McReynolds v. Sodexho Marriott Services, Inc.

208 F.R.D. 428, 2002 U.S. Dist. LEXIS 11744, 2002 WL 1400738

• Court: District Court, District of Columbia
• Decided: June 27, 2002
• Docket: No. Civ.A. 01-510(ESH)
• Status: Published

Opinion

MEMORANDUM OPINION

HUVELLE, District Judge.

In General Telephone Co. v. Falcon,1 the Supreme Court ruled that “a Title VII class action, like any other class action, may only be certified if the trial court is satisfied, after a rigorous analysis, that the prerequisites of Rule 23[] have been satisfied.” Id. at 161, 102 S.Ct. 2364. In accordance with that admonition, the parties in this case, a proposed class action brought under Title VII of the Civil Rights Act of 1964 (“Title VII”), as amended, 42 U.S.C. § 2000e, et seq., have submitted nearly 150 pages of pleadings and nine volumes of exhibits. Upon consideration of this comprehensive record, the Court concludes that plaintiffs have fulfilled the requirements of Fed.R.Civ.P. 23(b)(2) with respect to the liability phase of the case, and their motion for class certification will therefore be granted to that extent.

BACKGROUND

1. Sodexho’s Organizational Structure

The ten named plaintiffs, who are also the proposed class representatives, are current and former employees of defendant Sodexho Marriott Services, Inc. (“Sodexho”). Sodex-ho is the largest food and facilities management corporation in North America, with approximately 110,000 employees at 5,000 different client worksites, and over $4.7 billion in revenue last year. Sodexho was formed on March 27, 1998, from the merger of Marriott Management Services (“MMS”), a division of Marriott International, Inc., and Sodexho U.S.A. (“SUSA”), a branch of the French corporation Sodexho Aliance, S.A.² At the time of the merger, MMS employed approximately three times as many people as SUSA. (Def. Ex. 2, Declaration of Diana Newmier (“Newmier DecL”) 11114-5.)

The parties dispute the nature of Sodexho’s organizational structure. However, because the allegations in the complaint are presumed true for purposes of a motion for class certification, In re Lorazepam, & Clorazepate Antitrust Litigation, 202 F.R.D. 12, 14 (D.D.C.2001), appeal denied, 289 F.3d 98 (D.C.Cir.2002); Stevelman v. Alias Research Inc., 2000 WL 888385, at *1 (D.Conn. June 22, 2000) (citing Shelter Realty Corp. v. Allied Maintenance Corp., 574 F.2d 656, 661 n. 15 (2d Cir.1978)); Johnson v. Aronson Furniture Co., 1998 WL 641342, at *2 (N.D.Ill. Sept.11, 1998) (citing Johns v. DeLeonardis, 145 F.R.D. 480, 482 (N.D.Ill.1992)), the Court will assume the facts as set forth by plaintiffs.

According to the complaint, Sodexho is separated into six divisions — Corporate Services, Health Care, Campus Services,³ School Services, Canada, and Laundry — of which the first four are the largest. (Complaint fl 17; Pl.Mem. at 3.) The primary responsibility of most employees in the Corporate Services, Health Care, Campus Services, and School Services divisions is to provide food and facilities management services to their clients, and the names of these divisions reflect the nature of the clients’ businesses rather than the services that Sodexho provides to them.⁴ (Pl.Mem. at 3; Def.Opp. at 6.) Laundry Services cleans linens and uniforms for health care and lodging clients.⁵ Most Sodexho employees work on the premises of their clients. These positions are known as “in-unit” or “operations” jobs. (Complaint 1118.) The highest in-unit position is the Director or General Manager of a client facility. (Id.) The promotional pool for managerial positions is company-wide, and managers apply for and receive promotions across divisions. (Pl.Ex. 4, Deposition of Henry Watkis (“WatMs Dep.”) at 67-68.) However, each division maintains its own human resources department, and most promotion decisions are made and reviewed entirely within each unit. (Def.Ex. 1, Declaration of Ollie Lawrence (“Lawrence Deck”) 111121-23.)⁶

A smaller group of Sodexho employees works at its corporate offices and other locations as supervisors or in areas such as sales, human resources, or finance. Managers in these areas are said to work “above the unit,” and “generally have greater authority and receive higher salary, benefits, and perquisites than General Managers.” (Answer 1119.) In fact, General Managers typically report directly to an above-the-unit manager known as a District Manager. (Id.)

Until late 1998, Sodexho organized its salaried positions by grades, which ranged from 40 to 64. Most in-unit managers occupied grades in the 40s, while General Managers held grades 51 to 53, and District Managers held grades 51 to 54. Most General Managers were grade 51, and most District Managers were grade 52 or 53. (Pl.Ex. 9, Expert Report of Bernard Siskin (“Siskin Rep.”) at Tables 1-2.) In late 1998, Sodexho reclassified its salaried positions using a broader band system, and multiple grades under the former system were collapsed into a single band. That system is still in effect today. In-unit managers now occupy bands 75 to 77; lower out-of-unit employees hold bands 85 to 88; and above-the-unit positions are bands 89 to 95. (Siskin Rep. at 2.)

II. Sodexho’s Promotion Policies and Practices

Sodexho is a corporate giant, and as a result, its personnel decisions are largely decentralized. As defendant notes, “each manager at Sodexho has discretion to decide what criteria to apply in making decisions involving promotions and other personnel actions.” (Def.Opp. at 12-13; see Def.Ex. 28, Deposition of Cynthia Carter McReynolds (“McReynolds Dep.”) at 46-54.)

In 1994, in response to “a multi-million dollar discrimination suit,” MMS implemented a formal system for salaried positions called the Marriott Career Management System (“MCMS”). (Pl.Ex.13.) Under MCMS, employees in any division could apply for job vacancies throughout the country with the consent of their supervisor. This posting process applied company-wide, and employees could apply for jobs in other divisions.⁷ (Pl.Ex. 12, Deposition of William Anstee (“Anstee Dep.”) at 28, 133; Pl.Ex. 16, Deposition of George Koenig (“Koenig Dep.”) at 90-91.) This system survived the 1998 merger, and in 2000, Sodexho modified its posting system using a web-based program so that employees could immediately apply for job vacancies without their supervisors’ consent. (Pl.Ex. 19, Deposition of Dolores Coe (“Coe Dep.”) at 201-02; Complaint H 20.) This system is known as “Career Center” or “Talent Point.” (Id.) In other material aspects, the promotion system has not changed since 1994.

However, MCMS’s role in the promotion practices is somewhat unclear. A February 23,1999 Sodexho internal memorandum summarized the problems with MCMS that existed at the time. In particular, the memo noted that the system was “[vjiewed as customer unfriendly, not time efficient to use,” and that there was “[n]o adequate system for follow up to candidates who post.” (Pl.Ex. 18, at 2.) The company’s own statistics bear this out. In 1998, for instance, 5,215 positions were posted on MCMS, but only 1,671 were filled, while 2,595 were closed or deleted. (Id. at 3.) Moreover, at the time of the 1999 memorandum, Sodexho was still trying to determine “[w]hat positions should be posted [and h]ow long [positions should] be posted for.” (Id.)

Irrespective of how an employee applies for a promotion, defendant’s promotion decision-making process is decentralized and allows unfettered managerial discretion. (Def.Opp. at 12.) According to William An-stee, Sodexho’s witness on its promotion process, the hiring manager has complete discretion to decide whom to interview for each position. (Anstee Dep. at 81.) There are no company rules or guidelines regarding the promotion process, and there are no requirements to take any notes during the interview process, to rank applicants, or to explain the selection. (Id. at 90-92.) After a hiring manager makes his selection, he “would take that decision to whoever their boss was [and] say, hey, I’ve selected candidate A.... And then the district manager along with the human resource manager [puts] together an offer if that, indeed, was what it was.” (Id. at 92.) In short, Sodexho has “no company guidelines or fixed criteria or anything ... to follow in making those decisions.” (Id. at 182.)

III. Plaintiffs’ Allegations of Discrimination

Plaintiffs challenge these promotion practices as they relate to high-level managerial positions at or above the level of an above-the-unit manager. (Complaint H 22.) Specifically, they contend that the system for filling vacant managerial slots is characterized by three common elements: (1) positions are filled without being posted at all; (2) when jobs are posted, favored employees are preselected for the positions; and (3) when “more than one person is actually considered for an opening, hiring managers have virtually total discretion in their promotion decisions because they are not given job-related or objective criteria and are not required to document the reasons for their decisions, making them unreviewable.” (Pl.Mem. at 7.) Plaintiffs allege racial discrimination under both disparate treatment and disparate impact theories, and seek to certify a class

of all African-Americans who are or were salaried employees of Sodexho [] at any time from March 9, 1998, to the present, and who have held or sought to obtain (1) an upper-level managerial, supervisory, or professional position (above-the-unit or comparable level of responsibility) or (2) a job that would lead to such a position, and who have been, continue to be, or may in the future be adversely impacted by So-dexho’s racially discriminatory policies and practices affecting promotions or advancement.

