Campbell v. Nat'l R.R. Passenger Corp.

311 F. Supp. 3d 281

• Court: Court of Appeals for the D.C. Circuit
• Decided: April 26, 2018
• Docket: Civil Action No. 99–2979 (EGS); Civil Action No. 01–1513 (EGS)
• Status: Published

Opinion

Emmet G. Sullivan, United States District Judge

Plaintiffs-seventy-one African-American current or former employees or applicants *286for employment at defendant National Railroad Passenger Corporation ("Amtrak")-allege that Amtrak engaged in racial discrimination in its hiring, promotion, and disciplinary practices and created a hostile work environment. Plaintiffs bring this lawsuit on behalf of themselves and more than 11,000 African-American unionized Amtrak employees, former employees, and applicants for employment at Amtrak.

Pending before the Court are plaintiffs' motion for class certification, Amtrak's motions to exclude a number of plaintiffs' experts, Amtrak's motion to strike portions of the declarations filed by plaintiffs in support of class certification, Amtrak's motion to strike portions of plaintiffs' reply in support of their motion for class certification, and Amtrak's motion for partial summary judgement. As explained more fully below, because plaintiffs' class definitions make membership in plaintiffs' proposed class contingent on individualized merits determinations, and because plaintiffs have failed to meet their burden to establish that the claims of all class members are susceptible to common proof, plaintiffs' motion for class certification is DENIED . In addition, Amtrak's motion to exclude Jay Finkelman's expert report and testimony is GRANTED , Amtrak's motion to exclude Thomas Roth's expert report and testimony is DENIED , Amtrak's motion to exclude Edwin Bradley and Liesl Fox's expert report and testimony is DENIED , Amtrak's motion to strike portions of plaintiffs' declarations is GRANTED in part , Amtrak's motion to strike portions of plaintiffs' reply brief is GRANTED in part and DENIED in part , and Amtrak's partial motion for summary judgment is GRANTED .

In Part I of this opinion, the Court sets forth the procedural history of this litigation. Part II sets forth factual background regarding Amtrak's structure, hiring and promotions decisions, disciplinary system, and work environment. In Parts III and IV, the Court analyzes the admissibility of various experts and other evidence offered in support of plaintiffs' motion for class certification. Part V discusses whether class certification is warranted in this case and, finally, Part VI resolves Amtrak's partial motion for summary judgment on plaintiffs' disparate-impact claims.

I. PROCEDURAL HISTORY

A. The Initial And Amended Complaints

This employment discrimination class-action was filed on November 9, 1999 on behalf of current and former African-American employees of Amtrak's Intercity Strategic Business Unit or applicants for employment in that unit. Compl., ECF No. 1. Plaintiffs alleged claims for violations of the Civil Rights Act of 1866, 42 U.S.C. § 1981, and violations of Title VII of the Civil Rights Act of 1964 ("Title VII"), 42 U.S.C. § 2000e, against Amtrak and a myriad of labor unions representing certain plaintiffs. Id. An amended complaint was filed on March 13, 2000, adding a number of named plaintiffs and a handful of labor unions as defendants. First Am. Compl., ECF No. 30.

B. The Court Adjudicates A Series Of Dispositive Motions

The first round of dispositive motions was filed in May 2000 in response to the amended complaint. Although a number of labor-union defendants answered the amended complaint, a few moved to dismiss on the ground that the labor unions were not "indispensable parties" to the litigation and would be better joined in the liability phase of the lawsuit if plaintiffs prevailed on their discrimination claims against Amtrak. See Union Defs.' Mot. to Dismiss, ECF No. 40; Union Defs.' Mot. to *287Dismiss, ECF No. 48. Amtrak also moved to dismiss plaintiffs' class claims, arguing that no amount of discovery would render plaintiffs' proposed classes certifiable under Federal Rule of Civil Procedure 23. See Def.'s Mot. to Dismiss, ECF No. 47. Amtrak moved separately to dismiss the individual claims of plaintiffs on a variety of grounds or, in the alternative, for a more definite statement of those claims. See Def.'s Mot. to Dismiss, ECF No. 50.

Shortly after those motions were briefed, plaintiffs moved for a temporary restraining order and preliminary injunction in order to enjoin Amtrak from "discriminating, disciplining, intimidating, or in any other way retaliating" against plaintiffs and class members. See Pls.' Mot. for TRO/PI, ECF No. 51. The Court denied the request for temporary injunctive relief on June 12, 2000. See Order, ECF No. 62. Thereafter, the Court granted the motions of the union defendants to be dismissed from the case, subject to their being rejoined in the event plaintiffs are successful on their liability claims and the union defendants are necessary to the finalization of an appropriate remedy. See Order, ECF No. 63; Order, ECF No. 64.

A second amended complaint, filed August 22, 2000, added one named plaintiff and eliminated the labor-union defendants. See Second Am. Compl., ECF No. 79. On January 26, 2001, the Court denied Amtrak's motion to dismiss plaintiffs' class claims. See Mem. Op. and Order, ECF No. 92. The Court determined that dismissal of the class claims was premature given the early stage of the proceedings, particularly because additional discovery could permit plaintiffs to correct any fatal flaws in their class definition. Id. at 3.¹ Later that year, the Court denied Amtrak's motion to dismiss plaintiffs' individual claims. See Campbell v. Amtrak , 163 F.Supp.2d 19 (D.D.C. 2001). In so doing, the Court rejected all four of Amtrak's arguments for dismissal, namely that: "1) certain 42 U.S.C. § 1981 claims are barred by the statute of limitations; 2) claims of plaintiffs who previously filed a charge involving the same conduct complained of here, but failed to sue, are barred by the statute of limitations in their right-to-sue letters; 3) certain Title VII claims are barred by the statute of limitations; and 4) claims which do not allege a timeframe fail to state Title VII claims." Id. at 21. The Court granted in part, however, Amtrak's motion for a more definite statement, ordering "plaintiffs to include dates of alleged events, to the extent possible, in an amended complaint" and "to amend their pleading to include a more appropriate term to define the class, so as to exclude from the class definition the salaried managerial and professional positions that were included within the scope of the McLaurin class action discrimination case against Amtrak." Id. at 28.

Plaintiffs filed a third amended complaint on January 3, 2002 to address the concerns set forth in the Court's dismissal Order. See Third Am. Compl., ECF No. 100. On May 27, 2002, plaintiffs filed the fourth amended-and currently operative-complaint. See Fourth Am. Compl., ECF No. 145. The complaint was amended in response to a decision by the parties to merge twenty-one discrimination lawsuits filed by current and former Amtrak employees in the Eastern District of Louisiana into the putative Campbell classes. See Pls.' Mem. in Supp. of Mot. to Amend, ECF No. 143 at 3-4. The parties also agreed to add one plaintiff from the Louisiana actions-Joseph McDonald-as a named plaintiff in this action. See id. at 4-5.

*288On February 4, 2002, Amtrak moved to dismiss some of the individual claims contained in the third amended complaint, see Def.'s Mot. to Dismiss, ECF No. 104, which it supplemented in response to the fourth amended complaint on August 28, 2002, see Def.'s Suppl. Mem. in Supp. of Mot. to Dismiss, ECF No. 127. On September 26, 2002, the Court denied Amtrak's motion. See Campbell v. Amtrak , 222 F.Supp.2d 8 (D.D.C. 2002). Amtrak had sought to dismiss one plaintiff's claims on the grounds that the continuing-violations theory could not save those claims from being barred by the statute of limitations, to dismiss six other plaintiffs' claims as "based on expired right-to-sue notices," and to dismiss the claims of three other plaintiffs as barred by the settlement of another class-action lawsuit. See id. at 9. In denying Amtrak's motion to dismiss, the Court found that the continuing-violations theory could bring one plaintiff's claims within the statutory period, that further factual development was required to determine whether other plaintiffs were entitled to equitable tolling of the statute of limitations, and that plaintiffs' claims were not clearly covered by the settlement agreement. Id. at 10-14.

C. The Related Case Of Bethea v. Amtrak Police Department

On July 11, 2001, Loretta Bethea filed an individual employment-discrimination lawsuit against the Amtrak Police Department in this court. See Compl., Bethea v. Amtrak Police Department , No. 01-cv-01513, ECF No. 1. Ms. Bethea alleged that she had suffered discrimination on the basis of her race and gender in connection with promotions and discipline. See generally id. Amtrak answered the complaint on September 6, 2001. See Answer, Bethea v. Amtrak Police Department , No. 01-cv-01513, ECF No. 5. On July 11, 2011, the parties requested a continuance of the initial scheduling conference in view of a request to consolidate Bethea with Campbell for pretrial purposes, see Joint Mot. to Continue, Bethea v. Amtrak Police Department , No. 01-cv-01513, ECF No. 11, and on May 2, 2003, the cases were consolidated for pretrial purposes, see Order, ECF No. 139.

D. The Parties Proceed To Class-Certification Discovery

Meanwhile, discovery was well under way in Campbell . Immediately after denying Amtrak's 2002 motion to dismiss, the Court entered an Order directing the parties to propose "an appropriate schedule for the completion of discovery in this matter." Order, ECF No. 132 at 1. After receiving the parties' proposal, the Court entered a Scheduling Order on November 7, 2002. See Sched. Order, ECF No. 135. The Scheduling Order provided that class-certification discovery would be completed by November 5, 2003, with expert-discovery regarding class certification to be completed by February 5, 2004. See id. at 1-2. The parties had also requested that the Court set a schedule for summary-judgment briefing. See Joint Status Report, ECF No. 133. The Court directed that both the class-certification and summary-judgment motions be filed by April 5, 2004, with the motions to be ripe by July 6, 2004. See Scheduling Order, ECF No. 135 at 3.

This schedule was extended at the parties' request on many occasions. See Am. Sched. Order, ECF No. 155; Minute Order of March 26, 2004; Am. Sched. Order, ECF No. 186; Minute Order of Sept. 14, 2004; Am. Sched. Order, ECF No. 205; Minute Order of Jan. 14, 2005. The parties requested additional continuances to work through discovery disputes and to create a joint database of employment-related data. See Minute Order of Sept. 9, 2005; Pls.' Mot. for Sanctions, ECF No. 231; Minute Order of Nov. 8, 2006.

*289On December 30, 2010, the Court entered a Revised Scheduling Order providing that the motions for class certification and summary judgment would be fully briefed by December 23, 2011. Sched. Order, ECF No. 280. That schedule was again modified due to the Supreme Court's grant of certiorari in Wal-Mart Stores, Inc. v. Dukes , 564 U.S. 338, 354, 131 S.Ct. 2541, 180 L.Ed.2d 374 (2011). Minute Order of May 10, 2011.

E. The Parties Brief Their Motions For Class Certification And Summary Judgment

On February 21, 2012, plaintiffs filed their motion for class certification. See Pls.' Mot. to Certify Class, ECF No. 303. Amtrak filed its opposition on June 26, 2012, along with its motion for partial summary judgment. See Def.'s Opp. to Mot. to Certify Class, ECF No. 320; Def.'s Mot. for Summ. J., ECF No. 328. On the same day, Amtrak filed its motions to exclude the report and testimony of several of plaintiffs' experts. See Def.'s Mot. to Exclude Finkelman, ECF No. 319; Def.'s Mot. to Exclude Roth, ECF No. 329; Def.'s Mot. to Exclude Bradley and Fox, ECF No. 331. Amtrak further moved to partially strike the declarations of certain putative class members offered in support of plaintiffs' motion for class certification. See Def.'s Mot. to Strike, ECF No. 330. These motions were all ripe by January 4, 2013.

The parties had agreed to engage in private mediation for a period of ninety days following the exchange of expert reports, class-certification briefing, and summary-judgment briefing. See Joint Proposed Sched. Order, ECF No. 310 at 2. Accordingly, the Court held in abeyance the parties' motions pending the conclusion of mediation. See Minute Order of Sept. 27, 2012. On March 4, 2013, the parties filed a joint status report indicating that mediation had been unsuccessful. See Joint Status Report, ECF No. 363. The Court subsequently requested that the parties file supplemental briefing discussing any new legal authority regarding class certification. See Minute Order of December 2, 2013. The parties submitted their supplemental briefing in early 2014. See Pls.' Supp. Class Cert. Mem., ECF No. 370; Def.'s Resp. to Pls.' Supp. Class Cert. Mem., ECF No. 371.

II. BACKGROUND

A. Amtrak's Structure

Amtrak provides passenger rail service through forty-six states and the District of Columbia. See Report of Drs. Edwin L. Bradley and Liesl M. Fox ("Bradley/Fox Rep."), ECF No. 304-1 at 3. In the period after its inception in 1971, Amtrak was "basically a centrally managed corporation in D.C." Dep. of Gilbert Mallery ("Mallery Dep."), ECF No. 323-7 at 4. In 1994, Amtrak began creating "strategic business units" or "SBUs" with the goal of organizing the business "around the services that existed" as opposed to around a corporate headquarters. Dep. of Lee W. Bullock ("Bullock Dep."), ECF No. 323-1 at 6; see also Mallery Dep., ECF No. 323-7 at 4 (explaining that the SBUs were created "to decentralize decision making" and "to move decision making in the corporation closer to the customers"). While decisions relating to Amtrak's "financial targets" and "ultimate strategy" for the collective-bargaining process were still centered in Amtrak's corporate headquarters in the District of Columbia, other decisions, like those related to budgets and marketing, were delegated to the SBUs. Bullock Dep., ECF No. 323-1 at 9. For example, although the SBUs followed the "broad" human-resources policies set at the corporate level, each SBU had its own human-resource director and decisions with respect to "hiring and firing" employees were made at the SBU level. Mallery Dep., ECF

*290No. 323-7 at 4-5. Thus, while Amtrak's corporate headquarters endeavored to ensure that any "federal and company-wide mandates were complied with," "the day-to-day decisions were delegated to the HR professionals in the business units" who "basically enforced, monitored, controlled to make sure those corporate policies were followed." Id. at 5.

