the employment decisions entrusted to them, then they violated Company policy. Not only is racial discrimination against Company policy, but managers are also trained to act without regard to race. Employees have avenues to complain if they are mistreated. See supra Part III.A.1.e.

Because each class representative was aggrieved by some alleged misconduct by his or her own manager(s), there could be little in common among their claims. See supra Part III.B. For the same reason, the class representatives could have little in common with class members spread over so many stores in so many states with so many different managers responsible for making the challenged employment decisions. One manager's decision about the employment status of a representative plaintiff working at a store in Savannah, Georgia, or Lake City, Florida, would obviously have a different rationale than another manager's decision about the employment status of a prospective class member working at a store in Houston, Texas, or Manassas, Virginia.

The circumstances of each putative representative's case will depend on how a specific manager(s) treated him in his store. In turn, defendant will no doubt assert specific, non-discriminatory reasons why the local manager made that decision. Each plaintiffs claim will depend on an individualized factual inquiry. Multiply that inquiry by hundreds or thousands of potential class members — because each of their claims will depend on what a local manager did or did not do — and the complexity of this case quickly escalates with no commonality between the representative plaintiffs and the class and no typicality between the claims of the representative plaintiffs and the claims of the class.

The plaintiffs list eight cases that they assert stand for variations of the principle that, for typicality, "it is sufficient that the claims of the named Plaintiffs are substantially similar to the claims of the class, despite factual variations."⁷⁴ Pl.'s Br. [82] at 42-43. A review of the cases shows that, in contrast to the case at hand, all concern a primary, overt policy⁷⁵ or event⁷⁶ was unifying and largely determinative. The singularity of the policy or event made circumstantial differences among individual plaintiffs less significant.See, e.g., Appleyard v. Wallace, 754 F.2d 955, 959 (11th Cir. 1985) ("In this case, however, the challenged rule is the sole basis upon which the plaintiffs were denied Medicaid nursing home benefits."). Indeed, in one, the plaintiffs did not even ask for individual damages, id.; in another, the court excepted damages from certification because of individuality issues. Murray v. Auslander, 244 F.3d 807 (11th Cir. 2001).

74 Plaintiffs assert that this quotation comes from Cox, 784 F.2d at 1557; however, this statement and the rest of the material allegedly quoted from Cox do not appear in the opinion. The quote in fact comes from Drayton v. Western Auto Supply Co., 203 F.R.D. 520, 527 (M.D. Fla. 2000), rev'd in part, No. 01-10415, 2002 WL 518017 (11th Cir. Mar. 11, 2002).

75 Cox, 784 F.2d 1546 (overt policy of workforce gender segregation); Murray v. Auslander, 244 F.3d 807 (11th Cir. 2001) (single, overt state policy of "capping" Medicaid benefits); Piazza v. EBSCO Ind., Inc., 273 F.3d 1341 (11th Cir. 2001) (single practice of defendant operating competing companies, injuring members of profit sharing plan of defendant); Appleyard v. Wallace, 754 F.2d 955 (11th Cir. 1985) (Alabama's Medicaid level of care admission criteria);Drayton, 203 F.R.D. at 526-27 ("Plaintiffs are not bringing an across-the-board discrimination case. At the heart of the class' claim in this case is Defendants' alleged overriding, company-wide policy of employment segregation, namely the staffing of stores according to the racial makeup of their clientele. . . . [Defendant's] Regional Vice President . . .acknowledged the existence of this segregationist policy in his deposition."); Caridad, 191 F.3d at 291 ("The Class Plaintiffs . . . argue that it is this policy — the delegation to supervisors, pursuant to company-wide policies, of discretionary authority without sufficient oversight — that gives rise to common questions of fact warranting certification."), remanded to 197 F.R.D. 85 (S.D.N.Y. 2000), vacated by 267 F.3d 147 (2d Cir. 2001), cert. denied, ___ U.S. ___, 122 S.Ct. 1349 (2002).

76 Anderson v. Garner, 22 F. Supp.2d 1379 (N.D. Ga. 1997) (Murphy, J.) ("Shakedowns" at Hays State Prison on July 10, 1996); Kornberg v. Carnival Cruise Lines, Inc., 741 F.2d 1332 (11th Cir. 1984) (failure of the sanitary system during one-week cruise).

In this case, however, plaintiffs allege that Cracker Barrel engages in various discriminatory practices — in its "selection and compensation procedures, racially hostile reputation and working conditions, and unequal terms and conditions of employment" — that "channel" African-American applicants to back-of-the-house positions at hiring and then keep them there through practices involving PAR testing, cross-training and management training. Am. Compl. [92] ¶ 28. Rather than a singular, overt, centralized policy, plaintiffs point to multiple practices, primarily occurring at 450 stores in forty-one states by a multitude of local decision-makers. Therefore, a tangible unifying principle, which predominates over factual variations in the cases cited by plaintiffs, does not exist here.