(Pl.Mot. at 1.)

This system is discriminatory, according to plaintiffs, because high-level managers at So-dexho are predominantly white, and white employees are pre-selected for and promoted to managerial positions at the expense of more qualified African-American applicants, who receive no response to their job inquiries. To support this argument, plaintiffs offer both statistical and anecdotal evidence, including the testimony of a senior-level white manager who charges that Sodexho has intentionally discriminated against African-American employees in its promotion practices.

Although the parties have extensively briefed the merits of plaintiffs’ Title VII claims at this class certification stage, an abbreviated survey of the plaintiffs’ evidence that is relevant to the instant motion is set forth below. As noted, this evidence is the traditional combination of “anecdotal evidence of discrimination in combination with statistical evidence of minority underrepresentation.” Allen v. Chicago Transit Authority, 2000 WL 1207408, at *10 (N.D.Ill. 2000).

A. Anecdotal Evidence of Discrimination

Plaintiffs offer numerous anecdotes of discrimination from the testimony of high-level Sodexho officials. For example, they cite the testimony of Sodexho’s top official for equal employment opportunity, who stated that the number of African-American promotions “indicated a problem” and “would be considered low” (Pl.Mem. at 15; Watkis Dep. at 202-03); the President of Sodexho’s Health Care Division, who, in describing the lack of women among the 100 District Managers in that unit, noted, “Well, you think that’s bad, I think Henry [Richardson] is the only black one we have” (Pl.Mem. at 16; Pl.Ex. 26, Deposition of Kelly Hymes (“Hymes Dep.”) at 312-13); the deposition testimony of a named plaintiff, who recalled the statement of a Sodexho Human Resources Manager at a conference that there was no point in training other Human Resources Managers to recruit more African Americans, because “our district managers won’t hire [black people],” a statement with which others at the conference — including a Corporate Vice President — allegedly agreed (Pl.Mem. at 18-19; Pl.Ex. 28, Deposition of Tasha Hardy (“Hardy Dep.”) at 163,208-09); and the comment of Brenda Hendrickson, an African-American supervisor at the company, that it is “common practice at Sodexho for white managers to refer to African Americans as ‘you people’ or similar phrases.”⁸ (Pl.Mem. at 20; Pl.Ex. 7, Deposition of Brenda Hen-drickson (“Hendrickson Dep.”) at 113-14, 117-19.)

Plaintiffs also offer the testimony of Kelly Hymes, a white former District Manager, who testified to episodes of preselection of white managers by white decision-makers, the use of the posting process as a “formality,” and the derogatory statements about African-Americans attributed to a white upper-level manager, who allegedly said that “African Americans were genetically inferior to whites, and genetically most of the criminals in the world were African Americans, and that they didn’t deserve to have promotions.” (Pl.Mem. at 19; Hymes Dep. at 95-101,179,184.)

In addition, plaintiffs contend that Sodexho has ignored warnings to remedy this discrimination, but instead, it has tried to cover it up. For example, plaintiffs allege that beginning in 1995, MMS executives were made aware of the lack of African-American above-the-unit managers by members of the company’s diversity roundtable group, but did nothing to remedy the problem. (Pl.Ex. 8, Deposition of Chip Moss (“Moss Dep.”) at 200-02.) They also cite the deposition testimony of Ellen Early, a putative class member, who stated that a white Sodexho executive refused to let her look at the company’s workforce composition data during a Sodexho diversity meeting. “[T]he executive said to me that he could not fax the information to me. And when I stated why, he said, well, because the numbers really don’t look very good____He said, we’re not — we’re not letting any of this out....” (Pl.Ex. 25, Deposition of Ellen Early (“Early Dep.”) at 242-43.)

B. Statistical Evidence of Minority Un-derrepresentation

The parties devote much of then-time to jousting over the merits of the analyses of their respective statistical experts. While this Circuit has not squarely addressed the standard for evaluating the parties’ competing statistical evidence on a motion for class certification, it has recognized that a trial court should accept plaintiffs statistical evidence as true at this stage. See, e.g., Wagner v. Taylor, 836 F.2d 578, 594 (D.C.Cir.1987) (accepting as true plaintiffs proferred statistics for purpose of motion for class certification, and holding that “[plaintiffs seeking class certification are not required to prove the merits of their cases, but their presentations must be specific enough to allow the court to discern at least a rough outline of a class that properly can be certified”); Gonzalez v. Brady, 136 F.R.D. 329, 333 (D.D.C.1991) (accepting plaintiffs’ statistics as true for purposes of the Rule 23 inquiry). Moreover, the Second Circuit has considered this issue and held that, while “[a] district court must ensure that the basis of the expert [statistical] opinion is not so flawed that it would be inadmissible as a matter of law, ... [it] may not weigh conflicting expert evidence.” In re Visa Check/Mastermoney Antitrust Litigation, 280 F.3d 124, 135 (2d Cir.2001). In short, “this sort of statistical dueling is not relevant to the certification determination.”⁹ Caridad v. Metro-North Commuter Railroad, 191 F.3d 283, 292 (2d Cir.1999) (internal quotation omitted). This doctrine has been adopted by other courts around the country. See, e.g., DeLoach v. Philip Morris Cos., 206 F.R.D. 551, 2002 WL 559453, at * 13 (M.D.N.C. Apr.3, 2002); Wilfong v. Rentr-A-Center, Inc., 2001 WL 1728985 (S.D.Ill. Dec.27, 2001). Nonetheless, an examination of the statistics proferred by plaintiffs is pertinent to the question of class certification.¹⁰ Wagner, 836 F.2d at 592-94.

Plaintiffs’ statistical expert, Dr. Bernard Siskin, conducted two separate analyses of Sodexho’s promotion system. In his first study, he analyzed promotions between grades and bands. Siskin noted that as of January 2001, only 10 of 306 District Managers (3.3 percent) were African-American, despite the fact that 13.3 percent of Unit and General Managers were black. (Siskin Rep. 113.) Moreover, only 18 out of 692 above-the-unit jobs, or 2.6 percent, were held by African Americans as of September 2000. (PL Ex.43.) Examining Sodexho workforce composition data since 1995¹¹ and controlling for race and seniority, Siskin found that African Americans received 143 fewer promotions than they would have in a non-discriminatory system. (Siskin Rep. 117.) Siskin determined that the likelihood of this occurring without race being a factor was less than one in 82.7 billion. (Id. U33.) He also concluded that African-American employees were stuck at the lowest levels of management for their seniority. “If African American and white management employees who were similarly situated with respect to seniority had been equally likely in the year 2001 to be at a given band, there would be 174 more African Americans at bands 76 and 77, 27 more African Americans at bands 88 and 89, and 42 more African Americans at bands 90 and above.” (Id. 1124.) Siskin found the standard deviations for these disparities to be 8.58, 4.42, and 5.40, respectively.¹² (Id.) Next, Siskin examined the effect of these disparities on salaries, and determined that the typical African-American Sodexho employee received $4,934 per year less than his white counterpart, at a standard deviation of 6.16. (MU40.)

Finally, Siskin found that Sodexho segregates its employees by race. Plaintiffs’ statistics indicate that 596 of Sodexho’s 878 units have no black managers, and that nearly two-thirds of all African-American managers in the remaining 282 units are concentrated in fewer than ten percent of those accounts. (Watkis Dep. at Exs. 34, 37.) Siskin determined that black managers are two-to-three times as likely as white • managers to report to an African-American District Manager (Siskin Rep. 1139), and that the statistical evidence supports testimony that African-American managers are disproportionately assigned to “black accounts.”¹³ (Id. H 4; Hymes Dep. at 389-90.)

In his second study, Siskin also examined within-band promotions. With the addition of this new data, Siskin found that the shortfall in promotions for African Americans nearly doubled, rising from 143 to 265.¹⁴ (Siskin Supp.Rep. 116.) Siskin also analyzed the study of plaintiffs statistical expert, Dr. Joan Haworth, who had examined Sodexho’s promotion of African-American employees by division. Siskin found that there was a shortfall in each division, and that even though these shortfalls varied from division to division, the disparity was statistically insignificant. Siskin therefore concluded that the discrimination was company-wide, and that the level of bias did not statistically vary by division. (Id. 111122-23.) “In sum,” Siskin wrote, “the pattern of shortfalls in each division is consistent with a common promotion process, and hence supports the argument that there is a common discriminatory company-wide promotion process.” (Id. H 23.)

C. The Allegations of the Named Plaintiffs

In addition to their classwide allegations of discrimination, each of the ten named plaintiffs has set forth individual claims of racial discrimination in the complaint. Although concentrated primarily in the Health Care unit, the named plaintiffs also represent the Education, School Services, and Corporate Services divisions of Sodexho.¹⁵ In addition, while most of the named plaintiffs worked for Sodexho in the District of Columbia, Virginia and Maryland, one was employed in California, and another is a resident of Georgia. All of the named plaintiffs worked at MMS before it merged with SUSA, and all allege that they have been discriminated against because of their race in the course of applying for a promotion.