The SBUs were disbanded in 2002, and Amtrak returned to a more traditional structure under which it was organized by functional department at the corporate level and by operating division at the field level. Decl. of Patricia Kerins ("Kerins Decl."), ECF No. 328-7 ¶ 28; Dep. of Edward Valentine Walker, III ("Walker Dep."), ECF No. 309-4 at 3. Although Amtrak has eighteen departments, plaintiffs' expert Thomas Roth postulates that approximately ninety-seven percent of Amtrak's unionized workforce resides in one of five departments. Decl. and Expert Rep. of Thomas R. Roth ("Roth Rep."), ECF No. 304-2 ¶¶ 8, 15. According to Mr. Roth, these five departments coincide with five functional categories-or "craft" groups-that are useful "for analytical purposes": operating, equipment maintenance, maintenance of way, clerical/on-board services, and security. Id. ¶ 9. Mr. Roth opines that employees in these five craft groups have "a fundamentally shared function" and that the jobs in each of the categories "share common work sites and supervision." Id. ¶ 21. In addition, Mr. Roth asserts that, even though Amtrak's employees are represented by seventeen different unions, employees within each craft group tend to negotiate common terms and common work conditions in their collective-bargaining agreements. Id. ¶ 25. Finally, Mr. Roth notes that the "rules governing discipline and grievances are common to a substantial degree within each functional employee group." Id. ¶ 31.

B. Hiring And Promotions

Amtrak has a corporate hiring, promotion, and transfer policy that was created "to provide guidelines to Amtrak supervision on how jobs are filled through employment, promotion, and transfer of employees." See May 1, 1994 Amtrak Employment/Promotion/Transfer Policy ("1994 Amtrak Hiring Policy"), ECF No. 307-2 at 3. This policy has been the same since January 1, 1989. See, e.g. , Jan. 1, 1989 Amtrak Employment/Promotion/Transfer Policy, ECF No. 307-1 at 3; Sept. 2000 Amtrak Employment, Promotion and Transfer Policy, ECF No. 307-3 at 3.

Pursuant to that policy, positions covered by collective-bargaining agreements ("agreement-covered positions") must be "advertised for bid in accordance with the applicable labor agreement." 1994 Amtrak Hiring Policy, ECF No. 307-2 at 12. Local employees who are members of the union that covers the vacant position are eligible to bid on the position. Decl. of Sarah Ray ("Ray Decl."), ECF No. 322-5 ¶ 4. Generally, the most senior employee who places a bid and otherwise meets the qualifications is placed in the position. Id. ¶ 5. If no local employee bids on the position, then human resources will determine if a member of the relevant union in a different geographic location wishes to transfer to take the position. Id. ¶ 7. That employee would also be required to meet any qualification requirements before being awarded the position permanently. Id.

When positions are not filled after this internal bidding process, certain steps must be taken to fill a vacancy. See 1994 Amtrak Hiring Policy, ECF No. 307-2 at 13. The hiring process begins with the job requisition form, which provides detailed information regarding the duties and responsibilities associated with the position, the requisite qualifications and experience *291required, and any preferred qualifications and experience. Id. at 8; Ray Decl., ECF No. 322-5 ¶ 10. Generally, a hiring manager will determine the hiring criteria for an open position by reviewing a job description or prior requisition forms. Ray Decl., ECF No. 322-5 ¶ 11; Decl. of Suzanne Allan ("Allan Decl."), ECF No. 321-3 ¶ 5. The process of preparing and approving a job requisition form varies by department. Decl. of Barbara Wu ("Wu Decl."), ECF No. 322-8 ¶ 4.

The selection criteria for each position vary significantly and depend on the job description and requirements described in the job requisition form. Wu Decl., ECF No. 322-8 ¶ 9; Ray Decl., ECF No. 322-5 ¶ 12. Local applicants are preferred for certain positions, especially those for on-board crew. Wu Decl., ECF No. 322-8 ¶ 9; Ray Decl., ECF No. 322-5 ¶ 22. Someone in human resources is responsible for screening all the applications for a particular job to determine which applicants match the minimum requirements, have similar experience to that of the position at issue, and have a stable employment history. Wu Decl., ECF No. 322-8 ¶ 9. A hiring manager may ask that the human-resources recruiter provide the applications for all candidates that meet the minimum requirements of the position or may request applications from only the most qualified applicants. Ray Decl., ECF No. 322-5 ¶ 25.

Almost all agreement-covered positions require that the applicant pass a test or set of tests prior to becoming eligible for interviews. Wu Decl., ECF No. 322-8 ¶ 10. Applicants who meet the minimum requirements for a vacancy are invited to take the test. Kerins Decl., ECF No. 328-7 ¶ 8. The tests administered vary depending on the position and the union involved, and they have changed over time. Wu Decl., ECF No. 322-8 ¶ 10. Passing a test does not necessarily mean that the applicant will be interviewed for the position; rather, only the most qualified applicants are generally interviewed for each position. Id. ¶ 13. Typically, at least three to five applicants are selected to be interviewed for each vacancy. Ray Decl., ECF No. 322-5 ¶ 29; Allan Decl., ECF No. 321-3 ¶ 10.

The hiring manager, in consultation with others, develops a set of interview questions. Wu Decl., ECF No. 322-8 ¶ 15. The types of questions asked during an interview depend on the position at issue, any unique requirements relating to the particular opening, and the preferences of the hiring manager. Kerins Decl., ECF No. 328-7 ¶ 13. Each applicant who interviews for a particular position is asked the same set of questions. Wu Decl., ECF No. 322-8 ¶ 15. Interviews are conducted by panels of managers and, in some cases, a union representative. Id. ¶ 17. At some point before they start conducting interviews, most managers participate in a behavioral-based interview training led by a member of the human-resources department. Kerins Decl., ECF No. 328-7 ¶ 12; Decl. of Karen Broadwater, ECF No. 321-6 ¶ 21.

At the conclusion of the interview, the panel members provide each other with feedback on the candidate. Allan Decl., ECF No. 321-3 ¶ 17. For some positions, interviewers use a rating form to score the applicant's responses. Wu Decl., ECF No. 322-8 ¶ 20. If the scores of each panel member vary, a consensus form may be used to reach a final score. Id. Unless the applicant has a very low score in a key competency, the applicant with the highest total score is usually recommended for the position. Id. In other cases, panel members may simply take notes during the interview to record their opinions about applicants' responses. Kerins Decl., ECF No. 328-7 ¶ 15. The process of assessing candidates is "not a cut-and-dried type process," but rather involves a "discussion ...

*292among the panel members about the strengths and weaknesses of a candidate." Dep. of Sheila Davidson, ECF No. 306-2 at 16. Candidates are evaluated based on their experience, interview performance, and professionalism. Kerins Decl., ECF No. 328-7 ¶ 15.

While each member of the panel shares his or her thoughts about the qualifications of the candidates, the ultimate decision of which candidate to recommend for the vacancy lies with the hiring manager. Kerins Decl., ECF No. 328-7 ¶ 16. The hiring manager's selection may be reviewed by his or her supervisor, and the decision is ultimately approved by the human-resources department at Amtrak's corporate headquarters. Id. ¶ 18; Walker Dep., ECF No. 309-4 at 11-12.

Dr. Bradley and Dr. Fox, plaintiffs' statistical experts who analyzed Amtrak's hiring and promotion data, found that African-American individuals were hired and promoted for vacant positions at rates lower than their non-African-American counterparts. Bradley/Fox Rep., ECF No. 304-1 at 4. Specifically, Dr. Bradley and Dr. Fox concluded that 3,053 fewer African-American individuals were hired or promoted than would be expected from the pool of applicants, after removing those candidates in the pool who were not minimally-qualified for the position. Id. at 15-16. Dr. Bradley and Dr. Fox did not, however, consider other criteria-such as seniority, work experience, education, or whether the applicant had previously worked at Amtrak-that may have affected hiring or promotion decisions. Dep. of Edwin Bradley ("Bradley Dep."), ECF No. 331-3 at 23-24, 28, 56-57.

C. Discipline

The collective-bargaining agreements usually contain rules governing the discipline process. Decl. of Charles E. Woodcock, III ("Woodcock Decl."), ECF No. 322-7 ¶ 23. The discipline process at Amtrak generally progresses as follows: (1) verbal warning; (2) written warning; (3) disciplinary hearing if a formal charge is filed; (4) a second disciplinary hearing if a formal charge is filed; and (5) a third disciplinary hearing if a formal charge is filed, which may in turn lead to termination. Id. ¶ 20. Discipline decisions are generally made by and subject to the discretion of a local manager. Id. ¶ 22. The final decision to terminate an individual currently rests with the vice-president of human resources. Walker Dep., ECF No. 309-4 at 4-5.

This basic disciplinary process is similar for employees across all labor unions, though there are some limited differences. See Dep. of LaVerne Miller, ECF No. 308-6 at 34-35 (Amtrak corporate designee testifying that the claims and grievance procedures across craft groups are "equal across the board"); Woodcock Decl., ECF No. 322-7 ¶ 23; Roth Rep., ECF No. 304-2 ¶ 31. For example, each collective-bargaining agreement has "just cause" type provisions that afford employees the right to file an appeal of any disciplinary charges. Woodcock Decl., ECF No. 322-7 ¶ 23; see also Roth Rep., ECF No. 304-2 ¶ 32 (explaining that the language of the grievance procedures vary between collective-bargaining agreements but that they all "embody the principles of just cause, fair and impartial investigation, timeliness and [ ] other due process elements").

Despite these broad similarities, rules governing employee conduct may vary by position. Woodcock Decl., ECF No. 322-7 ¶ 21. For example, passenger engineers are subject to certain federal regulations and operating rules that other employees are not. Id. ¶ 21. Likewise, there may be different expectations for ticket clerks, who deal with customers on a daily basis, than for other employees whose jobs do *293not require interaction with the public. Id. ¶ 21.

Dr. Bradley and Dr. Fox compared the rates of disciplinary charges between African-American and non-African-American unionized employees at Amtrak. Bradley/Fox Rep., ECF No. 304-1 at 16. They found that, of the 24,136 disciplinary charges issued to Amtrak employees during the analysis time period, 10,651 charges were brought against African-American employees, even though one would have expected only 8,924 charges to be brought against African-American employees during that same period. Id. Notably, Dr. Bradley and Dr. Fox did not make these comparisons among employees that were similarly situated-for example, Dr. Bradley explained that his analysis did not consider the specific position or union to which the employee belonged, an employee's previous disciplinary history, the severity of the offense and discipline issued, or the employee's tenure at Amtrak. See Bradley Dep., ECF No. 331-3 at 65-67.

D. Work Environment

Amtrak, like many employers of its size, has corporate policies prohibiting discrimination, harassment, and retaliation. See Def.'s Opp. to Mot. to Certify Class, ECF No. 320 at 16-19; Dep. of Karen Broadwater Ex. 1, ECF No. 322-10 at 14-17 (Sept. 20, 2011 EEO and Affirmative Action Policy); id. Ex. 2, ECF No. 322-10 at 18-22 (Anti-Discrimination and Anti-Harassment Policy). In addition, as a result of the settlements entered in McLaurin v. Amtrak and Thornton v. Amtrak , Amtrak established a Dispute Resolution Office ("DRO") in 1999, which was located within the Business Diversity Department. Decl. of Dawn Marcelle ("Marcelle Decl."), ECF No. 322-2 ¶¶ 2, 9. The function of the DRO was to investigate internal complaints of harassment or discrimination raised by agreement-covered employees. Id. ¶ 10. Employees could initiate complaints internally in a variety of ways: they could raise complaints with supervisors, report complaints directly to their local DRO office, or call the DRO hotline. Id. ¶ 14.

Wanda Hightower, the Vice President of the Business Diversity Department between April 1999 and February 2001, testified that she and her staff attempted to aggressively investigate race discrimination complaints during her tenure at Amtrak. See Dep. of Wanda Hightower ("Hightower Dep."), ECF No. 309-9 at 7-8. Ms. Hightower testified that these efforts were met with resistance by both lower-level employees and upper management at Amtrak. See id. at 14-18, 22-23, 29. She also stated that racial discrimination "was bad across the system" at Amtrak, particularly among the "rank and file." Id. at 30. This testimony is supported by the declarations of named plaintiffs and putative class members, some of whom point to individual instances of racism and others of whom point to a more pervasive culture of racism during their tenure at Amtrak. See Pls.' Mot. for Class Cert. Ex. 8, ECF No. 304-8. These declarations detail instances of overt and obvious racism (e.g., use of racial epithets, hanging black dolls or monkeys from nooses in employee common areas, racially-charged physical threats), in addition to allegations of more subtle racism (e.g., assigning African-American employees more menial job assignments). See id.

In June 2007, Amtrak dissolved the Business Diversity Department, and the DRO was merged into the Human Resources Department. Marcelle Decl., ECF No. 322-2 ¶ 24. After the DRO moved to the Human Resources Department, it continued to receive and investigate internal complaints through May 2011, at which time the DRO was dissolved. Id. ¶ 25. All complaints related to discrimination are now addressed by the EEO Compliance *294Unit, which is part of Amtrak's Legal Department.²

III. EXCLUSION OF EXPERT TESTIMONY

Amtrak moves to exclude the testimony and reports of various experts proffered by plaintiffs in support of their motion for class certification. See Def.'s Mot. to Exclude Finkelman, ECF No. 319; Def.'s Mot. to Exclude Roth, ECF No. 329; Def.'s Mot. to Exclude Bradley and Fox, ECF No. 331. Amtrak contends that these experts must be excluded under Federal Rule of Evidence 702 and Daubert v. Merrell Dow Pharmaceuticals, Inc. , 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993). Plaintiffs argue that Daubert 's admissibility considerations are not relevant at the class-certification stage and, in any event, that their evidence is admissible.