It is not surprising, then, that other courts faced with employment decision-making far less decentralized than here have found commonality and/or typicality lacking. For example, in Webb v. Merck Co., African-American plaintiffs, both current and former employees of Merck, sought certification of a class under Title VII and § 1981 to challenge initial assignments, performance reviews, job postings, promotions, and compensation. 206 F.R.D. 399 (E.D. Pa. 2002). The proposed class included all African Americans employed or formerly employed at every level and grade in two corporate divisions at six separate facilities located in five separate states. Id. at 401. After review of each named plaintiff's situation, the court found that resolution of their claims would depend on their own discrete factual scenarios, and that Merck had defenses unique to each plaintiffs claims. The court stated that it could "only imagine how these factual disparities will be exponentially increased by the addition of at least some 5000 plaintiffs." Id. at 405. Because the claims were not susceptible to generalized proof or defenses, the fact-finder would have to consider each of the named plaintiffs claims on a case-by-case basis along with Merck's unique defenses to each claim. "[R]esolution of the merits of the claims would degenerate into an unmanageable plethora of multiple individual determinations for each individual proposed class member."Id. at 406. Although there were corporate-wide employment policies, the claims arose out of "individual decisions made by hundreds if not thousands of individual managers at all of these organizations all of whom had varying degrees of autonomy over compensation and promotion decisions." Id. The same reasoning holds true here, where, rather than six operations in five states, the decisions by hundreds or thousands of autonomous managers at over 450 stores in forty-one states are at issue.

The following cases are in accord with Webb in rejecting class certification where there was decentralized employment decision-making over separate facilities located in different geographic locations. The resulting need for individualized inquiries into each employment decision by each manager undermined commonality and/or typicality. See Bradford v. Sears, Roebuck Co., 673 F.2d 792, 795 (5th Cir. 1982) (reversing class certification in part because the employment practices complained of occurred in different retail stores where "different staffs supervise the stores and oversee the personnel decisions"); Stastny v. S. Bell Tel. Tel. Co., 628 F.2d 267, 278-80 (4th Cir. 1980) (where promotion and pay decisions entrusted to scores of different supervisors and managers at twenty-four geographically-dispersed locations in North Carolina, commonality lacking); Bacon v. Honda of Am. Mfg., Inc., 205 F.R.D. 466, 478 (S.D. Ohio 2001) (refusing to certify proposed class encompassing employees in thirty-nine departments at four separate Ohio manufacturing plants when employment decisions made by different coordinators and managers working in autonomous departments); Donaldson v. Microsoft Corp., 205 F.R.D. 558, 566-67 (W.D. Wash. 2001) (refusing to certify class where potential class members located throughout United States, engaged in wide variety of jobs, and evaluated by different decision-makers applying differing standards); Reid, 205 F.R.D. at 667-70 (refusing to allow named plaintiffs who all worked at one facility to represent persons employed at other company facilities; although corporate-wide employment policies existed, facilities had autonomy to implement them according to their needs); Wright v. Circuit City Stores, Inc., 201 F.R.D. 526, 540 (N.D. Ala. 2001) (class certification rejected for lack of commonality/typicality where purported class spread across fifteen states, involving 125-160 different stores, warehouses, and service centers, each of which had varying degrees of autonomy over hiring, evaluation, assignments, training, and promotion decisions, with those decisions being made by many different managers); Abram, 200 F.R.D. at 432 (given decentralized nature of employer's decision-making with regard to pay, the decisions of apparently hundreds of managers are implicated, requiring many individualized inquiries that no "only undermine 'commonality' but also have the potential to render class treatment unworkable"); Lott v. Westinghouse Savannah River Co., 200 F.R.D. 539, 554-56 (D.S.C. 2000) (management of employment practices, although facilitated by central Human Resources Division, entrusted to thousands of managers who exercised independent judgment; because such decentralization leads to individualized inquiries, commonality lacking).⁷⁷

77 See also Zachery v. Texaco Exploration Prod., Inc., 185 F.R.D. 230, 238-40 (W.D. Tex. 1999) (refusing to certify multi-facility class where proposed class consisted of seventeen business units, spread across fifteen states, each with varying degrees of local autonomy over employment decisions); Betts v. Sundstrand Corp., No. 97 C 50188, 1999 U.S. Dist. LEXIS 9743, at *18-23 (N.D. Ill. June 21, 1999) (lack of centralized hiring decision-maker, sheer number of managers entrusted with hiring decisions, and wide range of jobs in proposed class, show no commonality/typicality) Troupe v. Randall's Food Drugs, Inc., No. 3:98-CV-2462-P, 1999 U.S. Dist. LEXIS 11620, at *13 (N.D. Tex. July 28, 1999) (class certification denied where proposed class included employees and applicants in fifty stores located in two cities and their outlying suburbs, and where personnel decisions made in localized and decentralized manner); Bostron v. Apfel, 182 F.R.D. 188, 195-96 (D. Md. 1998) (certification of nationwide class rejected where facilities were widely dispersed, and there was substantial autonomy granted to local and regional supervisors to make employment decisions);Reyes v. Walt Disney World Co., 176 F.R.D. 654, 658 (M.D. Fla. 1998) (no commonality where plaintiffs employed by three separate departments with different decision-makers); Abrams v. Kelsey-Seybold Med. Group, Inc., 178 F.R.D. 116, 130 (S.D. Tex. 1997) (commonality and typicality lacking because "the members of the purported class were not subjected to the same decision-making authority"); Boykin v. Viacom Inc., No. 96 CIV. 8559 (DLC), 1997 U.S. Dist. LEXIS 17872, at *11-15 (S.D.N.Y. Nov. 12, 1997) (where employment decision-making decentralized and entrusted to at least fifteen different managers, resulting individualized inquiries into whether those managers engaged in discrimination defeated commonality/typicality); Appleton v. Deloitte Touche L.L.P., 168 F.R.D. 221, 231-32 (M.D. Tenn. 1996) (where most if not all employment decisions were made at local office level, certification of nationwide class denied); Zapata v. IBP, Inc., 167 F.R.D. 147, 159 (D. Kan. 1996) (absence of centralized employment decision-making or a showing of decentralization by the defendant may result in denial of class certification); Lumpkin v. E.I. DuPont De Nemours Co., 161 F.R.D. 480, 482 (M.D. Ga. 1995) (no commonality when plaintiffs worked in different departments, shifts, and job levels, were supervised by different people, and directed their demotion and promotion denial claims toward different decision-makers); Seidel v. Gen. Motors Corp., 93 F.R.D. 122, 124-25 (W.D. Wash. 1981) (denying motion for class certification where evidence demonstrated that employment decisions made at local branch level).