Cynthia McReynolds began working at MMS in 1984, and is currently employed in Sodexho’s Health Care Division as the General Manager of Howard University Hospital in the District of Columbia. (Complaint 114.) In her 17 years of employment at Sodexho, McReynolds has posted for vacant positions — primarily those outside of her unit— between 50 and 75 times. Although she has been promoted within her unit, McReynolds has never received an out-of-unit promotion, and contends that she has repeatedly been passed over in favor of less qualified white applicants. (Id. H 46a.) In 2000, for example, McReynolds posted for at least seven positions, but received no information regarding any of her applications. (Id. U 46b.) She alleges that several years before, she was passed over for promotions to Senior Manager of Human Resources and Human Resources Recruiting Manager in favor of less qualified white applicants. (Id. 111146c-e.)

Lisa Mitchell worked at MMS and Sodex-ho from 1988 to 1999, and was employed primarily in the Health Care Division. (Id. H 5.) Just prior to leaving the company, she worked as General Manager of Suburban Hospital in Bethesda, Maryland, where she alleges that she was discriminated against in a variety of ways by her white supervisor. (Id. UH45a-e.) Mitchell also claims that in 1994, she was passed over for a promotion in favor of a less-qualified white employee who not only was not one of the original applicants for the position, but in fact had been a member of the panel that was interviewing candidates for the job. Mitchell complained about this conflict to the white Regional Vice President, who refused to change the selection decision.¹⁶ (Id. H 45d.)

Robert Morris began working at MMS in 1990, and is currently employed in Sodexho’s Health Care Division as a General Manager at the Civista Medical Center in La Plata, Maryland. (Id. 116.) Morris contends that he has been passed over for promotions as a result of the preselection of white candidates, and specifically, that he was discriminated against when Sodexho hired a less qualified white candidate as Acting Director of Prince Georges Hospital, without even offering Morris the opportunity to interview for the acting or permanent position. (Id. U 40.)

Chip Moss worked in the Health Care Division of MMS and Sodexho in Maryland and Virginia from 1993 to 2000. (Id. f 117, 41.) In 1999, he applied for the position of District Manager in the Senior Services Division, which is part of the company’s Health Care Division in Gaithersburg, Maryland. However, Moss was not offered the opportunity to interview for the position, which instead was filled by a white male. (Id. U 41a.) A few months later, he attempted to apply to be the General Manager at Fairfax Hospital, although the job was not posted. Moss was rebuffed when the white District Manager failed to return his phone calls, and then hired a white applicant for the position. Moss never received an explanation as to why he was not even interviewed for either of these vacancies. (Id. HH 41a-b.)

Marilyn McNish has been employed by MMS and Sodexho since 1991, and currently works in the Health Care Division as a Clinical Nutrition Manager at Prince Georges Hospital in Cheverly, Maryland. (Id. 118.) McNish alleges that it took her ten years and an EEOC complaint to be promoted to her current position, while similarly situated white employees were promoted in a fraction of that time. (Id. 1142.)

Ezzie Henry worked primarily in the School Services Division of MMS-and Sodex-ho from 1993 to 2000, and currently lives in Port Deposit, Maryland. (Id. 119.) In her six years as a Human Resources Specialist, Henry posted for managerial positions at least ten times, but received only one temporary promotion. (Id. 1137a.) She alleges that she was discriminated against because of her race when she applied to become Senior Human Resources Manager for the Health Care Division (Id. H 37b), and contends that during her tenure in Human Resources, she frequently saw hiring managers circumvent MCMS by requesting that the paperwork of newly hired employees be processed before the job announcements were even posted. (Id. K 37c.)

Dwain Richardson began his employment at MMS and Sodexho in 1995, and is currently the General Manager at INOVA Fair Oaks Hospital in Fairfax, Virginia. (Id. 1110.) Richardson alleges that he repeatedly posted for vacant positions but never heard anything about his applications. (Id. 1138a.) In December 1999, he posted for his current position, but was denied the promotion even though he was the only qualified applicant.¹⁷ (Id. 1138b.)

Tasha Hardy was employed from 1997 to 2000 by MMS and Sodexho, where she worked in the Corporate Services Division in California as a Human Resources Manager. (Id. H11.) Hardy alleges that she was denied the opportunity to seek a higher-level position in the Health- Care Division because her white supervisor refused to allow her to apply for the promotion (Id. 1139a), and is aware of only three African-American managers in the West Coast Corporate Services Division, out of at least 350 managers. (Id. H 39e.) In 2000, Hardy attended the “Diversity Forum for Recruiters,” a national conference for Sodexho Human Resource Managers, where she and other Human Resource Managers agreed that it would be impossible to convince Sodexho’s District Managers, who are predominantly white, to promote more African Americans. (Id.)

Gail Darlington worked in the Health Care Division at MMS and Sodexho from 1988 to 1996. She returned to the company in 1999, and currently works in Georgia in the Corporate Services Division as an Information Systems Consultant. (Id. H12.) Darlington alleges, based on her experience, that “African-American managers in the Health Care Division work primarily at hospitals that serve mostly African-American populations and are staffed predominantly by African-American hourly employees.” (Id. H 43b.) She also contends that she has been passed over for promotions in favor of preselected white applicants, and offers evidence of discriminatory comments by white Sodexho employees. (Id HH 43a, c.)

Samuel Cokes’ employment with MMS dates back to 1975. Although he currently works in the Health Care Division at Howard University Hospital as a Production Manager, Cokes was previously employed for 15 years in the company’s Education Division. (Id. 1113.) Cokes has received only two promotions over his 26 years — -from Unit Supervisor to Production Manager, and from Production Manager to Director of Food Services at Virginia Commonwealth University in Richmond. Cokes claims that he was subsequently removed from his position as Director of Food Services despite good performance, and was replaced by a white employee. He has not been promoted in the 12 years since, although he has continued to seek promotions. (Id. H44a.) Cokes also charges that he has been interested in several positions that had been filled before they were even posted, and has been prevented from advancing at Sodexho by the pervasive discrimination at the company. (Id. 11H 44b-d.)

These ten named plaintiffs filed this action on March 12, 2001. The parties then conducted extensive discovery on the class certification issue before plaintiffs filed their Rule 23 motion. Following a discovery period of almost six months, the parties filed extensive briefs chronicling the facts as developed to date and analyzing the constantly expanding and complex body of law relating to Title VII class actions. The Court wishes to commend counsel for their comprehensive and helpful briefs.

LEGAL ANALYSIS

I. Standard of Review

The standards governing a motion for class certification in a Title VII action are of critical importance. Two cases that bind this Court — Falcon and Wagner — have discussed this issue in detail, and have provided an intricate blueprint for examining such a motion. The analysis begins with Rule 23(a) of the Federal Rules of Civil Procedure, which sets forth four prerequisites to a class action.

One or more members of a class may sue or be sued as representative parties on behalf of all only if (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). If these prerequisites are satisfied, a class action may be maintained if, inter alia,

(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; or

(3) the court finds that questions of law or fact common to the members of the class predominate over any questions affecting individualized members, and that a class action is superior to other available methods for the fair and efficient adjudication of the controversy.

Fed.R.Civ.P. 23(b). Plaintiffs bear the burden of proof on each element of Rule 23. McCarthy v. Kleindienst, 741 F.2d 1406, 1414 n. 9 (D.C.Cir.1984). “A district court exercises broad discretion in deciding whether to permit á • case to proceed as a class action.” Hartman, 19 F.3d at 1471 (citing Bermudez v. Dep’t of Agriculture, 490 F.2d 718, 725 (D.C.Cir.1973)). However, “a decision on class certification cannot be made in a vacuum. While, of course, a court does not possess ‘any authority to conduct a preliminary inquiry into the merits of a suit in order to determine whether it may be maintained as a class action,’ ... some inspection of the circumstances of the case is essential to determine whether the prerequisites of [ ] Rule 23 have been met.” Wagner, 836 F.2d at 587 (quoting Eisen v. Carlisle & Jacquelin, 417 U.S. 156, 177, 94 S.Ct. 2140, 40 L.Ed.2d 732 (1974)).

The level of inspection required in proposed Title VII class actions has been the subject of much debate, particularly with regard to the commonality and typicality prongs of Rule 23(a). As this Circuit noted,

Rule 23 inquiries in the context of Title VII racial discrimination suits have presented courts with a peculiar set of problems. These stem from the fact, as stated by the Supreme Court, that “racial discrimination is by definition class discrimination.” Recognition of this fact has left uncertain the degree of permissiveness tolerable in applying the requirements of Rule 23 in Title VII litigation, and has promoted a difference of opinion as to the proper standard for certifying Title VII classes.

Wagner, 836 F.2d at 588 (quoting Falcon, 457 U.S. at 157, 102 S.Ct. 2364). Prior to Falcon, many courts had adopted an “across-the-board” approach to Title VII class certification, whereby “any case featuring a proposed class composed of all members of a minority group connected in some fashion with a particular employer is deemed to present common questions of law and fact, regardless of individual variations in terms of discriminatory practices suffered or injuries sustained, merely by virtue of an allegation that racial discrimination had occurred.” Wagner, 836 F.2d at 588.