A. The Court Must Conduct A Full Daubert Inquiry Before Relying On Expert Testimony At The Class-Certification Stage

The issue of how to evaluate expert testimony at the class-certification stage "ha[s] beguiled the federal courts." Newberg on Class Actions § 7:24 (5th ed. 2014). The Supreme Court has strongly hinted that district courts should apply the same standard at the class-certification stage that they would apply to expert testimony offered at a later stage of proceedings. See Wal-Mart Stores, Inc. v. Dukes , 564 U.S. 338, 354, 131 S.Ct. 2541, 180 L.Ed.2d 374 (2011) (casting "doubt" on the conclusion "that Daubert did not apply to expert testimony at the certification stage of class-action proceedings"). Indeed, in 2013, the Supreme Court granted certiorari to resolve the issue but was unable to do so because "the question was not properly posed." See Newberg on Class Actions § 7:24 (citing Comcast Corp. v. Behrend , 569 U.S. 27, 133 S.Ct. 1426, 185 L.Ed.2d 515 (2013) ). The question is difficult "for the simple reason that certification is generally not the time to decide the merits of the case, yet expert witness testimony relevant to the merits often is proffered as also relevant to a prong of the certification inquiry." Newberg on Class Actions § 7:24.

The Court of Appeals for the District of Columbia Circuit ("D.C. Circuit") has not yet weighed in on whether a full analysis under Daubert is required at the class-certification stage. See In re Rail Freight Fuel Surcharge Antitrust Litig. , No. 07-0489, 2016 WL 2962186, at *2 (D.D.C. May 20, 2016) ; Moore v. Napolitano , 926 F.Supp.2d 8, 16, n.2 (D.D.C. 2013) ; Kottaras v. Whole Foods Mkt., Inc. , 281 F.R.D. 16, 26 (D.D.C. 2012). Most circuit courts that have addressed the issue have found that, where an expert's testimony is critical to class certification, "a district court must conclusively rule on any challenge to the expert's qualifications or submissions prior to ruling on a class certification motion"-i.e., "the district court must perform a full Daubert analysis before certifying the class." Am. Honda Motor Co. v. Allen , 600 F.3d 813, 815-16 (7th Cir. 2010) ; see also, e.g. , Ellis v. Costco Wholesale Corp. , 657 F.3d 970, 982 (9th Cir. 2011) (district court "correctly applied the evidentiary standard set forth in Daubert " at the class-certification stage); Sher v. Raytheon Co. , 419 Fed.Appx. 887, 890-91 (11th Cir. 2011) ("Here the district court refused to conduct a Daubert -like critique of the proffered experts' qualifications. This was error.");

*295In re Carpenter Co. , No. 14-0302, 2014 WL 12809636, at *3-4, 2014 U.S. App. LEXIS 24707, at *10-11 (6th Cir. Sep. 29, 2014) (district court did not abuse its discretion by analyzing expert testimony offered in support of class certification under Daubert ); In re Blood Reagents Antitrust Litig. , 783 F.3d 183, 187 (3d Cir. 2015) ("We join certain of our sister courts to hold that a plaintiff cannot rely on challenged expert testimony, when critical to class certification, to demonstrate conformity with Rule 23 unless the plaintiff also demonstrates, and the trial court finds, that the expert testimony satisfies the standard set out in Daubert ."); but see In re Zurn Pex Plumbing Prods. Liab. Litig. , 644 F.3d 604, 611-14 (8th Cir. 2011) (approving use of a "focused Daubert analysis" instead of a "full and conclusive Daubert inquiry").

The courts that have required a full Daubert inquiry generally focus on the "rigorous analysis" that a district court must apply to a plaintiff's request for class certification-a standard that, after Comcast , clearly applies to expert testimony that is proffered in support a request for certification. See Comcast Corp. v. Behrend , 569 U.S. 27, 34-35, 133 S.Ct. 1426, 185 L.Ed.2d 515 (2013). This means that the district court must discern whether a plaintiff has proven compliance with Rule 23(a)"in fact " and whether the plaintiff has " 'satisf[ied] through evidentiary proof at least one of the provisions of Rule 23(b).' " In re Blood Reagents Antitrust Litig. , 783 F.3d at 187. Under this approach, "[e]xpert testimony that is insufficiently reliable to satisfy the Daubert standard cannot 'prove' that the Rule 23(a) prerequisites have been met 'in fact,' nor can it establish 'through evidentiary proof' that Rule 23(b) is satisfied." Id. ; see also, e.g. , Messner v. Northshore Univ. HealthSystem , 669 F.3d 802, 812 (7th Cir. 2012) ("Failure to conduct [a Daubert ] analysis when necessary ... would mean that the unreliable testimony remains in the record, a result that could easily lead to reversal on appeal.").

The Eighth Circuit-the only Circuit to have reached a contradictory decision after Dukes -sanctioned a " 'tailored' Daubert analysis" that "examined the reliability of the expert opinions in light of the available evidence and the purpose for which they were offered." In re Zurn Pex Plumbing Prods. Liab. Litig. , 644 F.3d 604, 612 (8th Cir. 2011) ; see also id. at 611 (declining to "adopt a new rule, requiring a district court to determine conclusively at an early stage, not just whether or not expert evidence is sufficient to support class certification under Rule 23, but also whether that evidence will ultimately be admissible for trial"). This holding emphasized the "inherently preliminary nature of pretrial evidentiary and class certification rulings," and noted that the "main purpose of Daubert "-"to protect juries from being swayed by dubious scientific testimony"-does not arise in motions for class certification "where the judge is the decision maker." Id. at 613.

The Court is persuaded that it must conduct a full Daubert inquiry at the class-certification stage. Concerns regarding the tentativeness of class-certification rulings have been undermined significantly by the 2003 amendment to Rule 23, which removed language permitting a conditional class-certification ruling. See In re Zurn Pex Plumbing Prods. Liab. Litig. , 644 F.3d at 630 (Gruender, J., dissenting) (arguing that "the 2003 amendments to Rule 23 removed the provision that class certification 'may be conditional' " and that failing to conduct a full Daubert inquiry regarding evidence crucial to a certification decision would mean that "the case will proceed beyond class certification on the basis of inadmissible , unreliable expert testimony"). The fact that a class-certification ruling may be revisited, *296Fed. R. Civ. P. 23(c)(1)(C), or that merits-related discovery may lead to additional evidence that supports an expert's conclusions, does not warrant applying a relaxed standard to an expert's opinions at the certification stage. Moreover, after Dukes , "[t]he Court must consider merits questions when those questions overlap with Rule 23's requirements." Coleman through Bunn v. Dist. of Columbia , 306 F.R.D. 68, 77 (D.D.C. 2015) ; cf. Amgen Inc. v. Connecticut Ret. Plans & Tr. Funds , 568 U.S. 455, 466, 133 S.Ct. 1184, 185 L.Ed.2d 308 (2013) ("Merits questions may be considered to the extent-but only to the extent-that they are relevant to determining whether the Rule 23 prerequisites for class certification are satisfied."). Accordingly, the concern that conducting a full Daubert hearing would inappropriately prejudge a merits issue is less persuasive; if that merits issue overlaps with Rule 23, the Court must prejudge it to the extent necessary to assess plaintiffs' compliance with Rule 23.

In short, the Court agrees with the heavy weight of authority that, when a party moves to exclude expert testimony proffered in support of a motion for class certification, the district court must perform a full Daubert analysis before certifying a class. See In re Rail Freight Fuel Surcharge Antitrust Litig. , No. 07-0489, 2016 WL 2962186, at *2 (D.D.C. May 20, 2016) (addressing the "reliability of the experts' methodology under Daubert and Rule 702" at the class-certification stage); Kottaras v. Whole Foods Mkt., Inc. , 281 F.R.D. 16, 26 (D.D.C. 2012) ("the Court agrees with other courts that the Rule calls for careful and searching analysis of all evidence with respect to whether Rule 23's certification requirements have been met, including expert opinions").

B. Legal Standard For Admissibility Of Expert Testimony

A district court has " 'broad discretion in determining whether to admit or exclude expert testimony.' " United States ex rel. Miller v. Bill Harbert Int'l Constr., Inc. , 608 F.3d 871, 895 (D.C. Cir. 2010) (citation omitted). The exercise of that discretion is governed by Federal Rule of Evidence 702, which provides that:

A witness who is qualified as an expert by knowledge, skill, experience, training, or education may testify in the form of an opinion or otherwise if:

(a) the expert's scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue;

(b) the testimony is based on sufficient facts or data;

(c) the testimony is the product of reliable principles and methods; and

(d) the expert has reliably applied the principles and methods to the facts of the case.

Rule 702 requires the district court to act as a "gatekeeper" for expert testimony by ensuring that "any and all scientific testimony or evidence admitted is not only relevant, but reliable." Daubert , 509 U.S. at 589, 113 S.Ct. 2786 ; see also Ambrosini v. Labarraque , 101 F.3d 129, 133 (D.C. Cir. 1996) ("[t]he Daubert standard involves a two-prong analysis that centers on evidentiary reliability and relevancy"). In Kumho Tire Co. v. Carmichael , 526 U.S. 137, 141, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999), the Supreme Court made clear that this gatekeeping obligation applies to all expert testimony, not just scientifically-based testimony.

Under Rule 702, expert testimony is reliable if (1) it is based on sufficient facts or data; (2) it is the product of reliable principles and methods; and (3) the *297expert has applied the principles and methods reliably to the facts of the case. See, e.g. , Heller v. District of Columbia , 801 F.3d 264, 271 (D.C. Cir. 2015) ("[C]ourts are obligated to 'determine whether [expert] testimony has a reliable basis in the knowledge and experience of [the relevant] discipline.' ") (quoting Kumho Tire , 526 U.S. at 149, 119 S.Ct. 1167 ); Robinson v. Dist. of Columbia , 75 F.Supp.3d 190, 199 (D.D.C. 2014) (" '[t]he trial judge ... must find that [the proffered testimony] is properly grounded, well-reasoned and not speculative before it can be admitted' ") (quoting Fed. R. Evid. 702 advisory committee notes). In determining reliability, the district court must "focus solely on principles and methodology, not on the conclusions that they generate." Ambrosini , 101 F.3d at 133. The trial judge has "considerable leeway in deciding in a particular case how to go about determining whether particular expert testimony is reliable." Kumho Tire , 526 U.S. at 152, 119 S.Ct. 1167 ; see also Estate of Gaither ex rel. Gaither v. Dist. of Columbia , 831 F.Supp.2d 56, 62 (D.D.C. 2011) (" 'Exactly how reliability is evaluated may vary from case to case, but what remains constant is the requirement that the trial judge evaluate the reliability of the testimony before allowing its admission at trial.' ") (citation omitted).

"The second Daubert prong relates to relevance and is fairly straightforward." Rothe Dev., Inc. v. Dept. of Defense , 107 F.Supp.3d 183, 197 (D.D.C. 2015). The district court "must determine whether the proffered testimony is sufficiently tied to the facts of the case and whether it will aid the factfinder in resolving a factual dispute. Fed. Trade Comm'n v. Whole Foods Market, Inc. , No. 07-1021, 2007 WL 7632283, at *1 (D.D.C. July 17, 2007). "The Daubert Court described this consideration as one of 'fit.' " Ambrosini , 101 F.3d at 134. Although the district court assumes only a "limited" gate-keeping role under these standards, and " '[r]ejection of an expert's testimony is the exception rather than the rule,' " see Paige Int'l, Inc. v. XL Specialty Ins. Co. , No. 14-1244, 2016 WL 3024008, at *3 (D.D.C. May 25, 2016) (citation omitted), the "decision to receive expert testimony" cannot be " 'simply tossed off to the jury under a 'let it all in' philosophy,' " see Joy v. Bell Helicopter Textron, Inc. , 999 F.2d 549, 569 (D.C. Cir. 1993) (citation omitted). As such, "[t]he issue for the Court to determine is whether ... [the expert's] assumptions amount to 'rampant speculation' and should be excluded, or whether [the] assumptions merely represent a weak factual basis for [the expert's] testimony that is appropriately challenged on cross examination." Boyar v. Korean Air Lines Co. , 954 F.Supp. 4, 7 (D.D.C. 1996).

C. Amtrak's Motion To Exclude Dr. Finkelman's Testimony And Report Is Granted

1. Dr. Finkelman's Qualifications And Opinions

Dr. Jay Finkelman is an industrial-organizational psychologist and the Vice-President of Academic Affairs and Chief Academic Officer of the Chicago School of Professional Psychology. Expert Rebuttal Report of Jay Finkelman, PhD, ABPP, CPE ("Finkelman Rebuttal Rep."), ECF No. 344-2 at 2. He holds a Ph.D. in Industrial-Organizational Psychology from New York University and an M.B.A. in Industrial Psychology from the Bernard M. Baruch School of Business. Expert Rep. of Jay Finkelman, PhD, ABPP, CPE ("Finkelman Rep."), ECF No. 304-3 at 3. Dr. Finkelman "specializes" in a variety of topics, including human resources, staffing industry management practices, employment discrimination, adverse impact, performance appraisal, and psychometrics. Id. at 5. He has "had hundreds of retentions and *298depositions" and has "testified at trial over 46 times." Id. at 3.

Dr. Finkelman was retained by plaintiffs to "review the hiring, promotional, and discipline policies of Amtrak" and "determine whether or not they were consistent with generally accepted Human Resource Management practices and the principles of Industrial-Organizational Psychology." Finkelman Rep., ECF No. 304-3 at 17-18. In preparing his report, Dr. Finkelman relied on plaintiffs' third amended complaint, Amtrak's answer to that complaint, the depositions and related exhibits of three of Amtrak's corporate representatives, the deposition and related exhibits of one fact witness, the expert report of Dr. Bradley and Dr. Fox, a document titled "Selection Roulette," and the declarations of class members. See Finkelman Report at 26; Dep. of Jay M. Finkelman ("Finkelman Dep."), ECF No. 319-3 at 5.

Based on his review of these selected materials, Dr. Finkelman provided an expert report, the substance of which spans approximately eight pages. In his report, in a section titled "Opinions," Dr. Finkelman first sets forth background principles undergirding "good" human-resource management policies and practices and states that Amtrak "did not appear to have adequate mechanisms in place" to accomplish certain objectives of human-resource management. Finkelman Rep., ECF No. 304-3 at 19-20. Notably, he does not cite any studies, data, articles, or other academic sources supporting any of his observations.