Like Webb and the others, the case at hand involves "individual decisions made by hundreds if not thousands of individual managers" with varying degrees of autonomy at store locations throughout the United States. 206 F.R.D. at 406. The devolution of decision-making, beyond the general guidelines and policies, and the specific circumstances for each plaintiff necessitate individualized proof and defenses incommensurate with the type of central determination for which class certification was designed to create efficiency. Indeed, the guidelines and policies of the central office explicitly decry discrimination, so only the respective practices of the various managers could be at issue. Therefore, the undersigned must recommend that class certification be denied, given the decentralized decision-making here.

(3) The Individualized Nature of the Class Claims Undermine Commonality and Typicality

Plaintiffs assert both disparate treatment and disparate impact claims. They claim that the method of proof and trial will not involve individualized evidence on the issues of class-wide liability and injunctive relief. Pls.' Br. in Supp. of Class Certification [82] at 31. However, as discussed below, this issue is not as simple as described by plaintiffs, because the individualized factual issues that will arise in the trial of either type of claim undermine commonality and typicality.

(a) Disparate Treatment

An African-American plaintiff establishes disparate treatment by showing that the employer intentionally treated him less favorably than similarly situated White employees because of his race. United States Postal Serv. Bd. of Governors v. Aikens, 460 U.S. 711 (1983). In the class-action context, plaintiffs must show that "racial discrimination was the company's standard operating procedure — the regular rather than the unusual practice." Int'l Bhd. of Teamsters v. United States, 431 U.S. 324, 336 (1977). This is commonly referred to as a "pattern or practice" case. Such cases may be brought both under Title VII and § 1981. Rutstein, 211 F.3d at 1237.

In a pattern-or-practice case, the plaintiff attempts to prove that discrimination was the employer's standard operating procedure through a combination of statistical, historical, and anecdotal evidence.Teamsters, 431 U.S. at 338-39, 361; Kilgo, 789 F.2d at 874. This is called the liability, or Stage I, proceeding. At Stage I, the focus is not on individual decisions, but rather on a pattern of discriminatory decision-making. Teamsters, 431 U.S. at 360 n. 46. If successful at Stage I, the "primary relief afforded to the plaintiff class in a pattern or practice case is declaratory or injunctive." Rutstein, 211 F.3d at 1238 n. 18 (citing Teamsters, 431 U.S. at 361).

If class members desire individual relief (as they do here), they must come forward during the remedial, or Stage II, proceeding. Rutstein, 211 F.3d at 1238 n. 18. Here, because of the poof of the pattern or practice presented at Stage I, there is an inference⁷⁸ that this pattern or practice negatively affected each individual seeking, for example, a front-of-the-house job, cross training, a PAR test, PAR movement, or a management job. Teamsters, 431 U.S. at 360-61. The burden is upon the defendant employer to demonstrate that the individual was denied an employment opportunity for lawful reasons. Id. at 362. If the employer fails to carry that burden, the court must then fashion appropriate relief. Id. at 364-67; see also Rutstein, 211 F.3d at 1239. A plaintiffs task in establishing the requisite commonality and typicality is "more difficult" where disparate treatment is alleged. Nelson, 709 F.2d at 679 n. 9. In fact, commonality and typicality are often absent in class-wide disparate treatment claims because such claims are by their very nature individual. Washington v. Brown Williamson Tobacco Corp., 959 F.2d 1566, 1570 n. 10 (11th Cir. 1992).

78 Although Teamsters used the term "inference," the Eleventh Circuit in Rutstein noted the split among its cases regarding whether, subsequent to the establishment of a pattern or practice of discrimination, individual plaintiffs are entitled to a presumption in favor of individual relief, or whether the showing of class-wide discrimination creates an inference that any particular employment decision was made pursuant to the employer's discriminatory policies. Rutstein, 211 F.3d at 1237 n. 17. Because it was not necessary for the decision, the court did not resolve that conflict. Id.