In Falcon, however, the Court rejected this approach as insufficient.

Conceptually, there is a wide gap between (a) an individual’s claim that he has been denied a promotion on discriminatory grounds, and his otherwise unsupported allegation that the company has a policy of discrimination, and (b) the existence of a class of persons who have suffered the same injury as that individual, such that the individual’s claim and the class claims will share common questions of law or fact and that the individual’s claim will be typical of the class claims. For [plaintiff] to bridge that gap, he must prove much more than the validity of his own claim. Even though evidence that he was passed over for promotion when several less deserving whites were advanced may support the conclusion that [plaintiff] was denied the promotion because of his national origin, such evidence would not necessarily justify the additional inferences (1) that this discriminatory treatment is typical of [defendant’s] promotion practices, (2) that [defendant’s] promotion practices are motivated by a policy of ethnic discrimination that pervades [it], or (3) that this policy of ethnic discrimination is reflected in [defendant’s] other employment practices, such as hiring, in the same way it is manifested in the promotion practices.

Falcon, 457 U.S. at 157-58, 102 S.Ct. 2364. Nonetheless, the Falcon Court did identify in a footnote two possible types of across-the-board class actions that could satisfy the requirements of Rule 23.

If [defendant] used a biased testing procedure to evaluate both applicants for employment and incumbent employees, a class action on behalf of every applicant or employee who might have been prejudiced by the test clearly would satisfy the commonality and typicality requirements of Rule 23(a). Significant proof that an employer operated under a general policy of discrimination conceivably could justify a class of both applicants and employees if the discrimination manifested itself in hiring and promotion practices in the same general fashion, such as through entirely subjective decisionmaking processes. In this regard, it is noteworthy that Title VII prohibits discriminatory employment practice s, not an abstract policy of discrimination. The mere fact that an aggrieved plaintiff is a member of an identifiable class of persons of the same race or national origin is insufficient to establish his standing to litigate on their behalf all possible claims of discrimination against a common employer.

Id. at 159 n. 15, 102 S.Ct. 2364 (emphasis in original).

In Wagner, this Circuit interpreted the Court’s decision in Falcon.

What the Court demands from those seeking certification of a class cutting across employment status or job categories is a ‘specific presentation’ identifying the questions of law or fact common to the class representative and the members of the class proposed. The Court has identified two qualifying presentations, the first involving an employer who used a biased testing procedure to evaluate both applicants for employment and incumbent employees seeking promotions, and the second attacking an employer whose general policy of discrimination manifested itself in hiring and promotion in the same general manner, such as through entirely subjective decisionmaking processes. Significant proof of either practice, then, should assure certification of a class containing both actual and potential employees.

Wagner, 836 F.2d at 589 (quoting Falcon, 457 U.S. at 158, 102 S.Ct. 2364). The requisite showing to qualify as “significant proof’ was refined by this Circuit in Hartman, when it noted that “the plaintiff bears the ‘obligation to show, in at least a preliminary fashion, the required commonality between her claims and those of the putative class.’ ” Hartman, 19 F.3d at 1472 (quoting Nelson v. United States Steel Corp., 709 F.2d 675, 680 (11th Cir.1983)). “Nevertheless, a motion for class certification is not an occasion for examination of the merits of the case.” Caridad, 191 F.3d at 291.

Plaintiffs have brought claims for both disparate treatment and disparate impact. In Palmer v. Shultz, 815 F.2d 84 (D.C.Cir.1987), this Circuit explained the difference between the two types of claims.

Under Title VII a plaintiff can rely on either of two different theories to support a claim of unlawful sex discrimination. A “disparate treatment” claim alleges that the defendant intentionally based an employment decision on the sex of the plaintiffs. Disparate treatment claims can involve an isolated incident of discrimination against a single individual, or, as in this case, allegations of a “pattern or practice” of discrimination affecting an entire class of individuals. Id. A “disparate impact” claim alleges that the defendant based an employment decision on a criterion that although “facially neutral” nevertheless impermissibly disadvantaged individuals of one sex more than the other. Palmer, 815 F.2d at 90 (citing International Brotherhood of Teamsters v. United States, 431 U.S. 324, 335-36, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977)). However, “ ‘an important point of convergence’ between disparate treatment and disparate impact claims exists in class actions such as this one. Because both ... claims ‘are attacks on the systemic results of employment practices ... proof of each claim will involve a showing of disparity between the minority and majority groups in an employer’s workforce.’ ” Moore v. Summers, 113 F.Supp.2d 5, 19 (D.D.C.2000) (quoting Segar v. Smith, 738 F.2d 1249, 1267 (D.C.Cir.1984)). The Court will proceed with its Rule 23 analysis with regard to both types of claims.

II. Rule 23(a) Analysis

As noted, plaintiffs seek to certify a class consisting of all current and former African-American Sodexho employees from March 9, 1998 to the present who held or sought to obtain (1) a position at or comparable to the level of above-the-unit, or (2) a job that would lead to such a position, and who have been, are, or may be “adversely impacted by Sodexho’s racially discriminatory policies and practices affecting promotions or advancement.” (Pl.Mot. at 1.) With an estimated 2,000 or more members in the proposed class (Complaint 1159), defendant does not contest the numerosity requirement of Rule 23(a)(1), and the putative class clearly satisfies that standard. See Stewart v. Rubin, 948 F.Supp. 1077, 1088 (D.D.C.1996), affd, 124 F.3d 1309 (D.C.Cir.1997) (245-member class spread throughout the country satisfies numerosity requirement). The commonality, typicality, and adequacy prongs, however, require a more probing analysis.

A. Commonality

Fed.R.Civ.P. 23(a)(2) requires the plaintiffs to show that “there are questions of law or fact common to the class.” As with all other aspects of Rule 23, the trial court must engage in a “rigorous analysis” to determine whether this prerequisite has been fulfilled. Hartman, 19 F.3d at 1473.

As Falcon made clear, there is more to a showing of commonality than a demonstration that class plaintiffs suffered discrimination on the basis of membership in a particular group.... [A] plaintiff need not isolate the particular practice and prove that such practice caused the discrimination, [but] plaintiffs must make a significant showing to permit the court to infer that members of the class suffered from a common policy of discrimination that pervaded all of the employer’s challenged employment decisions.

Id. at 1472. As noted, plaintiffs must make a “specific presentation” that identifies the questions of law or fact common to the class representative and the putative class. Wagner, 836 F.2d at 589.

Plaintiffs’ argument for commonality is threefold. First, they contend that in spite of Sodexho’s decentralized system, the promotion process is the same throughout the company. Second, they assert that this process is unfettered and subjective. Third, they allege that, as a result, the members of the putative class have been affected by these practices in the same way, by being denied promotions for which they were better qualified than the white employees who received them. As recognized by Wagner, 836 F.2d at 592, plaintiffs may rely on a combination of statistical and anecdotal evidence to support these assertions. See Valentino v. USPS, 674 F.2d 56, 68 (D.C.Cir. 1982) (“[C]lass action plaintiffs offer a combination of statistical proof and individual testimony of special instances of discrimination.”).

The Court is satisfied that there are common questions of law and fact with respect to the class and its representatives. Plaintiffs have made a “significant showing” of a “common policy of discrimination” under Hartman, 19 F.3d at 1472. As previously explained, Sodexho’s own witness on its promotion policies and practices confirmed that there are no company rules, guidelines, or criteria for hiring managers to follow, and that these supervisors are not required to take notes on interviews or to explain their choices. (Anstee Dep. at 81, 90-92, 182.) While it may be true, as defendants contend, that “[p]laintiffs challenge tens of thousands of employment decisions ... made by thousands of managers, at thousands of work sites dispersed across the country” (Def.Opp. at 32), plaintiffs argue that these decisions are linked by Sodexho’s company-wide policy of not having any consistent promotion policy, guidelines, or requirements. The testimony of Sodexho’s own witness, as well as of other company employees, confirms that So-dexho’s policy and practice is to allow each manager to make promotion decisions based on whatever criteria he or she chooses, with no guidance or rules.

This anecdotal evidence is buttressed by plaintiffs’ statistical evidence of discrimination in promotions. In particular, plaintiffs’ statistics show that African Americans are significantly underrepresented in above-the-unit positions in comparison with lower-paying jobs at Sodexho, and that if promotions at the company were based on seniority, 265 more African Americans would hold above-the-unit jobs. This data also demonstrates that African-American managers at Sodexho are clustered in a small percentage of units, and are significantly more likely to report'to another African-American manager. Plaintiffs’ statistical expert concluded that this pattern of discrimination was company-wide and did not vary significantly by division. Thus, plaintiffs have made a “significant showing,” Hartman, 19 F.3d at 1472, that Sodexho’s lack of any uniform promotion policy or guidelines has had a disparate impact on the promotion of African-American employees, and has enabled or even fostered an environment at the company in which officials intentionally discriminate against blacks by denying them promotions to upper-level managerial positions.