Dr. Finkelman next makes "[a] few specific observations" with respect to this case. Id. at 18-19. Those observations consist of twenty bullet points that point out various problematic human-resources practices purportedly found at Amtrak. Id. at 19-24. Nine of those bullets are summaries of testimony of Amtrak managers provided to Dr. Finkelman by plaintiffs' counsel in a document titled "Selection Roulette," coupled with Dr. Finkelman's observations about the hiring practices described in those summaries. Compare id. at 20-23, with Finkelman Dep. Ex. 7, ECF No. 319-3 at 130-136. Based on his review of the summaries, Dr. Finkelman concludes that Amtrak's employment policies and practices failed to accomplish the "dual" goals of human-resource management: to protect employees from discrimination, harassment, and retaliation while "also protecting the organization[ ] from liability associated with improper policies and practices." Finkelman Rep., ECF No. 304-3 at 18. According to Dr. Finkelman, this failure is attributable to the fact (1) that Amtrak has "few if any controls against intentional or inadvertent bias" and (2) that Amtrak's hiring, promotional, and discipline policies are "not consistent with generally accepted Human Resource Management practices nor with the professional requirements of Industrial-Organizational Psychology." Id. at 24; see also Finkelman Dep., ECF No. 319-3 at 4-5 (opining that, although Amtrak had overarching policies in place governing hiring, promotion, and employee discipline, individual managers departed from those policies in a manner that "allowed for subjectivity and the potential for bias or discrimination").

2. Dr. Finkelman's Opinions Are Unreliable

Amtrak argues that Dr. Finkelman's report fails Daubert 's reliability prong because, among other reasons, Dr. Finkelman did not consider sufficient facts in forming his opinion. See Def.'s Mem. in Supp. Mot. to Exclude Finkelman, ECF No. 319-1 at 19-25. Amtrak asserts that Dr. Finkelman's opinions are supported only by "cherry-picked" documents selected by plaintiffs' counsel, and that Dr. Finkelman failed to request, much less review, a host of other evidence "pertinent to the question he purportedly sought to answer-whether *299Amtrak's policies, practices, and procedures are consistent with generally accepted human-resources practices and the general principles of industrial organizational psychology." Id.

Federal Rule of Evidence 702 requires expert testimony to be "based on sufficient facts or data" to be reliable. United States ex rel. Miller v. Bill Harbert Int'l Const., Inc. , 608 F.3d 871, 894 (D.C. Cir. 2010) (quoting Fed. R. Evid. 702 ). Although an expert "need not consider every possible factor to render a 'reliable' opinion, the expert still must consider enough factors to make his or her opinion sufficiently reliable in the eyes of the court." MicroStrategy Inc. v. Bus. Objects, S.A. , 429 F.3d 1344, 1355 (Fed. Cir. 2005) ; Estate of Gaither ex rel. Gaither v. Dist. of Columbia , 831 F.Supp.2d 56, 66 (D.D.C. 2011) (expert must provide "meaningful measure of detail " as to the expert's "experience with and knowledge of" the facts underlying his opinions).

After reviewing Dr. Finkelman's report and deposition testimony, the Court finds that Dr. Finkelman's expert opinion relies on insufficient facts and data and therefore lacks the reliability required under Rule 702. For example, although Dr. Finkelman purports to opine about the human-resource management practices at Amtrak, he testified that he did not review information critical to those opinions, including the depositions of any human-resource managers, see Finkelman Dep., ECF No. 319-3 at 5; documents related to any job-selection decisions by any Amtrak manager, see id. at 23; personnel files or documents related to any discipline decision made by Amtrak, see id. ; or Amtrak's discrimination complaint procedures or anti-discrimination training, see id. at 45.

Likewise, although he agreed that the collective-bargaining agreements applicable to plaintiffs "would have [an] impact" on Amtrak's hiring, promotion, and disciplinary policies, see Finkelman Dep., ECF No. 319-3 at 25, he did not review those agreements prior to forming his opinions, see id. at 22-23. Dr. Finkelman also testified that, although he "assume[s]" that the consent decrees imposed in previous employment-discrimination litigation involving Amtrak had an impact "on the hiring, promotional, or discipline policies at Amtrak," he "didn't read the consent decrees" and therefore did not "know the exact changes that may have been made" by Amtrak in response to the decrees. Id. at 25. Rather, Dr. Finkelman testified that he only considered limited "facts" in forming his opinion that Amtrak's practices and policies were inconsistent with the goals of good human-resource management practices:

Q. You say "Amtrak has not accomplished either objective, in my opinion." What's the basis for that statement?

A. Well, the basis is the allegations that have been leveled by both the plaintiffs and perhaps other class members. In this matter, my review of the statements by supervisors as to how it is that they engage in selection or promotion. And I suppose also by the $16 million that Amtrak had to pay in one of the earlier phases of litigation, which seemed to suggest that there is a problem and some wrongdoing.

Finkelman Dep., ECF No. 319-3 at 26.

Even in his consideration of these limited "facts," Dr. Finkelman did little to confirm their accuracy. For example, although he offers opinions on the selection process used by supervisors in hiring and promoting employees, Dr. Finkelman did not read the deposition testimony of any supervisors. See Finkelman Dep., ECF No. 319-2 at 31. Instead, Dr. Finkelman relied on a document titled "Selection Roulette" in which plaintiffs' counsel "summed up" the testimony of select supervisors. Id. at 5 *300("The information from managers came from the depositions that had taken place by managers. And I got that information through, I think, a document that's referred to as a "selection roulette" or something like that."); id. at 28 (affirming that he relied on the document titled "selection roulette" for management testimony). Dr. Finkelman conceded that he took no steps to verify the accuracy or the representativeness of the information in the "Selection Roulette" document. Id. at 28-29. Indeed, Dr. Finkelman acknowledged that, "to figure out what really took place" at Amtrak, he would likely need to "get into more detail." Id. at 29. For purposes of the report, Dr. Finkelman explained

From the descriptions that were given [in the "Selection Roulette" document], assuming that they are reasonably accurate , they are so far off in acceptable norm that I didn't need any more at this point. I will look more carefully, and if-if the roulette has misrepresented any of those issues or approaches, and if it didn't accurately characterize what managers said, you know, that would be a different story. But there are specific citations that are included, and yes, I will be looking at those.

Id. (emphasis added).

Dr. Finkelman simply "assum[ed]" that plaintiffs' counsel "accurately characterize[d]" the testimony of Amtrak managers regarding their hiring practices instead of independently reviewing that testimony himself-and, critically, proceeded to formulate the opinions set forth in his expert report based, in part, on that unverified testimony. Such blind reliance on "facts" provided by plaintiffs' counsel-combined with his failure to review other sources of information that he conceded could have affected Amtrak's hiring, promotion, and disciplinary practices-renders his expert report unreliable. See, e.g. , Parsi v. Daioleslam , 852 F.Supp.2d 82, 89 (D.D.C. 2012) (facts and data relied upon by expert were "patently insufficient" where expert "read only an apparently haphazard selection of defendant's sources"); Sommerfield v. City of Chicago , 254 F.R.D. 317, 321 (N.D. Ill. 2008) ("Acceptance of the notion that an expert can reasonably base his opinion on summaries of deposition testimony prepared by a party's lawyer would effectively eliminate Daubert 's insistence that an expert's opinion be grounded on reliable information."); Equal Emp't Opportunity Comm'n v. Rockwell Int'l Corp. , 60 F.Supp.2d 791, 795 (N.D. Ill. 1999) (expert report unreliable where expert's "sole source of information ... c[ame] from summaries prepared by one of the litigants" and expert failed to "review the entire depositions"); Equal Emp't Opportunity Comm'n v. Bloomberg L.P. , No. 07-8383, 2010 WL 3466370, at *14, 2010 U.S. Dist. LEXIS 92511, at *46 (S.D.N.Y. Aug. 31, 2010) (expert's reliance "solely on the information fed to him by [plaintiff] without independently verifying whether the information [wa]s representative undermine[d] the reliability of his analysis"); Makor Issues & Rights, Ltd. v. Tellabs, Inc. , No. 02-4356, 2010 WL 2607241, at *5, 2010 U.S. Dist. LEXIS 62114, at *14-15 (N.D. Ill. June 23, 2010) (excluding expert report related to defendant's financial forecasts in part because expert "did not read any testimony about how [defendant] prepared its forecast" and "did not even read the deposition transcript" of the head of defendant's forecasting division).

There are other problems with the reliability of Dr. Finkelman's proposed expert testimony as well. For example, Dr. Finkelman has not identified any particular principles or methodology he used in forming his opinions. Dr. Finkelman does not cite a single study, report, or other source for his opinions related to appropriate human-resources policies and practices. And, although he opines that Amtrak permitted an inappropriate degree of subjectivity in *301its hiring and promotion practices, Dr. Finkelman did not attempt to measure the degree of subjectivity at Amtrak. Finkelman Dep., ECF No. 319-3 at 27. Indeed, Dr. Finkelman did not conduct any independent research as part of his assessment of Amtrak's policies. Id. at 22. Rather, when asked how he prepared his report, Dr. Finkelman testified:

The process I use, which is the way I normally do it is, I go through all the documents initially to get a sense of it, and I tend to do that rapidly. And then I go back and start making determinations as to what fits into a report. And that's exactly the process I used. So I start by reading the complaint and the answer to the complaint, and then the depositions. And I start finding spots in the deposition or the declarations that seemed to be relevant to what I was asked to do, and then I just put that together as a report.

Finkelman Dep., ECF No. 319-3 at 24; see also id. at 27 (testifying that his report was "predominantly" based on his "review of the deposition testimony of managers as to how it is that they make decisions pertaining to hiring and promotion and to a lesser degree discipline"); id. at 28 (confirming that he did not read the deposition testimony of managers but instead relied on the "selection roulette" document provided by plaintiffs' counsel that summarized the testimony of selected managers).

In short, Dr. Finkelman appears to have uncritically relied on documents supplied to him by plaintiffs' counsel, cited to those pieces of evidence that supported his theories, and concluded that this selective evidence demonstrates that Amtrak's practices were inconsistent with generally-accepted human resource management practices. Finkelman Rep., ECF No. 304-3 at 25. To the extent that this may be considered a methodology at all, it does not meet the standards of reliability demanded by Rule 702 or Daubert . See, e.g. , Chesapeake Climate Action Network v. Exp.-Imp. Bank of the United States , 78 F.Supp.3d 208, 219 (D.D.C. 2015) (expert testimony inadmissible where expert failed to identify any "principles or methodology" used to arrive at his opinions, but rather "note[d] only that he reviewed certain documents and reached a series of conclusions"); Obrycka v. City of Chicago , 792 F.Supp.2d 1013, 1025 (N.D. Ill. 2011) (excluding expert witness who "did not conduct any independent research" to prepare his report and who failed to "investigate[ ] the veracity of the materials Plaintiff's counsel provided to him, or request[ ] additional materials from Plaintiff's counsel to further inform his opinion").

3. Dr. Finkelman's Opinions Do Not Constitute "Generalized Expert Testimony"

In response to these deficiencies, plaintiffs assert that "some of the content" of Dr. Finkelman's report and testimony consists of "explication of the principles" of social science that may be admitted as "generalized testimony." See Pls.' Opp. to Mot. to Exclude Roth and Finkelman ("Pls.' Roth/Finkelman Opp."), ECF No. 345 at 10-12. Plaintiffs contend that such testimony is appropriate because

it might also be important in some cases for an expert to educate the factfinder about general principles, without ever attempting to apply these principles to the specific facts of the case.... For this kind of generalized testimony, Rule 702 simply requires that: (1) the expert be qualified; (2) the testimony address a subject matter on which the factfinder can be assisted by an expert; (3) the testimony be reliable; and (4) the testimony 'fit' the facts of the case.

Id. at 11 (quoting Fed. R. Evid. 702 advisory committee notes). Plaintiffs do not specify which portions of Dr. Finkelman's report or testimony they seek to admit under this provision. See generally Pls.'

*302Roth/Finkelman Opp., ECF No. 345. Nor do plaintiffs address how, precisely, they believe such testimony about industrial-organizational psychology, without any application to the facts of this case, will assist the Court's class-certification inquiry. Id. Despite plaintiffs' argument to the contrary, Dr. Finkelman's report says little about the "role of unbridled subjectivity in employment selections," see Pls.' Roth/Finkelman Opp., ECF 345 at 12, that is "generalized" and does not "attempt[ ] to apply the[ ] principles to the specific facts of the case," Fed. R. 702 advisory committee notes. See Finkelman Rep., ECF No. 304-3 at 19-25; see also Fox v. Pittsburgh State Univ. , No. 14-2606, 2016 WL 4919463, at *3 (D. Kan. Sept. 15, 2016) (rejecting defendant's argument that expert was merely offering "generalized testimony" where defendant's aim was to "implicitly apply the[ ] principles [offered by the expert] to the specifics of the case").

Moreover, even if the Court were persuaded that Dr. Finkelman offers generalized testimony that could be helpful to the issue of class certification, plaintiffs have failed to show that Dr. Finkelman's testimony is, at bottom, reliable. As explained above, Dr. Finkelman does not cite any studies or other data supporting his opinions, and he did not attempt to measure the degree of subjectivity in Amtrak's hiring, promotion, or disciplinary decisions. Finkelman Dep., ECF No. 319-3 at 27. Moreover, Dr. Finkelman does not account for a host of information, some of which he admits is relevant to the very question he aims to answer. Id. at 22-23, 25, 45. Plaintiffs argument that the Court should accept Dr. Finkelman's testimony simply because he is "a very knowledge psychologist" and "gets it," Pls.' Roth/Finkelman Opp., ECF No. 345 at 27, is not only conclusory, but also inadmissible ipse dixit in its most classic form. See Gen. Elec. Co. v. Joiner , 522 U.S. 136, 146, 118 S.Ct. 512, 139 L.Ed.2d 508 (1997) ; Obrycka v. City of Chicago , 792 F.Supp.2d 1013, 1025 (N.D. Ill. 2011).