Plaintiffs assert that they and their class would proffer statistical and anecdotal evidence to prove liability in Stage I. Pls.' Br. in Supp. of Class Certification [82] at 33-34. However, they ignore the difficulties that would be encountered at any Stage II proceeding. Although a plaintiff or class member would be entitled to an inference that the pattern or practice led to denial of an employment opportunity, Cracker Barrel would be entitled to demonstrate that the individual was denied that opportunity for lawful reasons. This would lead to individualized factual determinations for each plaintiff and class member.

For example, assume plaintiffs proved a pattern or practice of bias at Cracker Barrel at Stage I. Then, representative plaintiff Workue sought to take advantage of the inference that arose in her favor at Stage II. At this point, the defendant would be entitled to present the evidence discussed above (i.e., her falsification of applications, failure to complete applications, and apparent inability to get to work in suburban or coastal Virginia), along with evidence relating to each of her unsuccessful applications, such as (1) the job for which she applied; (2) whether there was a job opening at the time she applied; (3) who the decision-maker was; (4) why s/he decided not to hire plaintiff; (5) who, if anyone, was hired instead of plaintiff; and (6) the qualifications, experience, race, etc., of the person who may have been hired instead of plaintiff. The Court has not even addressed the additional complications that would arise when the class representatives and individual named plaintiffs sought to raise their non-class claims, such as hostile work environment and constructive discharge.

Defendant is correct that this type of fact-intensive inquiry would be repeated continually because plaintiffs complain about employment decisions made by hundreds (if not thousands) of independent local decision-makers.⁷⁹ The individualized determinations necessary to resolve the claims of each plaintiff and prospective class member is the antithesis of the streamlined, group adjudication envisioned under Rule 23. See Reid, 205 F.R.D. at 660 ("The class action device envisions that liability can be determined in a single trial, and it anticipates that members of the class are entitled either to identical relief or to relief so similar that little more than a calculator or formula is required to reach the result.") (citing Falcon, 457 U.S. at 155, for the proposition that class action device saves the resources of both the courts and the parties by permitting an issue potentially affecting every class member to be litigated in an economical fashion).

79 "Where the alleged discrimination results from the actions of an individual employee, supervisor, or decision-maker, maintaining a lawsuit as a class action may not be appropriate." Reid, 205 F.R.D. at 660.

Class action certification is inappropriate where many individualized factual determinations are necessary. See Trevino v. Holly-Sugar Corp., 811 F.2d 896, 905 (5th Cir. 1987) (necessity for individualized proof indicates that class action is not economical or efficient way of processing complaints of proposed class); Nelson, 706 F.2d at 679. The court in Patterson v. General Motors Corp. rejected class certification when

[t]he issue of whether a particular job assignment or promotion denial was discriminatory would depend upon any number of factors peculiar to the individuals competing for the vacancy, including relative seniority, qualifications, availability for work and desire to perform the job. Each disciplinary action would present a different set of facts for each employee. "In other words, the plaintiffs claims do not relate to general policies or practices which are allegedly discriminatory, but rather to individualized claims of discrimination which could not possibly present common questions of law or fact sufficient to justify class action treatment."

631 F.2d 476, 481 (7th Cir. 1980) (citation omitted). The presence of such factors in the claims of many of the plaintiffs and prospective class members here makes class certification inappropriate.

(b) Disparate Impact

Under the disparate impact theory, plaintiffs must show that a facially neutral requirement or policy of defendants disproportionately and adversely affected members of a protected group. Where this showing is made, the employer then has the burden of demonstrating the necessity for and job-relatedness of the criterion in question. See Griggs v. Duke Power Co., 401 U.S. 424, 434 (1971). There is no requirement to prove discriminatory intent. EEOC v. Joe's Stone Crab, Inc., 220 F.3d 1263, 1273 (11th Cir. 2000). Disparate impact claims may be brought only under Title VII and not under § 1981. Reid, 205 F.R.D. at 667 n. 15.

Congress codified the burdens of proof in a Title VII disparate impact case in the Civil Rights Act of 1991. See 42 U.S.C.A. § 2000e-2(k) (West 1994). In the first stage of a disparate impact case, the "complaining party [must] demonstrate that a respondent uses a particular employment practice that causes a disparate impact on the basis of race." Id. § 2000e-2(k)(1)(A)(i).⁸⁰ See generally In re: Employment Discrimination Litig. Against Alabama, 198 F.3d 1305, 1311-16 (11th Cir. 1999). This burden of proof is carried primarily through use of statistics. Courts must be mindful, however, of proper comparisons, and the use of the proper qualified applicant or incumbent pools, with rebuttal by the employer using its own statistics or by pointing out errors in plaintiffs statistics. Id. at 1312-14.

80 The statute also provides that "if the complaining party can demonstrate to the court that the elements of a respondent's decision-making process are not capable of separation for analysis, the decision-making process may be analyzed as one employment practice." 42 U.S.C.A. 2000e-2(k)(1)(B)(i) (West 1994).