To rebut this showing of a common policy, defendant offers essentially four arguments, none of which has merit at the class certification stage. First, it contends that its promotions are not determined based on entirely subjective criteria, because some individual hiring managers do consider objective criteria, and because some jobs have objective requirements, including a satisfactory performance evaluation, which in turn is based on objective criteria. (See Def.Opp. at 35-36.)¹⁸ These arguments, however, miss the point of Falcon’s footnote 15, which carved out an exception to the across-the-board rule for “entirely subjective decisionmaking processes,” rather than requiring entirely subjective hiring criteria. Falcon, 457 U.S. at 159 n. 15, 102 S.Ct. 2364 (emphasis added). Whether a particular manager uses objective criteria in making particular promotion decisions is irrelevant to the commonality analysis; instead, what is significant is that the determination of which criteria to use is left entirely to the individual manager. That is, although individual promotion decisions may in isolated cases be made based on some objective criteria, Sodexho’s overall promotion practices remain entirely subjective, because the decision of whether to use objective criteria — and if so, which criteria to use — is a matter left to the discretion of the individual manager.¹⁹ See Griffin v. Dugger, 823 F.2d 1476, 1490 (11th Cir.1987) (“entirely subjective” prong of Falcon implicated where employees “are promoted on the basis of their performance as viewed by their superiors”).

Second, defendant argues that “a literal avalanche of decisions in the past three years has rejected class claims very similar to those of Plaintiffs,” in which the defendant’s decision-making processes were decentralized, or, for example, the putative class encompassed a broad geographic area. (Def.Surr. at 7-8.) In support, defendant cites numerous recent district cases from other jurisdictions in which class certification has been rejected. (See Def.Opp. at 31-32; Def. Surr. at 7-8.) This reliance is unpersuasive. In the first instance, this Circuit has spoken in detail on this issue, and has articulated the appropriate analysis for this Court.²⁰ Defendant’s pleadings are notable for their lack of reference to Wagner and Hartman, the two leading cases in this Circuit. Moreover, motions for class certification are fact-specific, and therefore, the outcome of these motions will vary based on the specific facts of the case. Thus, despite the parties’ protestations to the contrary, these factual distinctions mean that there is really no case on point.²¹ Not surprisingly, then, the cases cited by defendant are distinguishable.²² Plaintiffs have defined their claims and class narrowly to include only discrimination in promotions, and only those individuals who sought either an above-the-unit position or a job that would lead to an above-the-unit position. And, as noted, they have made a preliminary showing through statistics and testimony that Sodexho’s promotion practices with regard to this group of individuals is entirely subjective and perhaps discriminatory.²³ These differences distinguish this action from the cases cited by defendants. See Valentino, 674 F.2d at 68 (“The burden of comparing appropriate groups ... is far more tractable when all members of the class are professional, administrative, or technical employees with generally similar job skills and seek advancement to positions involving those same skills.”)

Perhaps more importantly, defendant’s view of the commonality test would preclude class certification in an action against any company that has decentralized its personnel practices. Sodexho contends that a class cannot be certified “without identifying any centrally unifying general policy of discrimination,” and that there can be no such policy when decision-making procedures are decentralized. (Def.Surr. at 7 (internal quotations omitted).) This position would permit companies to escape Title VII class actions by minimizing the amount of control that they exercise over individual managers. Such a holding would run afoul of the purpose of Title VII, which is “not to provide redress but to avoid harm,” by encouraging employers “to adopt antidiscrimination policies and to educate their personnel on Title VII’s prohibitions.” Kolstad v. American Dental Association, 527 U.S. 526, 545, 119 S.Ct. 2118, 144 L.Ed.2d 494 (1999) (internal quotation omitted).

Third, defendant attacks Siskin’s statistical evidence, both on its merits and as a matter of law. As noted, the Court should not pass on the merits of plaintiffs’ statistical analysis at this stage, because such an analysis is “not relevant to the [class] certification determination.” Caridad, 191 F.3d at 292; see Wagner, 836 F.2d at 593-95 (accepting plaintiffs’ presentation of statistics in evaluating question of commonality). While defendant’s contention that plaintiffs’ statistics are flawed as a matter of law because they “are merely descriptive, do not provide any appropriate benchmarks, and [are] premised on an analysis of seniority alone” (Def.Opp. at 33) must be addressed, it is ultimately unavailing. Sodexho relies on Coward v. ADT Sec. Sys., Inc. 140 F.3d 271, 274 (D.C.Cir. 1998), to support this argument. There, this Circuit held that the “ ‘[m]ajor factors’ that a regression analysis²⁴ must include depend on the facts and theory of the particular case.” Id. at 274 (quoting Bazemore, 478 U.S. at 400, 106 S.Ct. 3000). Using that standard, the Court found plaintiffs’ statistics to be invalid as a matter of law because they controlled only for race and seniority, even though “[ajccording to their own theory of the ease, [ ] job title, or some other measure of type of work, serve[d] as a major factor within the meaning of Bazemore, [because plaintiffs] claim that they should be compared to employees in other job categories who perform similar work but who earn more than they.” Id. Such is not the case here. In the first instance, Coward considered the use of statistics to establish a prima facie case of discrimination on a motion for summary judgment, not to meet the less demanding “significant proof’ standard on motion for class certification. Wagner, 836 F.2d at 589. Moreover, unlike Coward, plaintiffs’ statistics are not “so incomplete as to be inadmissible as irrelevant.” Id. Plaintiffs’ claim is that discrimination at Sodexho has prevented African-American employees from advancing to high-level managerial positions. The two major factors related to this claim are therefore race and seniority, and plaintiffs’ expert has controlled for these variables. Although Siskin did not control for education and experience by position, these factors are not essential to plaintiffs’ argument, and their omission “will affect the analysis’ probativeness, not its admissibility.” ²⁵ Bazemore, 478 U.S. at 400, 106 S.Ct. 3000. Plaintiffs’ statistics are therefore sufficient to support a finding of commonality.²⁶

Fourth, defendant asserts that the continuing violation doctrine defeats commonality. Under Sodexho’s interpretation of the complaint, plaintiffs are seeking damages for injuries that occurred before the limitations period. Whether they would be permitted to do this under the “continuing violation” doctrine would, defendant contends, need to be addressed on an individual basis. Although plaintiffs’ invocation of a “continuing violation” in both the complaint and its motion for class certification is somewhat ambiguous (see Complaint 1190; Pl.Mem. at 26 n. 26), plaintiffs have made clear in their reply brief that they seek damages only for injuries suffered beginning in March 1998 — though the injury may have been caused by a pattern of discrimination that started before then. (PL Reply at 13.) See Bazemore, 478 U.S. at 396, 106 S.Ct. 3000. To the extent that evidence of discrimination from before March 1998 is relevant to the determination of liability, there are questions of law and fact common to the class, because this evidence would be offered to prove the ongoing practice of discrimination that plaintiffs have alleged. This continuing violation doctrine thus has no effect on the question of commonality with respect to liability.²⁷

B. Typicality

Next, plaintiffs must show that “the claims or defenses of the representative parties are typical of the claims or defenses of the class.” Fed.R.Civ.P. 23(a)(3). “The commonality and typicality requirements of Rule 23(a) tend to merge.” Falcon, 457 U.S. n. 13, 102 S.Ct. 2364. “[T]be court must consider whether [the class representatives] suffered injury from a specific discriminatory promotional practice of the employer in the same manner that the members of the proposed class did, and whether [the class representatives] and the class members were injured in the same fashion by a general policy of employment discrimination.” Wagner, 836 F.2d at 591; see Paxton v. Union Nat’l Bank, 688 F.2d 552, 562 (8th Cir.1982) (“Typicality is not defeated because of the various promotional opportunities at issue, or the differing qualifications of plaintiffs and class members.”)²⁸

The analysis of commonality applies equally to typicality. Because plaintiffs have made a convincing showing of a common policy of discriminatory treatment that extends across divisions, units, and geographic regions, including those where the proposed representatives worked, they have made a sufficient demonstration that the claims of the class representatives are largely typical of those of the class.