4. Dr. Finkelman's Rebuttal Report Does Not Render His Analysis Reliable

Perhaps recognizing the deficiencies in Dr. Finkelman's report, plaintiffs finally contend that, "to the extent there were shortcomings in Dr. Finkelman's initial Expert Report, they are removed by his Expert Rebuttal Report" which "specifically cit[es] the Industrial-Organizational Psychology studies and publications that support his opinions in this case." Pls.' Roth/Finkelman Opp., ECF No. 345 at 28. Although Dr. Finkelman's rebuttal expert report does cite academic literature purportedly supporting his opinions, the rebuttal report fails to address the fundamental problem with his initial expert report: Dr. Finkelman did not review sufficient facts to develop his Amtrak-specific expert opinions. Although Dr. Finkelman need not "examine every document that was filed in this matter" to opine on the adequacy of Amtrak's human-resource management practices, see Finkelman Rebuttal Rep., ECF No. 344-2 at 16, he cannot simply rely on cherry-picked facts selected by plaintiffs' counsel in forming his opinions. See, e.g. , Shawler v. Ergon Asphalt & Emulsions, Inc. , No. 15-2599, 2016 WL 1019121, at *11 (E.D. La. Mar. 15, 2016) (excluding report as unreliable where expert simply "cherry-pick[ed] evidence favorable to [plaintiff] and dictate[d] what inferences and legal conclusions the Court should draw from that evidence").

D. Amtrak's Motion To Exclude Mr. Roth's Report and Testimony Is Denied

1. Mr. Roth's Qualifications And Opinions

Thomas R. Roth is a financial and economic advisor to labor organizations. See *303Roth Rep., ECF No. 304-2 ¶ 1. He holds a bachelor of science in economics and industrial relations and a master of science in labor and industrial relations. Roth Rep. Ex. A, ECF No. 304-2 at 15. Mr. Roth has worked with labor organizations in a variety of sectors, including-and perhaps especially in-the railroad industry. Id. Specifically, he has represented all fourteen standard railroad unions before the Presidential Emergency Board and in arbitration proceedings. Id. Indeed, Mr. Roth affirms that he has "been directly and intimately involved in every round of collective bargaining at Amtrak since 1978." Decl. & Expert Rebuttal Rep. of Thomas R. Roth ("Roth Rebuttal Rep."), ECF No. 344-1 ¶ 2. Currently, he is President of Labor Bureau Inc., a private-consulting firm that provides professional services in labor-relations matters. Roth Rep. Ex. A, ECF No. 304-2 at 15.

In his expert report, Mr. Roth classifies the seventeen different collective-bargaining units at Amtrak into five "functional categories" or "craft groups" that he contends "mirror Amtrak's management structure." Roth Rep., ECF No. 304-2 ¶¶ 8-15. He explains that these five functional groups share common work sites and supervision, along with certain terms in their collective-bargaining agreements. Id. ¶¶ 21-36. Based on these observations, Mr. Roth opines that "it makes sense" to analyze Amtrak's process for hiring, promoting, and disciplining its employees by these functional groups. Dep. of Thomas R. Roth ("Roth Dep."), ECF No. 329-3 at 52, 53-54. Mr. Roth's opinions are based on his personal experience with Amtrak and the labor organizations representing its workforce, his general knowledge of Amtrak's operation and the railroad industry, a review of certain Amtrak collective bargaining agreements, and "other pertinent statistical information" maintained by Mr. Roth or his staff. Roth Rep. ECF No. 304-2 ¶ 2; Roth Dep., ECF No. 329-3 at 5.

2. Mr. Roth's Opinions Are Reliable

Amtrak argues that Mr. Roth's report and testimony should be excluded because they are unreliable. See Def.'s Mem. in Supp. of Mot. to Exclude Roth ("Def.'s Roth Mem."), ECF No. 329-1 at 12-18. First, Amtrak complains that Mr. Roth did not review any pleadings, 30(b)(6) testimony, other expert reports, Amtrak organizational charts, descriptions of the reporting structure at Amtrak, or a host of documents related to specific employment decisions made at Amtrak with respect to plaintiffs. Id. at 14-15. For example, Amtrak points out that Mr. Roth failed to review all the collective-bargaining agreements, instead just reviewing "one [collective-bargaining agreement] per union" even though some unions "have multiple agreements based on geographical location ... and ... have had several applicable agreements during the alleged class period." Id. at 13. Second, Amtrak argues that Mr. Roth's classifications are not reliable because Mr. Roth acknowledged during his deposition that the functional categories may not "neatly" describe the range of employees in each group. Id. at 16-17. And third, Amtrak contends that Mr. Roth ignored contradictory information and focused only on purported commonalities in examining the collective-bargaining agreements. Id. at 17-18.

Plaintiffs respond that Mr. Roth appropriately relied on his "deep, thorough, and encyclopedic" knowledge of the railroad industry, Amtrak's unions, and collective-bargaining agreements in forming his opinions with respect to the functional categories. Pls.' Roth/Finkelman Opp., ECF No. 345 at 12-13. To this end, plaintiffs point out that Amtrak's criticism that Mr. Roth should have "reviewed each and every collective bargaining agreement" rings hollow given that Mr. Roth "was an active participant in negotiating those same collective *304bargaining agreements." Id. at 13. In addition, plaintiffs argue that, notwithstanding Amtrak's arguments to the contrary, Mr. Roth does not purport to opine in his initial expert report that the five functional groups should be used to analyze "every selection decision, discipline decision, and hostile work environment claim for virtually every unionized employee" at Amtrak over the class period. Id. at 17. Rather, plaintiffs claim that Mr. Roth's report does nothing more than "set[ ] forth the existence, nature, and significance of the Craft Groups at Amtrak." Id. Finally, plaintiffs contend that Amtrak's complaints about Mr. Roth's treatment of variations between craft groups do nothing to undermine Mr. Roth's opinion regarding the similarities between the craft groups. Id. at 18.

The Court finds that Amtrak's objections go to the weight to be given to Mr. Roth's testimony and not its reliability. Amtrak does not attack Mr. Roth's qualifications, and the Court finds that Mr. Roth is certainly qualified "by knowledge, skill, experience, training, or education." Fed. R. Evid. 702. Indeed, as Mr. Roth's rebuttal report makes clear, he has "direct and thorough" knowledge of collective bargaining in the railroad industry and is "specifically" knowledgeable about the craft structure that he discusses in his report. Roth Rebuttal Rep., ECF No. 344-1 at 2.

Moreover, Amtrak has not demonstrated that Mr. Roth's failure to review certain materials-namely, pleadings, depositions, other expert reports, personnel files, job descriptions, or the entire universe of collective-bargaining agreements pertaining to Amtrak unions-has rendered his opinions regarding the structure of craft groups at Amtrak so unreliable as to be excluded. See Joy v. Bell Helicopter Textron, Inc. , 999 F.2d 549, 567 (D.C. Cir. 1993) (admission of expert testimony "does not constitute an abuse of discretion merely because the factual bases for an expert's opinion are weak"); Stryker Spine v. Biedermann Motech GmbH , 684 F.Supp.2d 68, 101 (D.D.C. 2010) (expert's failure to review certain records did not render his opinion unreliable where his opinions were based "on his expertise in the relevant field").

Mr. Roth's opinions stand in contrast to those offered by Dr. Finkelman. Whereas Dr. Finkelman's reliance on unverified summaries of cherry-picked deposition testimony provided to him by plaintiffs' counsel rendered the foundation of his testimony unreliable, Mr. Roth merely relied on fewer "facts" than Amtrak prefers. Amtrak is free to challenge the factual bases of Mr. Roth's opinions through "[v]igorous cross-examination" and "presentation of contrary evidence," which are "the traditional and appropriate means of attacking shaky but admissible evidence." Daubert , 509 U.S. at 596, 113 S.Ct. 2786. Moreover, the Court can determine what weight to afford Mr. Roth's opinions as part of its class-certification analysis. See In re Rail Freight Fuel Surcharge Antitrust Litig. , 292 F.Supp.3d 14, 61, 2017 WL 5311533, at *27 (D.D.C. 2017) ("The Court may admit expert opinion even where-as here-the factual bases for the opinion are weak. The Court will determine what weight to afford [the expert's] opinion, given its limited support, under Rule 23.").

3. Mr. Roth's Opinions May Be Relevant

Amtrak also argues that, even if Mr. Roth's testimony is reliable, it should still be excluded because "Mr. Roth has no basis for concluding that the groupings he identifies are relevant to the claims at issue in this case." Def.'s Roth Mem., ECF No. 329-1 at 19. In particular, Amtrak argues that the functional groups set forth in Mr. Roth's report "are in no way based upon the manner in which Amtrak managers *305make selection and discipline decisions" but instead only relate to "collective bargaining issues." Id. at 18-19. Contrary to Amtrak's contentions, the Court finds that Mr. Roth's testimony is "sufficiently tied to the facts of this case" such that it will aid the factfinder in resolving a factual dispute. Ambrosini v. Labarraque , 101 F.3d 129, 134 (D.C. Cir. 1996). According to Mr. Roth, the functional categories are, in fact, appropriate for analyzing hiring, promotion, and discipline processes and procedures at Amtrak. See Roth Dep., ECF No. 329-3 at 52 (hiring procedures and promotion process); id. at 53-54 (disciplinary procedures). The parties' disagreements as to these conclusions go to the weight to be given the evidence and not its admissibility. See Ambrosini , 101 F.3d at 140 (D.C. Cir. 1996) (" Daubert instructs that the admissibility inquiry focuses not on conclusions, but on approaches[.]").

In short, the Court finds that Mr. Roth's opinions are reliable and relevant to plaintiffs' class-certification motion and should not be excluded.

E. Amtrak's Motion To Exclude Dr. Bradley And Dr. Fox's Report and Testimony Is Denied

1. Dr. Bradley And Dr. Fox's Qualifications And Opinions

Dr. Edwin L. Bradley, Jr. and Dr. Liesl M. Fox are statisticians at Quantitative Research Associates, a firm that provides statistical and computing consulting services. See Bradley/Fox Rep., ECF No. 304-1 at 2-3. Dr. Bradley and Dr. Fox were asked to examine the hiring of external job applications, selection of internal candidates for promotions and transfers, and the disciplinary charges levied and resolved against Amtrak employees to ascertain the differences in hiring, promotion, and discipline rates between African-American and non-African-American employees. Id. at 4. The purpose of this analysis was to determine "whether the policies and practices used by Amtrak have had adverse impact against its African-American ... employees" between April 4, 1996 and December 31, 2008 (the "Analysis Period").Id.

To conduct their analysis, Dr. Bradley and Dr. Fox relied on certain databases that were prepared by Amtrak from electronic data sources at Amtrak ("Joint Databases"). Id. at 5-7. The Joint Databases provided information regarding hiring and termination dates, rates of pay, changes in job assignment, race, records of disciplinary violations, and applicant pools for select vacancies. Id. Dr. Bradley and Dr. Fox assigned each employee with a record in the Joint Database to one of four craft groups based on the employee's union membership. Id. at 6. Dr. Bradley testified that he was instructed to aggregate the data based on these craft groups by plaintiffs' counsel. Bradley Dep., ECF No. 331-3 at 5.

In analyzing hiring and promotion decisions, Dr. Bradley and Dr. Fox divided their analysis into two groups: vacancies for which they had "applicant flow data"-i.e., information regarding pools of individuals who actually applied for vacancies between July 2003 and December 2008-and vacancies for which no such data existed. Bradley/Fox Rep., ECF No. 304-1 at 8-16. For those vacancies for which applicant flow data did exist, Dr. Bradly and Dr. Fox compared the selection for each job opening against the specific pool of candidates who applied for the opening, after removing candidates who were not minimally qualified for the position because they, for example, failed a drug screen or failed a relevant skills test. Id. at 8-9. Dr. Bradley testified that, in his analysis of hiring and promotion decisions, he did not control for other types of qualifications *306that could conceivably influence a hiring or promotion decision:

Q. But you don't try to analyze when you're trying to figure out whether or not a component or the overall selection process has adverse impact, you don't consider the types of qualifications that a decision-maker might have looked at when making the decision, like experience and other types of qualifications?

A. No, I'm not thinking of that. I'm looking only at minimum qualifications.

Bradley Dep., ECF No. 331-3 at 28; see also id. at 56-57 (testifying that he did not take educational attainment into consideration in evaluating hiring or promotion decisions). Dr. Bradley also admitted that, although internal candidates were "always preferred" in the employment selection process, he did not control for that in his analysis of Amtrak's selection process. Id. at 23-24. Based on this analysis, Dr. Bradley and Dr. Fox found that there were a total of 6,193 individuals selected for vacancies for which there was Applicant Flow Data across all craft groups. Bradley/Fox Rep., ECF No. 304-1 at 10. Of those, only 2,335 individuals were African-American individuals, even though one would have expected 2,589 African-American selections based on the proportion of African-American candidates in the pool of applicants. Id.

For the majority of vacancies-approximately 49,000 of them-there was no Applicant Flow Data. Id. at 11. For those vacancies, Dr. Bradley and Dr. Fox analyzed selections using proxy benchmarks based on the vacancies for which applicant data did exist to represent African-American availability. Id. at 11-12. The same extrapolated benchmark was applied to every internal selection or external hire decision within a particular craft group. Id. 11-14, 26-27. For the vacancies for which Amtrak made external hires, Dr. Bradley and Dr. Fox found that there were a total of 10,074 individuals selected across all craft groups; of these, only 3,577 of the individuals selected were African-American, although 4,312 African-American selections were expected based on the benchmarks. Id. at 13. With respect to positions that were eventually filled through internal promotions rather than external hires, Dr. Bradley and Dr. Fox found that there were a total of 39,548 vacancies across all craft groups. Id. at 14. Of those, 12,834 were filled by African-American individuals, even though one would have expected 14,899 African-American selections. Id. In other words, Dr. Bradley and Dr. Fox concluded that African-American individuals were hired less and selected for fewer competitive promotions than their non-African-American counterparts.