If the plaintiff meets his burden of demonstrating that a challenged employment practice causes disparate impact, then the burden shifts to the defendant employer to "demonstrate that the challenged practice is job related for the position in question and consistent with business necessity." 42 U.S.C.A. § 2000e-2(k)(1)(A)(i) (West 1994). "Alternatively, the complaining party can demonstrate that other tests or selection devices, without a similarly undesirable racial effect, would also serve the employer's legitimate interest in efficient and trustworthy workmanship." In re: Employment Discrimination Litig., 198 F.3d at 1314-15 (quotation marks deleted); 42 U.S.C.A. § 2000e-2(k)(1)(A)(ii) (West 1994).

If it is ultimately determined that the employer violated the disparate impact provisions of Title VII, a court may order a wide range of equitable relief under 42 U.S.C. § 2000e-5 (g)(1). In re: Employment Discrimination Litig., 198 F.3d at 1315.

As for individual relief, if an individual plaintiff has shown that he or she was within the class of persons negatively impacted by the unlawful employment practice, then the employer must be given an opportunity to demonstrate a legitimate[,] nondiscriminatory reason why, absent the offending practice, the individual plaintiff would not have been awarded the job or job benefit at issue anyway.

Id. If the employer cannot so demonstrate, then individual relief may be awarded. Id. For an employee to obtain individual relief (e.g., a front-of-the-house job, cross training, a PAR test, PAR movement, or a management job), an inquiry similar to the remedial stage II of a pattern or practice disparate treatment claim is required. Robinson v. Metro-North Commuter R.R. Co., 267 F.3d 147, 161-62 (2d Cir. 2001),cert. denied, ___ U.S. ___, 122 S.Ct. 1349 (2002).

Assuming that plaintiffs could show prima facie disparate impact of any of Cracker Barrel's separate employment practices (which is not evident from the above-discussed statistical evidence), and Cracker Barrel could not demonstrate that a particular practice is job related and consistent with business necessity,⁸¹ Cracker Barrel could still avoid liability to an individual plaintiff or class member negatively impacted by the unlawful employment practice if it could demonstrate a legitimate nondiscriminatory reason why, absent the offending practice, the aggrieved individual would not have been awarded the job or job benefit at issue. It is at this point (which is akin to the Stage II proceeding of a pattern-or-practice case) that the numerous, individualized, factual determinations that prevented class certification of any disparate treatment claim also counsel against class certification of a disparate impact claim as well. These individualized determinations are, as previously noted, the antithesis of the streamlined process that Rule 23 envisions.

81 Plaintiffs contend that Dr. John Veres attests that Cracker Barrel has not satisfied the standards of job relatedness and business necessity (Pls.' Br. in Supp. of Class Certification [82] at 35), but they failed to list which specific employment practices were analyzed by Dr. Veres and failed to include his Declaration in the record. The only record evidence concerning job relatedness and business necessity is the report by a consultant hired by defendant, which validates the PAR program. Stinson Decl. ¶ 18, Ex. PAR-11.

(4) The Statistical Evidence Shows No Consistent Pattern of Adverse Treatment of African Americans

Statistical analyses of an employer's workforce by the parties' experts is often relevant to a class action determination. Plaintiffs assert that the Court must accept the assertions of their statistical experts as true and disregard any "dueling" between the parties' experts at the certification stage. Pls.' Reply Br. at 17 (citing, inter alia, Caridad, 191 F.3d at 292). However, the Second Circuit's construction of Rule 23 in Caridad, which assumes that whatever plaintiffs or their experts allege is true, as it would in testing the legal sufficiency of a complaint under Fed.R.Civ.P. 12(b)(6), ignores Falcon's rigorous analysis requirement. 457 U.S. at 161. Blind acceptance of a plaintiffs statistical evidence is not warranted. Instead, the appropriate standard is as follows:

Although on a class certification motion this court should not conduct a preliminary inquiry into the merits of the suit, inquiries into whether plaintiffs have met their burden of showing that class certification is appropriate often becomes intertwined with a limited discussion of the merits. As the Eleventh Circuit has noted, the principle that the court may not reach the merits of the claim "should not be talismanically invoked to artificially limit a trial court's examination of the factors necessary to a reasoned determination of whether a plaintiff has met [the] . . . burden of establishing each of the Rule 23 class action requirements." For that reason, an examination of what the statistical evidence shows and does not show, and preliminary judgments regarding the credibility of the showing made by the plaintiff in support of the motion for class certification are necessary in order for the court to undertake a sifting examination of the Rule 23 requirements.

Cullens v. Ga. Dep't of Transp., 38 Empl. Prac. Dec. P 35, 579, 1985 WL 307, at *14 (M.D. Ga. Aug. 19, 1985) (quoting Love v. Turlington, 733 F.2d 1562, 1564 (11th Cir. 1984)); see also Reid, 205 F.R.D. at 661-62 (statistical evidence reviewed and weighed); accord Cooper, 205 F.R.D. at 611-15.

In accordance with the above-quoted standard, the Court has conducted a rigorous analysis of the various statistical analyses presented (see supra Part IV), weighing the competing positions and rejecting those analyses that do not assist the Court in determining whether class certification is appropriate. The Court notes initially that, even if those statistics showed a pattern consistently adverse to African Americans, the decentralized nature of Cracker Barrel's employment decision-making and the individualized nature of the class claims, discussed supra, would prevent class certification.