Plaintiffs have failed, however, to show typicality in one respect. The ten class representatives worked at MMS before it merged with SUSA on IVlarch 27, 1998, and then continued their employment with the merged entity. None of them worked at SUSA. Yet plaintiffs’ class allegations date from March 9, 1998 — 18 days before the formation of Sodexho. While plaintiffs have satisfied their burden of showing a common policy of discrimination with respect to their own claims and those of the class members who worked at Sodexho, they have offered no evidence about the policies and practices of SUSA except for a limited description of that entity’s posting system. Moreover, plaintiffs’ evidence of discrimination before the merger is somewhat limited. Plaintiffs have not explained whether their statistics that predate the merger apply to MMS, to SUSA, or to both companies, and the majority of plaintiffs’ anecdotal evidence of discrimination describes events occurring after March 27, 1998. Without more, the Court is unwilling to infer that the statistical and anecdotal evidence of discrimination put forth by the plaintiffs with regard to Sodexho applies as well to MMS and SUSA. And since no named plaintiff worked at SUSA, there is no representative of the putative class who was employed at that entity before the merger. The Court therefore finds that the class representatives’ claims are typical of those of all members of the class following the formation of Sodexho on March 27, 1998, but not of claims that predate the merger.²⁹ As a result, the Court will limit the class period to include only claims against Sodexho beginning on March 27,1998.³⁰

C. Adequacy

“The adequacy inquiry under Rule 23(a)(4) serves to uncover conflicts of interest between named parties and the class they seek to represent. A ‘class representative must be part of the class and possess the same interest and suffer the same injury as the class members.’ ” Amchem Products, Inc. v. Windsor, 521 U.S. 591, 625-26, 117 S.Ct. 2231, 138 L.Ed,2d 689 (1997) (quoting East Tex. Motor Freight System, Inc. v. Rodriguez, 431 U.S. 395, 403, 97 S.Ct. 1891, 52 L.Ed.2d 453 (1977)) (internal citations omitted). This Circuit has explained that “[b]asic consideration of fairness require[s] that a court undertake a stringent and continuing examination of the adequacy of representation by the named class representatives at all stages of the litigation where absent members will be bound by the court’s judgment.” National Association for Mental Health, Inc. v. Califano, 717 F.2d 1451, 1457 (D.C.Cir.1983) (citing National Association of Regional Medical Programs, Inc. v. Mathews, 551 F.2d 340, 344-45 (D.C.Cir.1976)). In evaluating whether a named plaintiff will fairly and adequately protect the interests of the class members, this Circuit has “considered two principal requirements: 1) the named representative must not have antagonistic or conflicting interests with the unnamed members of the class, and 2) the representatives must appear able to vigorously prosecute the interests of the class through qualified counsel.” Id. at 1458. Defendant attacks only the first prong of this test, since they do not dispute counsel’s impressive experience and qualifications.

The leading case in this Circuit on the question of conflict of interest is Phillips v. Klassen, 502 F.2d 362 (D.C.Cir.1974). There, the Circuit traced the development of the use of a class action as background for its discussion of the necessity of adequate representation by the lead plaintiffs.

Class members whose interests are antagonistic in fact to, or even “potentially conflicting” with, the interests of the ostensibly representative parties cannot be bound, consistent with the requirements of due process to an adjudication taken in their name. Where courts discern that the interests of the named plaintiff are in significant part antagonistic to' those of the class he purports to represent, they decline to entertain the action as a class action.

Id. at 366 (quoting Hansberry v. Lee, 311 U.S. 32, 41 — 42, 61 S.Ct. 115, 85 L.Ed. 22 (1940)) (internal citation omitted). Given the difficulty in divining the interests of the absent class members, the Phillips Court instructed trial courts to focus on the remedies sought by the named plaintiffs and to determine whether the same relief would also likely be desired by the rest of the class.

In order to maintain a class suit, plaintiffs must be truly representational. Unless the relief sought by the particular plaintiffs who bring the suit can be thought to be what would be desired by other members of the class, it would be inequitable to recognize plaintiffs as representative, and a violation of due process to permit them to obtain a judgment binding absent plaintiffs.

Id. (quoting Dierks v. Thompson, 414 F.2d 453, 456 (1st Cir.1969)) (internal citation omitted).

Implicit in Phillips is the requirement that only those conflicts of interest that relate to the subject matter of the lawsuit are relevant to the adequacy determination. This principle has been widely recognized, see, e.g., Johns v. Rozet, 141 F.R.D. 211, 217 (D.D.C.1992); In re Cardizem CD Antitrust Litigation, 200 F.R.D. 297, 306 (E.D.Mich.2001); Caranci v. Blue Cross & Blue Shield of Rhode Island, 194 F.R.D. 27, 41 (D.R.I.2000); Walsh v. Northrop Grumman Carp., 162 F.R.D. 440, 445—47 (E.D.N.Y.1995); Riordan v. Smith Barney, 113 F.R.D. 60, 64 (N.D.Ill.1986), including in discrimination class actions. See Allen v. Isaac, 99 F.R.D. 45, 55 (N.D.Ill. 1983); Turner v. A.B. Carter, Inc., 85 F.R.D. 360, 364 (E.D.Va.1980). Sodexho contends that the proposed class representatives are inadequate because some of them are “claiming race discrimination and seeking damages based on the actions [of] other class representatives or class members.” (Def.Opp. at 13.) Specifically, defendant argues that several of the named plaintiffs seek damages for discrimination involving decisions by other proposed representatives or African-American managers. (Id.) Sodexho also asserts that the named plaintiffs and members of the putative class have necessarily given performance evaluations that impacted promotion decisions. (Def.Surr. at 9.) Defendant alleges that these conflicts are insurmountable, and that the proposed representatives are therefore inadequate.

Sodexho relies on Wagner, in which this Circuit affirmed the lower court’s denial of a motion to certify a class of all “professional, administrative, and technical employees graded GS-9 and above” at the Interstate Commerce Commission. Wagner, 836 F.2d at 581. Although the Court rested largely on plaintiffs’ failure to establish commonality, it also questioned the adequacy of the proposed representatives due to potential conflicts within the class.

Supervisory employees are often inappropriate representatives of nonsupervisory employees because the structure of the workplace tends to cultivate distinctly different interests between the two groups. Although each group shares the interest in freedom from discrimination, potential conflicts may and do arise within a class including both. In the instant case, supervisors who are members of the proposed class have been responsible for evaluating the performances of other members of the class; in.deed, the very individual on whose behalf injunctive relief is sought in this case was allegedly discharged by a black employee who is a potential member of the class. Additionally, [plaintiff] has accused his own supervisor, who is a potential class member, of racial discrimination....

Id. at 595. The holding of Wagner has been interpreted to mean that “the existence of a supervisory relationship between class members could undermine the adequacy of the representation of the class members in that relationship.” In re PEPCO Employment Litigation, 1992 WL 442759, at *22 (D.D.C. Dee.4, 1992) (emphasis added). In PEPCO, one proposed class representative alleged that his “black supervisor, a member of the class, diseriminatorily denied him promotions, and [a second] proposed class representative [] allegefd] his supervisor, [the first proposed class representative], discrimi-natorily denied him promotions arid raises.” Id. Faced with these existing conflicts, the court refused to certify the proposed class.

The allegations presented here differ significantly from the facts in Wagner and PEPCO. Although their deposition testimony is confusing in spots, plaintiffs have not alleged discriminatory treatment by other potential class members with respect to promotions. (Compare McNish Dep. at 13-29 with 179-80; Howell Dep. at 15-16, 102-03 with 206-07; Henry Dep. at 26-29 %oith 50-51, 71-72; Hardy Dep. at 64-67 with 72-73.) To the extent that plaintiffs, as defendant notes, “supervised other African-American class members[,] bore significant [human resources] responsibilities over them, ranging from performance evaluations to discipline and termination^] and ... reported to African-American managers at some time” (Def.Opp. at 14), these work relationships do not rise to the level of conflicts that preclude the named plaintiffs from adequately representing the interests of the class. The mere fact that some putative class members were involved in the supervision and' rating of other class members does not mean that the supervising class members perpetuated or contributed to any of Sodexho’s alleged discriminatory policies.³¹ See Hyman v. First Union Corp., 982 F.Supp. 1, 5 (D.D.C.1997) (adequacy element satisfied where no plaintiffs brought claims of discrimination against other plaintiffs and “even if the evaluations [of some supervisor plaintiffs] had an[] impact on the [] decisions [to terminate other class members], it was a negligible impact”).

In light of the absence of specific allegations of discrimination by class members against other class members, and because “the relief sought by the particular plaintiffs who bring the suit can be thought to be what would be desired by other members of the class,” Phillips, 502 F.2d at 366, the holding of Judge Lamberth in Neal v. Moore, 1994 U.S.Dist. LEXIS 21339 (D.D.C. Dec. 23, 1994), applies. “[Supervisors and non-supervisors alike will be able to come forward with evidence of their specific injuries and appropriate relief. Furthermore, an injunction against a few supervisory members of the class — who most likely did not exert significant influence over departmental policy-making — is fairly characterized as de minimis relative to the value of such an injunction in protecting those same supervisors from epidemic discrimination.” Id. at *46-47. In fact, Neal applies with particular force here because plaintiffs’ allegations— which the Court has found to be sufficiently substantiated to support a finding of commonality — are that Sodexho’s discriminatory practices have subjected African Americans to a glass ceiling. To thwart a class action because of the presence of both supervisors and non-supervisors in the class would actually reward defendant for its alleged discrimination.³²

III. Rule 28(b) Analysis

Plaintiffs have moved to certify the class under Fed.R.Civ.P. 23(b)(2), or alternatively, under Fed.R.Civ.P. 23(b)(3).³³ Central to the Rule 23 analysis are the remedies sought by plaintiffs. In this case, they have requested a declaratory judgment; permanent injunctions significantly restructuring Sodexho’s system for promoting employees and implementing anti-discrimination initiatives; ³⁴ job relief, such as reassignment, for plaintiffs and class members; and monetary relief, including back pay, and compensatory and punitive damages. (Complaint at 51-54.)