Finally, Dr. Bradley and Dr. Fox analyzed disciplinary charges and resulting outcomes for Amtrak employees. Id. at 16-17. They found that there were 10,796 employees who were issued a total of 24,136 disciplinary charges during the Analysis Period. Of those charges, 10,651 charges were issued to 4,175 African-American employees, even though one would expect only 8,924 charges to be issued to that group if the disciplinary process was race-neutral. Id. at 16. Dr. Bradley and Dr. Fox also found that African-American employees were terminated and received formal reprimands or deferred suspensions at statistically higher rates than non-African-American employees. Id. at 17. Based on their review, Dr. Bradley and Dr. Fox concluded that African-American employees were charged with disciplinary violations at a rate higher than their non-African-American counterparts. Id.

2. The Bradley/Fox Report Is Sufficiently Reliable

Amtrak does not dispute that Dr. Bradley and Dr. Fox are qualified to offer *307statistical expert testimony. Instead, Amtrak argues that the Court should not consider Dr. Bradley and Dr. Fox's expert report and testimony because their opinions are unreliable. Amtrak's argument focuses on Dr. Bradley and Dr. Fox's use of extrapolated benchmarks to assess racial disparities in Amtrak's hiring and promotional decisions. Def.'s Mem. in Supp. Mot. to Exclude Bradley and Fox (Def.'s Bradley/Fox Mem."), ECF No. 331-1 at 26-32. Specifically, Amtrak argues that, before extrapolating benchmarks for African-American hiring and promotion from the applicant flow data, Dr. Bradley and Dr. Fox were required to ensure that the applicant flow data was representative of the applicant pool to which the extrapolated benchmarks would apply. Id. at 28-29. Amtrak also argues that Dr. Bradley and Dr. Fox did not assess whether the sample size was sufficient to extrapolate the benchmarks. Id. at 29. Finally, Amtrak contends that Dr. Bradley and Dr. Fox erred by treating each benchmark as an "exact, known value" rather than "a sample estimate within a margin of error." Id. at 29-30.

While Amtrak points to potential problems with Dr. Bradley and Dr. Fox's extrapolation techniques, it fails to establish that these experts used a methodology so unreliable as to warrant exclusion of their report. There is no evidence that Dr. Bradley and Dr. Fox cherry-picked the data points in constructing the benchmarks for African-American availability where that data was not kept in the regular course by Amtrak. It is, of course, clear that the experts' extrapolated benchmarks in areas where no applicant flow data was available is less precise than Amtrak's actual applicant flow data. It is also clear, however, that plaintiffs may rely on reliable estimates when actual data is unavailable. See General Elec. Co. v. Joiner , 522 U.S. 136, 146, 118 S.Ct. 512, 139 L.Ed.2d 508 (1997) (stating the well-established principle that "[t]rained experts commonly extrapolate from existing data"); see generally Ramona L. Paetzold & Steven L. Willborn, The Statistics of Discrimination: Using Statistical Evidence in Discrimination Cases § 4.03 (2014) (describing the use of proxy data when actual data is unavailable or unreliable). It may be that Dr. Bradley and Dr. Fox's sample size of approximately 6,200 was too small, and perhaps a larger sample would have revealed fewer differences between the hiring and promotion of African-American individuals as compared to their non-African-American counterparts. Such a criticism can be brought out in cross-examination and does not render Dr. Bradley and Dr. Fox's methodology so unreliable that it should not be admitted. See, e.g. , Equal Emp't Opportunity Comm'n v. Texas Roadhouse, Inc. , 215 F.Supp.3d 140, 155 (D. Mass. 2016) ("Even when statistical analysis has involved general population census data to show discriminatory intent, it has not been precluded on Fed. R. Evid. 702 grounds.").

3. The Bradley/Fox Report Has Limited Probative Value

Amtrak spends the bulk of its brief arguing that plaintiffs' statistical evidence is irrelevant to its class-certification motion because Dr. Bradley and Dr. Fox did not study a particular employment practice or the decisions of any common decision-maker. See Def.'s Bradley/Fox Mem. ECF No. 331-1 at 9-26. As such, Amtrak argues that Dr. Bradley and Dr. Fox's analysis will be unhelpful to the trier of fact because these experts cannot opine that any specific employment practice caused the alleged statistical disparities. Id. at 16. Amtrak notes that plaintiffs' experts could have utilized the job files produced in discovery that contained candidate records, applications, selection criteria, rating sheets, and other records relating *308to each of the selections contained in the Joint Database. Id. at 7, 11, 36.

Plaintiffs respond that the components of Amtrak's selection process "were not able to be separated for analysis because they were interwoven and overlapping parts of a singular process." Pls.' Opp. to Mot. to Exclude Bradley and Fox, ECF No. 342 at 10-15. Plaintiffs further claim that data to do such an analysis was not available. Id. at 10, 15-21.

Plaintiffs bear the initial burden of making out a prima facie case of discrimination. Cooper v. Fed. Reserve Bank of Richmond , 467 U.S. 867, 874, 104 S.Ct. 2794, 81 L.Ed.2d 718 (1984). And because plaintiffs allege a system-wide pattern or practice of discrimination, plaintiffs have "to prove more than the mere occurrence of isolated or 'accidental' or sporadic discriminatory acts." Int'l Bhd. of Teamsters v. United States , 431 U.S. 324, 336, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977). Rather, plaintiffs have "to establish by a preponderance of the evidence" that racial discrimination was Amtrak's "standard operating procedure-the regular rather than the unusual practice." Id.

In a case such as this, then, statistical data is relevant because it can be used to establish a general discriminatory pattern in an employer's hiring or promotion practices. As the Supreme Court explained,

[s]tatistics showing racial or ethnic imbalance are probative ... because such imbalance is often a telltale sign of purposeful discrimination; absent explanation, it is ordinarily to be expected that nondiscriminatory hiring practices will in time result in a work force more or less representative of the racial and ethnic composition of the population in the community from which the employees are hired. Considerations such as small sample size may, of course, detract from the value of such evidence, and evidence showing that the figures for the general population might not accurately reflect the pool of qualified job applicants would also be relevant.

Int'l Bhd. of Teamsters , 431 U.S. at 340, 97 S.Ct. 1843. Importantly, in most cases in which plaintiffs allege a disparate impact, plaintiffs must do more than simply "show that there are statistical disparities in the employer's work force." Watson v. Fort Worth Bank & Trust , 487 U.S. 977, 994, 108 S.Ct. 2777, 101 L.Ed.2d 827 (1988). Rather, plaintiffs are responsible for "isolating and identifying the specific employment practices that are allegedly responsible for any observed statistical disparities." Id.

In this respect, plaintiffs' statistical evidence suffers from a number of shortcomings. For example, Dr. Bradley admitted that he did not study whether a particular employment practice at Amtrak caused an adverse impact. Bradley Dep., ECF No. 331-3 at 27. Moreover, although plaintiffs' claim that Amtrak's employment practices are "incapable of being separated for analysis" and therefore subject to a bottom-line analysis as to disparate impact, 42 U.S.C. § 2000e-2(k)(1)(B)(i), Dr. Bradley and Dr. Fox's report does nothing to demonstrate this fact. Their report does not grapple with the data contained in Amtrak's job files or explain how it was inadequate to render a statistical analysis as to a particular employment practice.

Nonetheless, although the statistical study proffered by plaintiffs' experts may, ultimately, be inadequate to satisfy plaintiffs' burden on the merits, the Court declines to exclude it as irrelevant at this time. As other courts have found, a "statistical study may fall short of proving the plaintiff's case, but still remain relevant to the issues in dispute." Obrey v. Johnson , 400 F.3d 691, 695 (9th Cir. 2005). For example, in Obrey , the defendant challenged *309the admission of the plaintiff's expert report because, inter alia , the statistical analysis was irrelevant to plaintiff's claims that the defendant engaged in discriminatory hiring practices. Id. at 694. The statistical analysis in that case only analyzed the race of the managers selected by the employer as compared to the race of those who applied for managerial positions-and, just like in this case, did not take into account the relative qualifications of the applicant pool or evaluate any specific employment practice. Id. at 694-698. While this evidence "by itself" could not "constitute proof that the [employer] has discriminated against [the plaintiff]," the court explained that "it should have been admitted for whatever probative value it had." Id. at 697. In other words, "defendant's objections to the admission of [the statistical evidence] went to weight and sufficiency rather than admissibility." Id. ; see also, e.g. , Puffer v. Allstate Ins. Co. , 255 F.R.D. 450, 462 (N.D. Ill. 2009) (although plaintiff's statistical expert "fail[ed] to link any pay differential that she found to any [employer] policy or practice," that deficiency did not render the report "irrelevant" but rather simply limited its probative value). Accordingly, the Court declines to exclude Dr. Bradley and Dr. Fox's testimony or report.

IV. EXCLUSION OF OTHER EVIDENCE

Amtrak has also moved to strike portions of the declarations submitted by plaintiffs in support of their request for class certification. Def.'s Mot. to Strike Decls. Of Pls., ECF No. 330. Plaintiffs oppose that motion, arguing both that the admissibility standards on which Amtrak relies are inapplicable at the class-certification stage, and that, in any event, the evidence is admissible. Pls.' Opp. to Mot. to Strike Decls. Of Pls., ECF No. 340.

Although the Supreme Court has not directly addressed the issue of whether evidence proffered in support of a motion for class certification must be admissible if it is to be considered by the court in determining class certification, its dicta in Eisen v. Carlisle & Jacquelin , 417 U.S. 156, 94 S.Ct. 2140, 40 L.Ed.2d 732 (1974), led some courts to find that lax evidentiary standards were appropriate.³ In Eisen , the Supreme Court emphasized that a court's decision at the class-certification stage "is not accompanied by the traditional rules and procedures applicable to civil trials." Id. at 178, 94 S.Ct. 2140. This statement came in the context of the Court's oft-cited conclusion-arrived at in the course of overruling a district court's decision to direct a defendant to cover some of the costs of providing notice to the class on the basis that plaintiffs were likely to succeed on the merits-that "[w]e find nothing in either the language or history of Rule 23 that gives a court 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." Id. at 177, 94 S.Ct. 2140.

Relying upon the statement regarding the absence of "traditional rules and procedures" at the class-certification stage, the D.C. Circuit held in an unpublished opinion that the rules of evidence do not apply at the class-certification stage. See *310In re Rand Corp. , No. 02-8007, 2002 WL 1461810, at *1 (D.C. Cir. July 8, 2002) ("[T]he propriety of a district court's refusal to scrutinize for admissibility and probative value evidence proffered to demonstrate that the requirements of Federal Rule of Civil Procedure 23(a) are satisfied is well-settled"). District courts, as plaintiffs recite, came to similar conclusions. See, e.g. , Disability Rts. Council v. Wash. Metro. Area Transit Auth. , 239 F.R.D. 9 (D.D.C. 2006).

The Supreme Court's decisions in Dukes and Comcast , however, have since shifted this landscape. In Dukes , the Supreme Court made clear that "[a] party seeking class certification must affirmatively demonstrate his compliance with [ Rule 23 ]-that is, he must be prepared to prove that there are in fact sufficiently numerous parties, common questions of law or fact, etc." Wal-Mart Stores, Inc. v. Dukes , 564 U.S. 338, 350, 131 S.Ct. 2541, 180 L.Ed.2d 374 (2011). The Court further explained that the "rigorous analysis" a district court must perform of a plaintiffs' claim for certification " '[f]requently ... will entail some overlap with the merits of the plaintiff's underlying claim.' " Id. at 351, 131 S.Ct. 2541. The Court elaborated:

A statement in one of our prior cases, Eisen v. Carlisle & Jacquelin , 417 U.S. 156, 177 [94 S.Ct. 2140, 40 L.Ed.2d 732] (1974), is sometimes mistakenly cited to the contrary: 'We find nothing in either the language or history of Rule 23 that gives a court 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.' But in that case, the judge had conducted a preliminary inquiry into the merits of a suit, not in order to determine the propriety of certification under Rule 23(a) and (b)... but in order to shift the cost of notice required by Rule 23(c)(2) from the plaintiff to the defendants. To the extent the quoted statement goes beyond the permissibility of a merits inquiry for any other pretrial purpose, it is the purest dictum and is contradicted by our other cases.

Id. 351 n.6, 131 S.Ct. 2541.

In Comcast , the Supreme Court reiterated the importance of considering all merits questions that may bear on any of the Rule 23 factors. See Comcast Corp. v. Behrend , 569 U.S. 27, 35, 133 S.Ct. 1426, 185 L.Ed.2d 515 (2013). In so doing, the Court reiterated the requirement that a party seeking certification affirmatively demonstrate compliance with Rule 23"through evidentiary proof ." Id. at 33, 133 S.Ct. 1426. The D.C. Circuit has recognized that, post- Comcast , "[i]t is now indisputably the role of the district court to scrutinize the evidence before granting certification, even when doing so 'requires inquiry into the merits of the claim.' " In re Rail Freight Fuel Surcharge Antitrust Litig. , 725 F.3d 244, 253 (D.C. Cir. 2013) (quoting Comcast , 569 U.S. at 35, 133 S.Ct. 1426 ).

These cases do not address the precise question before the Court-i.e., whether class-certification evidence must meet the standards of admissibility as set forth in the Federal Rules of Evidence. Moreover, although the Supreme Court granted certiorari in Comcast to decide "[w]hether a district court may certify a class action without resolving whether the plaintiff class had introduced admissible evidence," it ultimately did not decide that question because defendants had failed to raise an appropriate objection on that ground before the trial court and thus forfeited the issue. Comcast , 569 U.S. at 33 n.4, 133 S.Ct. 1426. These decisions suggest, however, that when a party objects to evidence provided in support of class certification, a district court must assess the admissibility of that evidence before certifying a class. As such, the Court concludes that, to the extent it relies on any controverted portions of plaintiffs' proffered declarations to *311support its class-certification ruling, it must first address Amtrak's evidentiary objections.⁴ Accordingly, in the course of its analysis, the Court will resolve Amtrak's objections to the portions of plaintiffs' declarations that the Court relies on in reaching its class-certification decision. To the extent the Court does not rely on the declarations, the Court will not address Amtrak's evidentiary objections, and Amtrak must re-raise them in future proceedings.