However, the Court is unable to discern from plaintiff's nationwide, aggregated statistics a consistent pattern of treatment adverse to African Americans. For example, although African Americans were over-represented in back-of-the-house positions compared to their workforce percentage, those jobs are not undesirable. Many named plaintiffs rated back-of-the-house jobs as among the best in the store.See, e.g., Rhodes Dep. at 53-54. Most significantly, the evidence shows that back-of-the-house jobs are among the best paying in the Company and sought by African Americans. Jones Decl. ¶ 22; H/T Report at 27-28 Fig. 2. Indeed, from the job interest survey, back-of-the-house jobs were of greater interest to African Americans than front-of-the-house jobs. H/T Report at 2-3, 57-58 Table 18. African Americans also desired to move from one back-of-the-house job to another rather than from a back-of-the-house job to one in the front-of-the-house. Id. at 3, 58-61 Table 19. Analyses by both experts agreed that there were no more African Americans employed in back-of-the-house jobs than one would have expected when job interest was taken into account.

That African Americans more often sought back-of-the-house jobs was also borne out by the survey of a sample of employee applications. See supra Part IV.C. Cracker Barrel hired the overwhelming percentage of persons into the jobs they sought with little variation by race. H/T Report at 49. Surprisingly, African Americans were more likely to be placed in front-of-the-house jobs when they had not so requested. Id. at 49-50. Again, applicant preference, not race, seems to account for the racial distribution that exists at Cracker Barrel. Because the positions into which applicants are placed depend in great part upon applicant preference, prior experience, and what job openings are available, the Court is cautious about drawing negative conclusions from plaintiffs' experts' opinion that the proportion of African Americans assigned to back-of-the-house positions was higher than the proportion of Whites assigned to those jobs. See B/F Report at 11.

As for the time that African-American employees spend in various PAR levels (see supra Part IV.E.), the Court is again cautious about drawing negative inferences because taking a PAR test and moving to another level is. again dependent on employee initiative in having a successful job evaluation, taking the PAR test, and studying sufficiently to pass it. H/T Report at 15-16, 64. Even ignoring this tremendous caveat, the aggregated statistics generated by plaintiffs' experts showed a shortfall for African-American employees of only 1.6% of 225, 254 PAR moves nationwide over six years. Id. at 66. When PAR movement is examined on a storelevel basis, most stores had no difference between the actual number of African-American PAR moves and the number predicted. Id. at 67. This is a strong indication that no pattern adverse to African Americans is occurring across the Company. When one looks at actual PAR test passing, the shortfall of African Americans passing is minuscule. See id. at 70.

Moreover, when one examines store-level results, there again is no pattern adverse to African Americans of taking or passing PAR tests. H/T Report at 72. Controlling for store, job, and PAR level, African-American employees take an average of 1.05 tests and non-African Americans an average of 1.04 tests. H/T Rebuttal Report at 38. Further, in 98.9% of the comparisons, there was no statistically significant difference on the basis of race in the number of times employees took tests. Id. at 38-39. Plaintiffs' experts asserted that for each PAR level from 1995 to 2000, the percentage of African-American test failures among first-time takers was greater than the percentage of non-African Americans in the first-time category to a significant statistical degree. B/F Rebuttal Report ¶¶ 56-57 Ex. 24. The report did not disclose the number of tests making up the difference in pass rates between African Americans and non-African-American test-takers. However, on the basis of plaintiffs' experts' analyses, defendant's experts calculated a difference of less than one-tenth of one percent between the actual and predicted number of passes and fails for African-American employees. H/T Rebuttal Report at 33. Finally, though plaintiffs assert that the failure to pass the PAR tests by African-American employees prohibits them from moving upward in PAR level, only 3.4% of all stores had statistically significantly fewer African-American test-takers moving upward than predicted, the majority (96.6%) of stores having test results that were either neutral or favorable to African Americans. H/T Rebuttal Report at 43-44. In light of these statistics, there is little evidence of a systematic bias in African-American PAR testing.

With regard to compensation, if Cracker Barrel were systemically discriminating against African Americans, one would expect to see significant negative pay differentials. However, an analysis that does not unfairly double the tip income of servers shows that African-American employees had statistically significantly higher compensation than White employees every year studied. Compare H/T Rebuttal Report at 47 with B/F Rebuttal Report ¶¶ 64-69.

With regard to hiring, the statistics generated by plaintiffs' experts showing that Cracker Barrel hired fewer African Americans than one would have expected must be disregarded because of the inappropriate comparison to EEO-9 census data. See supra Part IV.H.; B/F Rebuttal Report ¶ 25. Cracker Barrel does not hire from that broad a pool and typically hires experienced over inexperienced persons. See, e.g., H/T Rebuttal Report at 19; Bradley Dep. Vol. 11 (5/29/02) at 561.

Although the statistics for promotion of African-American employees into the Management Internship Program were not favorable to defendant, serious questions abound about the methodology that plaintiffs' experts used to obtain statistical significance. When both experts ran calculations limiting more appropriately the pool of candidates available for promotion, the numbers were closer to two standard deviations, which is barely significant. In light of the foregoing, the Court is not inclined to view this as evidence of discrimination. See Maddox, 764 F.2d at 1552.