As noted, Rule 23(b)(2) permits class certification if “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.” Fed.R.Civ.P. 23(b)(2). “While the rule itself is silent as to whether and to what extent monetary relief may also be sought, the Advisory Committee Notes on Rule 23 state that (b)(2) certification ‘does not extend to cases in which the appropriate final relief relates exclusively or predominantly to money damages.’ ” Taylor v. District of Columbia Water & Sewer Authority, 205 F.R.D. 43, 47 (D.D.C.2002) (quoting Fed. R.Civ.P. 23 (Advisory Committee Notes)). The Notes also “explicitly state that ‘cases in the civil-rights field’ are ‘illustrative’ of the class actions meant to be certified under Rule 23(b)(2).” Stewart v. Rubin, 948 F.Supp. 1077, 1089 (D.D.C.1996) (quoting Fed.R.Civ.P. 23 (Advisory Committee Notes)); see Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 614, 117 S.Ct. 2231, 138 L.Ed.2d 689 (1997). Consequently, “Title VII and other civil rights class actions are frequently certified pursuant to Rule 23(b)(2).” Eubanks v. Billington, 110 F.3d 87, 92 (D.C.Cir.1997).

At this stage, it is appropriate to certify this class pursuant to Rule 23(b)(2) for the purpose of determining liability. Any finding of liability is tied to the actions of Sodexho with respect to the class as a whole, and any subsequent injunctive or declaratory relief would have a significant impact on the company’s policies and practices in relation to the entire plaintiff class. At the same time, however, plaintiffs’ request for damages — including back pay — makes (b)(2) certification of the remedy phase unwise at this stage. As this Circuit noted, “the underlying premise of (b)(2) certification — that the class members suffer from a common injury that can be addressed by classwide relief — begins to break down when the class seeks to recover back pay or other forms of monetary damages to be allocated based on individual injuries.” Eubanks, 110 F.3d at 95. The Court will therefore bifurcate the case, certifying a(b)(2) class for liability. If liability is established, the Court will then determine the most appropriate mechanism for determining remedies.³⁵ See Eubanks, 110 F.3d at 96 n. 14; Pigford v. Glickman, 182 F.R.D. 341, 351 (D.D.C.1998); Neal, 1994 U.S.Dist. LEXIS 21339, at *42-44, *57.

CONCLUSION

For the aforementioned reasons, plaintiffs’ motion for class certification is granted as to liability, although the class period will be limited to begin with the formation of Sodex-ho on March 27,1998. ■

1 . 457 U.S. 147, 102 S.Ct. 2364, 72 L.Ed.2d 740 (1982).

2 . In June 2001, Sodexho Alliance, S.A. purchased the remaining shares of Sodexho Marriott and changed the name of the company to "So-dexho, Inc."

3 . This division is identified in the complaint as "Education.” (Complaint V 17.)

4 . Sodexho contends that its "distinct divisions, and the services each provide, reflect real differences in management functions and competencies. For example, ... managers in Health Care accounts often are required to have more sophisticated knowledge of dietary and nutritional planning than in other accounts [and], due to regulatory compliance issues, [are] intimately and actively involved in the administration of the client hospital. Corporate Services accounts, in contrast, have no correlative regulations ... but tend to provide a lower volume, but higher quality, of food services to its clients.” (Def.Opp. at 6 (internal citations omitted).) However, as discussed infra at note 23, whether these distinctions exist is of little relevance to the class certification motion.

5 . No information has been provided to the Court about the nature of the "Canada” division. Defendant also characterizes Sodexho as having six divisions, but states that the sixth division is its corporate headquarters, rather than the Canada unit. (Def.Opp. at 6.)

6 . Defendant does not dispute that vacancies are posted company-wide, but contends that in practice, promotions occur almost entirely within a division, noting that only four percent of employees move between divisions in any given period. (Def.Opp. at 17-18; Def.Ex. 10, Report of Joan Haworth ("Haworth Rep.”) 11 68.) In light of plaintiffs' claim that 28 percent of all applicants for open positions come from outside the division in which the job is posted (PI. Reply at 12-13; Pl.Ex. 50, Supplemental Report of Dr. Bernard Siskin ("Siskin Supp.Rep.”) at 9 n. 8), defendants’ statistics about the lack of promotions among divisions may actually support plaintiffs’ allegations of preselection for open positions. See infra pp. 433.

7 . SUSA had a similar system in place before the merger. Job openings were posted through a central administration point, and individual managers wrote instructions describing how to apply for each position. SUSA’s system did not have the "ability to track candidates.” (See PI. Ex. 18, AT 1.)

8 . Some of the anecdotal evidence of discrimination offered by plaintiffs predates the class period. Nonetheless, this testimony may still be relevant to a finding of liability and to injuries suffered during the class period if plaintiff shows an ongoing pattern of discrimination. See infra note 16.

9 . Defendant's argument that the Court should weigh the statistical analysis offered by the defendant against that of the plaintiffs at the class certification stage lacks support in the case law. Neither of the two cases that defendant relies on for this proposition addressed a motion for class certification; rather, both cases considered conflicting statistics on a dispositive motion. See Coward v. ADT Sec. Sys., Inc., 140 F.3d 271 (D.C.Cir.1998); Koger v. Reno, 98 F.3d 631 (D.C.Cir.1996).

10 . This Circuit has noted that ”[w]hile statistics can generally be probative of the [questions presented under the Rule 23 analysis, it] would feel uncomfortable in resting on the trial statistics in the present record for a final determination of commonality.” Hartman v. Duffey, 19 F.3d 1459, 1473 (D.C.Cir.1994). In Hartman, however, the factual record contained a "paucity" of statistical evidence, and no real anecdotal evidence of discrimination. Id. at 1473-74. In contrast, the record before this Court is thorough and well-developed, making certification based on anecdotal and statistical evidence appropriate.

11 . Although the class period extends back only to 1998, Siskin’s analysis on the effects of the promotion process begins in 1995, which is the beginning of "the period for which [he had] data to identify promotions.” (Siskin Rep. It 7; see Pl.Mem. at 26 n. 26)

12 . Standard deviations above 1.96 units have been found sufficient to establish prima facie cases of both disparate treatment and disparate impact. See Anderson v. Zubieta, 180 F.3d 329, 339-40 (D.C.Cir.1999).

13 . For example, 29 of Sodexho's 33 managers at Tuskegee University, Florida A & M, Howard University, Prairie View A & M, and North Carolina A & T — all of which are historically black colleges — are African-American, whereas only 4 of its 72 managers at American University, Marquette University, Wellesley College, Providence College, and Harvard Law School are black. (Id. 114.)

14 . In his earlier report, Siskin analyzed 5,226 promotions; in the supplemental study, he examined 10,002 promotions. (Id. 11 6.)

15 . There are no named plaintiffs representing either the Laundry or Canada division. See infra note 30.

16 . Although plaintiffs do not seek damages for this and other incidents that occurred before March 1998, evidence of these events may be relevant to injuries suffered after that time if the injuries were caused by an ongoing pattern of discrimination. See Bazemore v. Friday, 478 U.S. 385, 396, 106 S.Ct. 3000, 92 L.Ed.2d 315 (1986).

17 . According to the Complaint, Richardson was promoted to this position six months later only after Sodexho’s client intervened and insisted on the promotion. (Id. H 38c.)

18 . Defendant’s examples of objective criteria include, inter alia, testimony of a supervisor that ’’[i]f I'm hiring for a management position, I ask them how many years of management have you had. If they say zero, that screened them out” (Def.Ex. 40, Dep. of Ezzie Henry (“Henry Dep.”) at 88); the testimony of plaintiff Moss that education and supervisory experience are considered in some promotion decisions for salaried positions (Def.Ex. 19, Dep. of Chip Moss ("Moss Dep.”) at 75-76); and the fact that certain job descriptions and the general MCMS guidelines include some objective criteria. (Def.Ex. 50-51 (MCMS guidelines state that employee "can only post for a job that is no more than 2 grade levels higher than your current grade [and] must have been in your current position a minimum of one year”).) This final point about MCMS in particular misses the mark because it ignores plaintiffs' argument — and the statistics in support of this argument — that supervisors were not required to adhere to MCMS in making promotion decisions. See supra pp. 432.

19 . Defendant's interpretation would, in effect, eviscerate the "entirely subjective” exception in footnote 15 of Falcon, because it would find commonality only in an entity in which a defined group of personnel actions existed in which every decision was based on purely subjective criteria.