V. CLASS CERTIFICATION

"The class action is 'an exception to the usual rule that litigation is conducted by and on behalf of the individual named parties only.' " Comcast Corp. v. Behrend , 569 U.S. 27, 33, 133 S.Ct. 1426, 185 L.Ed.2d 515 (2013) (citation omitted). Class certification is governed by Federal Rule of Civil Procedure 23, which requires a plaintiff to demonstrate that the requirements of Rule 23(a) are met and that the class is maintainable pursuant to one of Rule 23(b)'s subdivisions. See Richards v. Delta Air Lines, Inc. , 453 F.3d 525, 529 (D.C. Cir. 2006).

The D.C. Circuit has made clear that class certification "is far from automatic." In re Rail Freight Fuel Surcharge Antitrust Litig. , 725 F.3d 244, 249 (D.C. Cir. 2013). Rather, a plaintiff seeking certification of a class must "affirmatively demonstrate his compliance" with the requirements of Rule 23. Wal-Mart Stores, Inc. v. Dukes , 564 U.S. 338, 350, 131 S.Ct. 2541, 180 L.Ed.2d 374 (2011). This is done not by pleading compliance with Rule 23, but by demonstrating "compliance ... in fact . " Id. At times, determining whether the proponent of a class has satisfied the requirements of Rule 23"resembles an appraisal of the merits, for 'it may be necessary for the court to probe behind the pleadings before coming to rest on the certification question.' " In re Rail Freight Fuel Surcharge Antitrust Litig. , 725 F.3d at 244 (citation omitted). The Court may not, however "consider merits questions that do not overlap with Rule 23's requirements." Coleman through Bunn v. Dist. of Columbia , 306 F.R.D. 68, 77 (D.D.C. 2015). "Ultimately, the district court's determination must rest on a 'rigorous analysis' to ensure that all the requirements are satisfied, and '[a]ctual, not presumed, conformance' with Rule 23 is indispensable." Burton v. Dist. of Columbia , 277 F.R.D. 224, 228 (D.D.C. 2011) (citation omitted).⁵

*312A. Plaintiffs' Proposed Classes

The named plaintiffs in this case are seventy-one African-American Amtrak employees, former Amtrak employees, or applicants for employment at Amtrak. Plaintiffs move to certify the following classes or, in the alternative, subclasses:

(1) All Black employees of Amtrak who are represented for purposes of collective bargaining by any labor union (except those who have worked only in the Northeast Corridor and are represented for purposes of collective bargaining by the Pennsylvania Federation of the Brotherhood of Maintenance of Way Employees) (herein, "Black CBA employees") who, since April 4, 1996, have been discriminated against because of their race or color in regard to competitive promotion selections; and/or in the alternative,

(a) a subclass of all Black CBA employees who, since April 4, 1996, have worked for Amtrak in any of the Shop Crafts who raise such promotion selection claims;

(b) a subclass of all Black CBA employees who, since April 4, 1996, have worked for Amtrak in any of the Engineering Crafts who raise such promotion selection claims;

(c) a subclass of all Black CBA employees who, since April 4, 1996, have worked for Amtrak in any of the Operating and Police Crafts who raise such promotion selection claims;

(d) a subclass of all Black CBA employees who, since April 4, 1996, have worked for Amtrak in any of the Clerical and On-Board Services Crafts who raise such promotion selection claims;

(2) All Black CBA employees of Amtrak who, since April 4, 1996, have been exposed to a racially hostile work environment, as embodied in racial harassment and/or racial discrimination in regard to training, job assignments, work assignments, noncompetitive transfers, scheduling, and other terms and conditions of employment; and/or in the alternative,

(a) a subclass of all Black CBA employees who, since April 4, 1996, have worked for Amtrak in any of the Shop Crafts and have been exposed to such a racially hostile work environment;

(b) a subclass of all Black CBA employees who, since April 4, 1996, have worked for Amtrak in any of the Engineering Crafts and have been exposed to such a racially hostile work environment;

(c) a subclass of all Black CBA employees who, since April 4, 1996, have worked for Amtrak in any of the Operating and Police Crafts and have been exposed to such a racially hostile work environment;

(d) a subclass of all Black CBA employees who, since April 4, 1996, have worked for Amtrak in any of the Clerical and On-Board Services Crafts and have been exposed to such a racially hostile work environment;

(3) All Black CBA employees of Amtrak who, since April 4, 1996, have been discriminated against in regard to discipline *313or termination; and/or in the alternative,

(a) a subclass of all Black CBA employees who, since April 4, 1996, have worked for Amtrak in any of the Shop Crafts and have been discriminated against in regard to discipline or termination;

(b) a subclass of all Black CBA employees who, since April 4, 1996, have worked for Amtrak in any of the Engineering Crafts and have been discriminated against in regard to discipline or termination;

(c) a subclass of all Black CBA employees who, since April 4, 1996, have worked for Amtrak in any of the Operating and Police Crafts and have been discriminated against in regard to discipline or termination;

(d) a subclass of all Black CBA employees who, since April 4, 1996, have worked for Amtrak in any of the Clerical and On-Board Services Crafts and have been discriminated against in regard to discipline or termination; and

(4) All Black CBA who have applied to work for Amtrak for any position(s) that would be represented for purposes of collective bargaining by any labor union since April 4, 1996, and been denied employment because of their race.

Pls.' Mot. for Class Cert., ECF No. 303 at 1-3.

B. Plaintiffs' Proposed Classes And Subclasses Are Impermissibly Fail-Safe

Although not specifically mentioned in Rule 23, there is an "implied requirement" that the class be "adequately defined" and "clearly ascertainable" before it can be certified. Thorpe v. Dist. of Columbia , 303 F.R.D. 120, 139 (D.D.C. 2014) (citation omitted). This "common-sense requirement" demands that the plaintiff "be able to establish [that] the general outlines of the membership of the class are determinable at the outset of the litigation." Id. (citation omitted); see also, e.g. , Johnson v. Dist. of Columbia , 248 F.R.D. 46, 52 (D.D.C. 2008) ("It is axiomatic that for a class action to be certified a 'class' must exist."); In re Rail Freight Fuel Surcharge Antitrust Litig. , 292 F.Supp.3d 14, 92, 2017 WL 5311533, at *51 (D.D.C. 2017) (explaining that an "implied" requirement of Rule 23 is that the putative class be " 'sufficiently defined so as to be identifiable as a class' ") (citation omitted). "Accordingly, a class may be certified only when 'an individual would be able to determine, simply by reading the [class] definition, whether he or she [is] a member of the proposed class.' " Artis v. Yellen , 307 F.R.D. 13, 23 (D.D.C. 2014) (citation omitted).

One aspect of this requirement that the class be adequately defined is that it not be a "fail-safe class"-i.e., that the class definition not depend on the merits of the underlying claim. See, e.g. , McCaster v. Darden Rests., Inc. , 845 F.3d 794, 799 (7th Cir. 2017) (a fail-safe class is one that " 'is defined so that whether a person qualifies as a member depends on whether the person has a valid claim' ") (citation omitted); Young v. Nationwide Mut. Ins. Co. , 693 F.3d 532, 538 (6th Cir. 2012) (defining a fail-safe class as "a class that cannot be defined until the case is resolved on its merits"). A fail-safe class is impermissible because "a class member either wins or, by virtue of losing, is defined out of the class and is therefore not bound by the judgment." Messner v. Northshore Univ. HealthSystem , 669 F.3d 802, 825 (7th Cir. 2012). Moreover, by " '[u]sing a future decision on the merits to specify the scope of the class,' " a fail-safe class definition " 'makes it impossible to determine who in the class until the case ends.' " Artis v. Yellen , 307 F.R.D. 13, 24 (D.D.C. 2014) (quoting *314Bolden v. Walsh Const. Co. , 688 F.3d 893, 895 (7th Cir. 2012) ); see also Williams v. Glickman , No. 95-1149, 1997 WL 33772612, at *4 (D.D.C. Feb. 14, 1997) (class not clearly defined because, to ascertain class membership, the court would be required to "answer several fact-intensive questions").

Plaintiffs move to certify the following classes:

(1) All Black CBA employees who ... have been discriminated against because of their race or color in regard to competitive promotion selections;

(2) All Black CBA employees who ... have been exposed to a racially hostile work environment , as embodied in racial harassment and/or racial discrimination, in regard to training, job assignments, work assignments, non-competitive transfers, scheduling, and other terms and conditions of employment;

(3) All Black CBA employees who ... have been discriminated against in regard to discipline or termination ; and/or in the alternative,

(4) All Black CBA who have applied to work for Amtrak for any position(s) that would be represented for purposes of collective bargaining by any labor union ... and been denied employment because of their race .

Pls.' Mot. to Certify Class, ECF No. 303 at 1-3 (emphases added).

Each of these proposed class definitions makes membership in the class contingent on individualized merits determinations as to whether the individual suffered discrimination because of his race, was exposed to racial harassment and/or racial discrimination, or was denied employment because of his race. In other words, to determine whether any individual is a member of one of these putative classes, the Court would be required to answer a critical question that goes directly to the merits of the litigation: did the individual suffer racial discrimination at the hands of Amtrak?

Moreover, should the Court or a jury conclude that Amtrak's employment decisions were not based on race, plaintiffs' proposed classes would contain no members. Each of the supposed class members would then be free to file new lawsuits attributing their adverse employment decisions to some other impermissible criteria, thereby depriving the judgment of any preclusive effect. See, e.g. , Bolden v. Walsh Const. Co. , 688 F.3d 893, 895 (7th Cir. 2012) (class defined as persons who did not earn more "because of their race" made "it impossible to determine who was in the class until the case ends," creating the prospect that, if the employer prevailed on the merits, any former worker "could file a new suit, given that the losing 'class' lacked any members"); Lucas v. Vee Pak, Inc. , 68 F.Supp.3d 870, 880 (N.D. Ill. 2014) ("All variations set forth in the complaint beg the question of liability, in that the class is made up of certain African-American workers who were not hired 'because of their race.' It is therefore a 'fail-safe' class: 'one that is defined so that whether a person qualifies as a member depends on whether the person has a valid claim.' ") (citation omitted).

The parties have failed to address these problems with plaintiffs' class definition, perhaps because the problems are repairable. Plaintiffs could, for example, redefine their classes so that membership is not contingent on whether the individual suffered racial discrimination. Cf. Messner , 669 F.3d at 825 ("Defining a class so as to avoid, on one hand, being over-inclusive and, on the other hand, the fail-safe problem is more of an art than a science. Either problem can and often should be solved by refining the class definition rather than by flatly denying class certification on that basis."); In re AutoZone, Inc., Wage & Hour Emp't Practices Litig. , 289 F.R.D. 526, 546 (N.D. Cal. 2012) ("Rather than denying certification on the basis of the fail-safe definition, the Court would *315have discretion here to redefine the class as 'all employees who sought and did not receive reimbursement for mileage,' which seems to avoid both ascertainability problems. This problem is therefore not insurmountable."). Accordingly, in the interest of judicial economy, the Court will address the parties' arguments regarding the requirements of Rule 23.

C. Plaintiffs Have Failed to Demonstrate That The Proposed Class Action Satisfies The Commonality Requirement Of Rule 23(a)

A plaintiff seeking class certification must establish that "there are questions of law or fact common to the class." Fed. R. Civ. P. 23(a)(2). The Supreme Court's decision in Wal-Mart Stores, Inc. v. Dukes , 564 U.S. 338, 131 S.Ct. 2541, 180 L.Ed.2d 374 (2011), "changed the landscape" that the district court must navigate when considering whether a putative class action satisfies Rule 23(a)'s commonality requirement. D.L. v. Dist. of Columbia , 713 F.3d 120, 126 (D.C. Cir. 2013). In Dukes , the plaintiffs alleged that Wal-Mart discriminated against female employees by denying them equal pay or promotions as compared with male employees. 564 U.S. at 343, 131 S.Ct. 2541. Specifically, the plaintiffs claimed that their local managers' discretion over pay and promotions was exercised disproportionately in favor of men, leading to an unlawful disparate impact on, and disparate treatment of, female employees. Id. at 344-45, 131 S.Ct. 2541. The plaintiffs claimed that this discrimination was "common to all Wal-Mart's female employees" and that "a strong and uniform 'corporate culture' permit[ted] bias against women to infect, perhaps subconsciously, the discretionary decisionmaking of each one of Wal-Mart's thousands of managers." Id. at 345, 131 S.Ct. 2541.

The district court certified a class consisting of Wal-Mart's female employees who "have been or may be subjected to Wal-Mart's challenged pay and management track promotions policies and practices." Id. at 346, 131 S.Ct. 2541. The Court of Appeals for the Ninth Circuit substantially affirmed, but the Supreme Court reversed, denying class certification for failure to satisfy the commonality requirement of Rule 23(a)(2). Id. at 345-59, 131 S.Ct. 2541.

In so doing, the Supreme Court first explained that Rule 23's commonality requirement "is easy to misread, since any competently crafted class complaint literally raises common 'questions.' " Id. at 349, 131 S.Ct. 2541 (citation and internal quotation marks omitted). Commonality requires more than common questions; it requires "the plaintiff to demonstrate that the class members 'have suffered the same injury.' " Id. at 350, 131 S.Ct. 2541 (quoting Gen. Tel. Co. v. Falcon , 457 U.S. 147, 157, 102 S.Ct. 2364, 72 L.Ed.2d 740 (1982) ). In other words, the claims of the putative class members "must depend upon a common contention" that is "of such a nature that it is capable of classwide resolution"-i.e., "that determination of its truth or falsity will resolve an issue that is central to the validity of each one of the claims in one stroke." Id. The Dukes plaintiffs identified only a general policy "of allowing discretion by local supervisors over employment matters"-effectively "a policy against having uniform employment practices." Id. at 355, 131 S.Ct. 2541. Resolution of the legality of any one manager's exercise of discretion, then, would have no bearing on the legality of any other manager's action, absent "some glue holding the alleged reasons for all those decisions together." Id. at 351, 131 S.Ct. 2541. The Supreme Court noted that such glue could be provided "if the employer 'used a biased testing procedure' " or upon " '[s]ignificant proof that an employer operated under a general policy of discrimination.' " Id. at 353, 131 S.Ct. 2541 (quoting Falcon , 457 U.S. at 159 n.15, 102 S.Ct. 2364 ).