Finally, although cross-training statistics were not examined (see supra Part IV.I), the named plaintiffs seemed to be able to cross-train at will (but, as the Court notes, many of them sought and obtained cross-training in another back-of-the-house job, not cross-training for a front-of-the-house job). See supra Part III.A. 1(b)(2).

Simply put, the statistics, properly analyzed, show no consistent pattern adverse to African-American applicants or employees at Cracker Barrel. On the contrary, the statistics show favorable treatment to African Americans in terms of compensation.

(5) Cracker Barrel's Varying Employment Practices Defeat Commonality/Typicality

Plaintiffs allege that "channeling" created the alleged imbalances in the racial distribution of Cracker Barrel's workforce between front-of-the-house jobs and back-of-the-house jobs and then perpetuated those imbalances. It appears to the Court that, by couching their claim as one of simple "channeling," plaintiffs disguise the fact that they are bringing the type of "across the board" class action attacking all of Cracker Barrel's employment practices that was found improper in Falcon. 457 U.S. at 161. However, channeling under plaintiffs' theory is not a discrete employment practice. Instead, it involves many practices, such as hiring (both at new stores and existing stores), PAR movements, cross-training, compensation, and promotion. Each of these practices is governed by a different set of Cracker Barrel policies and/or procedures which are applied, as noted above, primarily by local management. (For a description of these various practices, see supra Part III.A. 1.)

A trial attacking these different policies will not raise common issues. For example, as defendant correctly contends, the trial of whether Cracker Barrel engages in discriminatory initial job assignment will not raise issues common to the question of whether Cracker Barrel discriminates in PAR movement. To determine if there was discrimination in initial job assignment, evidence must be presented of what jobs were available, who applied, their races, and their qualifications, and why the hiring managers in each situation made the decisions they made.

In contrast, in a challenge to PAR movement, evidence must be presented of a plaintiffs PAR level, whether he successfully passed a performance evaluation (and if not, whether his manager evaluated him in a non-discriminatory manner), whether the employee studied sufficiently to pass the PAR test, and whether he obtained the next PAR level. Evidence about the performance evaluation scores and PAR testing results of other employees in the store in which the plaintiff worked would also be probative.

Thus, simply labeling the claim as one for "channeling" does not insure that it is worthy of class treatment. Although plaintiffs cite cases that certified channeling claims (Pls.' Br. in Supp. of Class Certification [82] at 5-6), those decisions are distinguishable from the facts here. In two of the cases, Cox, 784 F.2d at 1551, and Drayton, 203 F.R.D. at 527, the employer had an avowed, overt policy of workforce segregation. There is no allegation or proof of an overt policy of racial segregation here. On the contrary, the official Company policy is one of non-discrimination.

The other cases cited by plaintiffs rely on the so-called Falcon footnote 15 to support class certification in channeling cases.⁸² The courts in those cases found that the multiple practices that allegedly constituted channeling were united by entirely subjective employment practices. See Carpenter v. Stephen F. Austin State Univ., 706 F.2d 608, 617 (5th Cir. 1983) (challenges to pay, promotion and termination practices at single location united by entirely subjective employment practices); McClain v. Lufkin Indus., Inc., 187 F.R.D. 267, 279-80 (E.D. Tex. 1999) (class certified at single industrial complex where employment decisions were either based on entirely subjective employment practices or centrally controlled); Beckmann v. CBS, Inc., 192 F.R.D. 608, 614 (D. Minn. 2000) (employment practices at five television stations united by entirely subjective employment practices).

82 In Falcon, the Supreme Court found that it would still be possible to certify a class action challenging multiple employment practices in the rare instance where "discrimination manifested itself in [different employment] practices in the same general fashion, such as through entirely subjective decision-making processes." 457 U.S. at 159 n. 15.

Plaintiffs concede that Cracker Barrel's selection procedures involve a combination of both subjective and objective components. Pls.' Br. in Supp. of Class Certification [82] at 36 n. 13. The presence of both components takes this case out of Falcon footnote 15. See Vuyanich v. Republic Nat'l Bank of Dallas, 723 F.2d 1195, 1199-1200 (5th Cir. 1984) (processes not entirely subjective when employer used objective inputs);Cooper, 205 F.R.D. at 627 (employment practices not entirely subjective where, although managers exercised discretion, they were guided by written policies and procedures).