20 . For example, this Circuit noted that plaintiffs will often offer a combination of statistical and anecdotal evidence to demonstrate a common pattern of discrimination, and concluded that “it is hard to imagine how it could be otherwise." Wagner, 836 F.2d at 592-93. Defendant’s reliance on two Northern District of Georgia cases, therefore, for the proposition that anecdotal evidence is not probative of discrimination is unhelpful, because the views expressed in those opinions conflict with the law of this Circuit. (See Def.Opp. at 39-40) (citing Cooper v. Southern Co., 205 F.R.D. 596, 626-27 (N.D.Ga.2001); Reid v. Lockheed Martin Aeronautics Co., 205 F.R.D. 655, 667-68 (N.D.Ga.2001).)

21 . As a result, and in light of the large number of Title VII class actions filed over the past several years, it is hard to imagine an area in which it is easier for the parties to find district court case law to cite for almost any proposition. Both parties have certainly done this. In response to the 15 district court cases denying class certification from the last three years cited by defendant (Def.Opp. at 31-32), plaintiffs have identified five other cases where commonality was found (PI. Reply at 3-5), and they have also effectively distinguished this case from the ones cited by defendant. (Id. at 6-9.)

22 . The Court need not address all of the cases cited by defendant, because it is governed by Wagner and Hartman. However, it is notable that defendant relies heavily on Allen v. Chicago Transit Authority, 2000 WL 1207408 (N.D.Ill. July 31, 2000), for the proposition that "[tjhere are too many individual issues for there to be factual or legal issues common to the entire class.” Id. at *8. Notwithstanding the difference between the interpretation of Falcon by courts in the Seventh Circuit, see id. at *8, and this Circuit, Allen is distinguishable from the instant action because in that case, 1) the fact that defendant's "managers [were] required to put the reasons for their decisions in writing,” id. at *9, ensured that the decision-making process was not entirely subjective and unreviewable; 2) plaintiffs also claimed discrimination in pay, which made the commonality requirement more difficult to fulfill; and 3) plaintiffs appear to have offered no anecdotal evidence of discrimination beyond the specific claims of the class representatives, instead relying largely on statistics of under-representation. Id. at *9-10.

23 . This showing renders irrelevant to the class certification motion any substantive differences in the work done by employees of the different divisions at Sodexho, especially when combined with plaintiffs’ evidence that employees can and do apply for promotions across divisions.

24 . Regression analysis is "a statistical method for making quantitative estimates of the effects of different factors on some variable of interest. The regression technique has come into vogue in employment discrimination cases as a means of estimating the discrete influence factors such as sex, experience, and education have had on determining salary level.” Valentino, 674 F.2d at 70 (internal quotation and citation omitted).

25 . In Bazemore, on which this Circuit relied in Coward, the Supreme Court considered the use of statistical evidence in a discrimination case at trial. The Court found that "a regression analysis that includes less than all measurable variables may serve to prove a plaintiff’s case.” Bazemore, 478 U.S. at 400, 106 S.Ct. 3000. Moreover, plaintiffs explain that they failed to control for education and experience because Sodexho has no computerized data that would allow them to be evaluated quantitatively. (PI. Reply at 17 n. 26; Def.Ex. 11, Haworth Response to Siskin Supp.Rep. ("Haworth Resp.”) 1128 n. 14.) Here again, plaintiffs should not be penalized for defendant’s lack of centralized controls and recordkeeping.

26 . Relying on Moore and Valentino, defendant also challenges plaintiffs’ statistics on the ground that they confuse under-representation for discrimination. In Moore, plaintiffs based their discrimination claim on a comparison of the percentage of African Americans at one salary level with those at the next highest level. The court found no likelihood of success on the merits in large part because the analysis from that data assumed, without basis, that all African Americans at the lower level were eligible for promotion. Moore, 113 F.Supp.2d at 20. In Valentino, the Court found that plaintiffs had not established a prima facie case of classwide gender discrimination because the statistics "generally repeat[ed] the same theme: ... there are a larger percentage of women in the lower levels than in the upper levels, ... [but] provide[d] no basis for comparing men and women similarly educated and equipped to pursue the same occupations.” Valentino, 674 F.2d at 69-70 (internal quotation omitted). See Metrocare v. WMATA, 679 F.2d 922, 930 (D.C.Cir.1982) (statistics failed to establish a prima facie case of race discrimination because they failed to "compare the percentage of blacks hired for given jobs with the percentage of blacks qualified for those positions”). However, whether that critique applies to plaintiffs’ statistical evidence is not determinative of a motion for class certification, because at this stage plaintiffs' burden "entails more than the simple assertion of commonality and typicality, but less than a prima facie showing of liability.” Reid, 205 F.R.D. at 666 (internal quotation omitted). In both Valentino and Metrocare, the courts rejected plaintiffs’ statistical evidence at trial, long after plaintiff classes had been certified; in Moore, the court was determining the likelihood of success on the merits at the preliminary injunction stage. Whether plaintiffs can ultimately make out a prima facie case of classwide discrimination is not the issue here; rather, they must make only a “significant showing” to that effect. Hartman, 19 F.3d at 1472.

27 . With regard to the remedial phase, events occurring before March 1998 may affect the damages of different class members in different ways, which could in turn impact the Court's ability to certify the class under Rule 23(b) for the purpose of determining a remedy. Because the class is being certified only for liability at this time, however, the Court need not address the effect of the continuing violation doctrine on the class for remedial purposes.

28 . Contrary to plaintiffs’ assertion (see Pl.Mem. at 37), the statement articulated in Paxton was not adopted by this Circuit, but was instead merely noted in Wagner in a footnote as a “see also” cite. Wagner, 836 F.2d at 591 n. 82.

29 . For the same reasons, plaintiffs have failed to show commonality with respect to claims that arose before the merger. Moreover, as the unre-butted testimony offered by defendant of Sodex-ho’s Vice President of Human Resources indicates, MMS and SUSA operated under disparate payroll and pay scale systems prior to the merger. (Newmier Decl. 1Í1Í 6-7.) The administrative burden of extending the class period back to include the predecessor companies for 18 days before the merger therefore far outweighs any benefit to the plaintiffs, and would necessarily present "undue [] complication in the presentation of evidence.” Fed.R.Civ.P. 23(d)(1).

30 . Although none of the named plaintiffs worked in the Laundry or Canada divisions, the Court finds that the claims of the proposed representatives are typical of employees of those divisions, because, as noted, plaintiffs have demonstrated that the discriminatory practices they challenge occur throughout the company.

31 . Moreover, defendant’s suggestion that the adequacy prong could not be met where "class members competed against other class members for promotions” (Def.Opp. at 43), has been characterized as an "absurd proposition [that] would of course doom almost every class action charging discrimination in promotion or hiring....” Hartman v. Duffy, 158 F.R.D. 525, 547 (D.D.C. 1994) (citing Meiresonne v. Marriott Corp., 124 F.R.D. 619, 625 (N.D.Ill.1989)). Indeed, defendant’s view of adequate representation is so narrow that the only potential Title VII class that could be certified would have to be composed of black employees on exactly the same pay scale, none of whom ever supervised — or were supervised by — another black employee. This view is plainly far too restrictive.

32 . Defendant's reliance on Wagner, PEPCO, and the other post -Wagner cases it cites in support of its adequacy argument — Donaldson v. Microsoft, 205 F.R.D. 558, 568 (W.D.Wa.2001); Allen v. Chicago Transit Authority, 2000 WL 1207408, at *11, 2000 U.S.Dist. LEXIS 11043, at *36 (N.D.Ill.2000); and Appleton v. Deloitte & Touche L.L.P., 168 F.R.D. 221, 233 (M.D.Tenn.1996) — is misplaced for another reason. None of these cases denied class certification solely on the basis of adequacy by intraclass conflicts. In each case, the Rule 23(a)(4) analysis was merely "another factor that weighted] against certification.” Appleton, 168 F.R.D. at 233.

33 . Because the Court will certify a class under Rule 23(b)(2) for the purpose of determining liability, it need not address the parties’ Rule 23(b)(3) arguments.

34 . Some of the aspects of the proposed injunction — regarding, for example, "the selective and arbitrary use of merit bonuses" (Complaint at 52, UB(8)) — appear to go beyond the scope of the classwide discrimination alleged by plaintiffs.

35 . At the remedy phase, for example, "the court may adopt a 'hybrid' approach, certifying a(b)(2) class as to the claims for declaratory or injunc-tive relief, and a(b)(3) class as to the claims for monetary relief ... [or] the court may conclude that the claims of particular class members are unique or sufficiently distinct from the claims of the class as a whole, and that opt-outs should be permitted on a selective basis.” Eubanks, 110 F.3d at 96. See Thomas v. Albright, 139 F.3d 227, 234-36 (D.C.Cir.1998). Because the Court will not certify the class for the remedial stage at this time, it need not address defendant’s argument that this Court should follow the Fifth Circuit’s decision in Allison v. Citgo Petroleum Corp., 151 F.3d 402 (5th Cir.1998), which held that "[m]onetary relief predominates in (b)(2) class actions unless it is incidental to requested injunctive or declaratory relief.” Id. at 415.