6. Subclasses Are Not a Solution

The three subclasses plaintiffs propose if their original proposal was rejected still fail to solve the above-mentioned problems. Pls.' Mot. for Class Certification [82] at 2. Each proposed subclass still combines discrete employment practices that cannot be challenged absent an overt policy of discrimination or an entirely subjective decision-making process, neither of which is present. Each proposed subclass would still involve individualized claims of discrimination by hundreds or thousands of persons employed at hundreds of stores in over forty-one states in which the decisions of hundreds or thousands of autonomous decision-makers would be challenged. The lack of commonality and typicality would still be a bar to certification of any subclasses.

c. Adequacy of Representation

Given the undersigned's report that plaintiffs fail to prove the requisite commonality and typicality and recommendation that class certification be denied, it is unnecessary to analyze this final Rule 23(a) requirement. Nevertheless, the Court will do so briefly to address issued raised by defendant against plaintiffs' counsel. Defendant asserts that plaintiffs' attorneys could not adequately represent the proposed class because of (1) their argument that servers may have under-reported tip income, and (2) potential conflicts arising out of their representation of plaintiffs in the wage-hour litigation pending in this Court, styled McDerrnitt v. Cracker Barrel, No. 4:99-CV-0001-HLM (N.D. Ga. filed Jan. 4, 1999), and Stanley v. Cracker Barrel, No. 4:01-CV-0326-HLM (N.D. Ga. filed Dec. 13, 2001).

To satisfy the adequacy of representation requirement, the Court must determine (1) that the plaintiffs' attorneys are qualified, experienced, and will competently and vigorously prosecute this suit; and (2) that the interests of the class representative are not antagonistic to or in conflict with other members of the class. Griffin v. Carlin, 755 F.2d 1516, 1533 (11th Cir. 1985). Assuming that defendant's arguments fall within the first element above, the Court rejects them. Plaintiffs' counsel did not accuse any specific server of under-reporting tips, but simply argued (albeit unsuccessfully) from their experts' report that servers generally under-report tips. Moreover, any conflicts between the representation by plaintiffs' counsel of this putative class and the wage-hour plaintiffs in the two above-referenced cases are speculative and remote.

2. Rule 23(b)(2)

Plaintiffs move for certification under Fed.R.Civ.P. Rule 23(b)(2), which provides as follows:

(b) Class Actions Maintainable. An action may be maintained as a class action if the prerequisites of subdivision (a) are satisfied, and in addition:

. . . .

(2) the party opposing the class has acted or refused to act on grounds generally applicable to the class, thereby making appropriate final injunctive relief or corresponding declaratory relief with respect to the class as a whole.

Given the undersigned's report that plaintiffs have not satisfied Rule 23(a) and recommendation that class certification be denied, the Court need not determine whether plaintiffs have proven entitlement to certification under Rule 23(b)(2). Nevertheless, because plaintiffs challenge thousands of decisions by hundreds or perhaps thousands of different decision-makers across six years in over 450 Cracker Barrel stores located in forty-one states, it is difficult to see how defendant "has acted or refused to act on grounds generally applicable to the class." Abrams, 178 F.R.D. at 134 (rejecting 23(b)(2) certification where individual instances of discrimination involving different managers, jobs, and locations made class wide injunctive relief inappropriate).

VI. CONCLUSION

For the reasons stated above, the undersigned recommends that Plaintiffs' Motion for Class Certification [82] be DENIED.

ORDER

Before the Court are the following Motions relating to whether certain evidentiary submissions made in support of, or in opposition to, Plaintiffs' Motion for Class Certification^1a should be considered:

1. Defendant's Motion To Strike Plaintiffs' Declarations in Support of Class Certification [100]; and

2. Plaintiffs' Motion To Strike Cracker Barrel Employee-Witness Declarations Submitted in Opposition to Class Certification [111].

In support of these Motions, the parties submitted Briefs [100, 111], Response Briefs [107, 112], and Reply Briefs [109, 113]. Defendant also filed an objection to the reply brief that plaintiffs submitted in support of their Motion To Strike [116] on the grounds that it filed two days late, to which plaintiffs responded [117].

With regard to Defendant's Motion To Strike Plaintiffs' Declarations [100], the undersigned recognizes that certain statements made by plaintiffs in their Declarations contradicted their deposition testimony, were hearsay, or were not based on personal knowledge. However, in finding the facts related to each plaintiffs claim in the RR, the Court disregarded those portions of the Declarations that did not comply with Fed.R.Civ.P. 5 ¶ (e) or violated the Eleventh Circuit's decision inVan T. Junkins Assocs. v. U.S. Indus., 736 F.2d 656 (11th Cir. 1984). Because the Court can determine admissibility, there is no need to take the drastic step advocated by defendant and strike the Declarations in their entirety. Accordingly, Defendant's Motion To Strike Plaintiffs' Declarations [100] is DENIED.

As for Plaintiffs Motion To Strike Cracker Barrel Employee Witness Declarations, the Court rejects plaintiffs' contention that the Declarations should be stricken because they were not produced as part of defendant's initial Disclosures. Defendant acted properly on the basis of the Court's Case Management Scheduling Order. Accordingly, Plaintiffs' Motion To Strike [111] is DENIED.

Although the disputed Declarations may remain in the record, they are irrelevant to the issue of whether the plaintiffs have satisfied the requirements of Fed.R.Civ.P. 23. The racial discrimination about which plaintiffs complain could exist despite Declarations from 233 African-American employees containing statements favorable to their employer. Therefore, the undersigned has not considered these Declarations in preparing the RR.

In conclusion, Defendant's Motion To Strike Plaintiffs' Declarations in Support of Class Certification [100], and Plaintiffs' Motion To Strike Cracker-Barrel Employee-Witness Declarations Submitted in Opposition to Class Certification [111], are both DENIED.