*316Plaintiffs do not seriously contend that they can demonstrate commonality under Falcon 's first scenario here. While plaintiffs assert that "a biased testing procedure" need not be limited to "paper-and-pencil tests" and suggest that Amtrak's "selection interview process, ratings, rank-orderings, input from other managers, amorphous decision making, and the disqualifying discipline criterion" all qualify, plaintiffs give that argument short shrift, and with good reason. See Pls.' Mem. in Supp. of Mot. for Class Cert. ("Pls.' Class Cert. Mem."), ECF No. 303 at 25. For one, plaintiffs nowhere explain how some of these practices-for example, "amorphous decision making" or seeking "input from other managers" before hiring or promoting individuals-can be considered "non-subjective" criteria. Id. Even if they had, plaintiffs do not allege, much less demonstrate, that all members of the putative class were subject to the same set of objective procedures or policies. To the contrary, the evidence shows that the interview process, assignment of ratings, and rank orderings varied depending on the position and the panel of interviewers involved. See, e.g. , Ray Decl., ECF No. 322-5 ¶¶ 11-41; Wu Decl., ECF No. 322-8 ¶ 4; Allan Decl., ECF No. 321-3 ¶ 5.

Moreover, plaintiffs do not explain with any detail how any such policies "operated in a biased way." Burton v. Dist. of Columbia , 277 F.R.D. 224, 229 (D.D.C. 2011) (plaintiffs allegation that defendant "used a biased testing procedure" was insufficient where plaintiffs provided no "detail about how those examinations operated in a biased way"); see also Ross v. Lockheed Martin Corp. , 267 F.Supp.3d 174, 197-98 (D.D.C. 2017) (despite plaintiffs' assertion that defendant's "companywide evaluation method" resulted in employee ratings that were "poorly correlated with job performance," plaintiffs did not provide "an account of how those procedures themselves resulted in the racially disparate outcome that Plaintiffs have observed in [defendant]'s overall workforce").

Plaintiffs' theories of commonality under Falcon 's second scenario are far from clear. Plaintiffs acknowledge that Amtrak's express policies forbid racial discrimination and retaliation. Pls.' Class Cert. Mem., ECF No. 303 at 14. Instead, plaintiffs argue that "Amtrak's centrally-imposed policies, practices, and procedures," while uniform and racially-neutral, were open to "variations in practice" by supervisors which led to racial bias in employment decisions. Id. at 26; see also id. at 24 (asserting that "all of Amtrak's employment policies are entirely uniform nationwide, but are poorly implemented"); id. at 30 (there were "deviations or variations" from "commonly applicable procedures" that permitted "the influx of racially discriminatory bias"); id. at 7 ("Depositions of scores of Amtrak managers reflected that the standard selection process was followed throughout the country across all crafts; however, they also revealed numerous variations which allowed for the infusion of subjective qualities.").

Stated this way, plaintiffs' theory of commonality rests on the contention that Amtrak allowed lower-level employees to deviate from standard policies, which resulted in employment decisions being infected by bias. This theory is the same as the theory of commonality rejected in Dukes . Compare Pls.' Class Cert. Mem., ECF No. 303 at 26 ("Plaintiffs here show the pattern or practice [of racial discrimination] by demonstrating that Amtrak's centrally-imposed policies, practices, and procedures were in place and implemented throughout Amtrak's system, that Amtrak headquarters mandated that the polices be used companywide, and that variations in practice open the process up to the influences of bias[.]"), with Dukes v. Wal-Mart Stores, Inc. , 222 F.R.D. 137, 145 (N.D. Cal. 2004) ("Plaintiffs present evidence that *317Wal-Mart's policies governing compensation and promotions are similar across all stores, and build in a common feature of excessive subjectivity which provides a conduit for gender bias that affects all class members in a similar fashion."). As the Dukes court explained, such a "policy" of delegating discretion is "just the opposite of a uniform employment practice that would provide the commonality needed for a class action." Dukes , 564 U.S. at 355, 131 S.Ct. 2541.⁶

Indeed, in the wake of Dukes , courts "have generally denied certification when allegedly discriminatory policies are highly discretionary." Tabor v. Hilti, Inc. , 703 F.3d 1206, 1229 (10th Cir. 2013). But Dukes "did not set out a per se rule against class certification where subjective decisionmaking or discretion is alleged"; rather, "to satisfy commonality, a plaintiff must demonstrate that the exercise of discretion is tied to a specific employment practice, and that the 'subjective practice at issue affected the class in a uniform manner.' " Scott v. Family Dollar Stores, Inc. , 733 F.3d 105, 113 (4th Cir. 2013) (citation omitted). In other words, the requisite "glue" may be provided by "unit[ing] acts of discretion under a single policy or practice, or through a single mode of exercising discretion." In re Countrywide Fin. Corp. Mortg. Lending Practices Litig. , 708 F.3d 704, 708 (6th Cir.2013) ; see also Tabor , 703 F.3d at 1229 (to satisfy commonality, plaintiffs must "point to 'a common mode of exercising discretion that pervades the entire company' ") (citation omitted).

Plaintiffs suggest that they provide such "glue" because Amtrak's common and uniform employment policies-as set out in corporate policies governing promotions and transfers and through various collective-bargaining provisions-unite the individual acts of discretion of lower-level employees. See Pls.' Reply in Supp. of Mot. to Certify Class, ECF No. 344 at 6-14.⁷ To be sure, the Dukes Court recognized that, " 'in appropriate cases,' giving discretion to *318lower-level supervisors can be the basis of Title VII liability ... since 'an employer's undisciplined system of subjective decisionmaking [can have] precisely the same effects as a system pervaded by impermissible intentional discrimination.' " Dukes , 564 U.S. at 355, 131 S.Ct. 2541 (quoting Watson v. Fort Worth Bank & Trust , 487 U.S. 977, 990-91, 108 S.Ct. 2777, 101 L.Ed.2d 827 (1988) ). But "demonstrating the invalidity of one manager's use of discretion will do nothing to demonstrate the invalidity of another's," and therefore, in the usual course, a party seeking class certification "will be unable to show that all the employees' Title VII claims will in fact depend on answers to common questions." Id. at 355-56, 131 S.Ct. 2541 ; cf. Garcia v. Johanns , 444 F.3d 625, 632 (D.C. Cir. 2006) ("Establishing commonality for a disparate treatment class is particularly difficult where, as here, multiple decisionmakers with significant local autonomy exist."). Thus, to establish commonality under such a theory, plaintiffs must show how Amtrak's uniform polices resulted in a common mode of exercising discretion that pervaded the entire company and led to the discrimination about which plaintiffs complain.

Plaintiffs try to make such a showing by pointing to (1) the testimony of their expert that Amtrak's uniform employment policies were vulnerable to bias; (2) statistical evidence that shows disparities in selection and discipline rates between African-American and non-African-American individuals; (3) anecdotal evidence from members of the putative class describing instances of racial discrimination; and (4) the testimony of a former Amtrak employee discussing the inadequacies in Amtrak's handling of discrimination complaints. As explained more fully below, the Court finds that plaintiffs have failed to put forward "significant proof" that any alleged disparate outcomes in Amtrak's hiring, promoting, and disciplinary decisions are the result of a common mode of exercising discretion. Accordingly, plaintiffs have not satisfied Rule 23(a)'s commonality requirements, and plaintiffs' proposed classes cannot be certified.

1. Dr. Finkelman's Testimony Is Unreliable And Therefore Does Not Support Plaintiffs' Theory of Commonality

Plaintiffs assert that the opinions of their industrial-organizational psychology expert, Dr. Jay Finkelman, support their contention that Amtrak's human-resources practices made the company's employment decisions vulnerable to bias. Pls.' Class Cert. Mem., ECF No. 303 at 17-18, 26. Dr. Finkelman opines that individual managers departed from Amtrak's uniform hiring, promotion, and disciplinary policies in a way that "allowed for subjectivity and the potential for bias or discrimination." Finkelman Dep., ECF No. 319-3 at 4-5; see also Finkelman Rep., ECF No. 304-3 at 25 ("There is a disturbing and pervasive randomness to the evaluation, selection and discipline procedures that Amtrak apparently uses throughout the system. There are few is any controls against intentional or inadvertent bias or discrimination. The process appears to be highly subjective and unstructured."). As previously explained, however, Dr. Finkelman's failure to verify the facts supporting his opinions render his report and testimony unreliable under the standards set forth in Daubert v. Merrell Dow Pharmaceuticals, Inc. , 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993). See supra Part III.C. Therefore, the Court will not consider his report or testimony in evaluating plaintiffs' commonality arguments.⁸

*3192. Plaintiffs' Statistical Evidence Does Nothing To Establish That Amtrak's Employment Practices Led To Any Alleged Disparate Outcomes

Plaintiffs next point to statistical evidence to show that Amtrak's facially-neutral employment policies resulted in a disparate racial impact. Pls.' Class Cert. Mem., ECF No. 303 at 18-19, 26, 30. Specifically, plaintiffs' statistical experts, Dr. Bradley and Dr. Fox, found that African-American individuals were hired and promoted for vacant positions at rates lower than their non-African-American counterparts and were disciplined at rates higher than their non-African-American counterparts. See generally Bradley/Fox Rep., ECF No. 304-1.

Statistical evidence may, of course, be used to prove discrimination on a disparate-impact theory. See, e.g. , Watson v. Fort Worth Bank and Trust , 487 U.S. 977, 991-1000, 108 S.Ct. 2777, 101 L.Ed.2d 827 (1988). Dukes did not change this standard, but rather reiterated that statistical correlation cannot substitute for a specific finding of class-action commonality. See Dukes , 564 U.S. at 356, 131 S.Ct. 2541. In other words, "merely proving that the discretionary system has produced a racial ... disparity is not enough " where plaintiffs are unable to identify a specific employment practice that is responsible for the alleged disparity. Id. This is particularly true when, as here, the challenged employment practices combine both objective and subjective components. See Watson , 487 U.S. at 994, 108 S.Ct. 2777 ("Especially in cases where an employer combines subjective criteria with the use of more rigid standardized rules or tests, the plaintiff is in our view responsible for isolating and identifying the specific employment practices that are allegedly responsible for any observed statistical disparities.").

Here, the critical question is whether Dr. Bradley and Dr. Fox's statistical analysis shows that subjective employment decisions at Amtrak led to racially disparate outcomes. Dr. Bradley admitted, however, that the statistical analysis did not focus on any particular employment practice, and he acknowledged that he therefore cannot opine that a particular employment practice caused any alleged racial disparities:

Q. Dr. Bradley, I'd like to ask whether or not you can give a professional statistical opinion or do you give a professional statistical opinion in your report that a particular employment practice at Amtrak caused adverse impact against African-Americans?

A. I cannot.

Q. Did you study whether a particular employment practice at Amtrak caused adverse impact?

A. I did not.

Bradley Dep., ECF NO. 331-3 at 27.

Dr. Bradley reiterated this conclusion later in his testimony:

Q. If we start at a global level, how do I know that the employees that I'm looking at ... were affected by some type of *320criteria that had adverse impact on them that was the same criteria?

A. Well, it may not be the same criteria. You get differences I think like we do across all of these jobs in a particular craft. And blacks are showing a smaller rate and it is statistically significant, that indicates to me there's some problem somewhere and we need to investigate where that problem is.

Id. at 32-33. In other words, plaintiffs' statistical experts do little more than establish that African-American candidates are underrepresented in Amtrak's hiring and promotion decisions, and overrepresented in Amtrak's disciplinary decisions. This is precisely the sort of statistical evidence that was rejected as insufficient in Dukes . See Dukes , 564 U.S. at 356-57, 131 S.Ct. 2541.

Moreover, in conducting their analysis, Dr. Bradley and Dr. Fox examined employment decisions across four craft groups, each of which contain numerous positions with different responsibilities, that are overseen by different supervisors, that are in different locations, and that are covered by different labor unions. See Expert Rep. of Jerrold A. Glass, ECF No. 320-4 ¶¶ 14-24. When asked about his approach, Dr. Bradley conceded that his aggregated analysis would not permit any conclusions about the potential causes of any racially-disparate impact seen in the statistical analysis:

Q. Wouldn't you want to ... try to find the jobs that are similar to each other and aggregate them?

A. At some point once you you've got adverse impact, you want to try and drill down and find out where the problems are occurring.

Q. Did you do that in your study, try to drill down?

A. I have not done that.

Q. Why not?

A. I wasn't asked to do that in this particular case.

Bradley Dep., ECF No. 331-3 at 32. Other courts have rejected Dr. Bradley's expert opinions for similar reasons. See Anderson v. Westinghouse Savannah River Co. , 406 F.3d 248, 262-63 (4th Cir. 2005) (district court did not err in excluding Dr. Bradley's opinions because his statistical analysis did not compare similarly-situated employees and therefore was not probative of whether or not there was a disparate impact). Indeed, Amtrak's statistical expert-whose qualifications and opinions plaintiffs do not challenge-found that there was no consistent pattern of adverse outcomes for African-American individuals when decisions were analyzed based on job-specific selection criteria. See Expert Rep. of Donald Deere ("Deere Rep."), ECF No. 328-5 at 28-32.

Nor did Dr. Bradley and Dr. Fox consider objective factors like seniority, previous work experience, or education in examining Amtrak's hiring and promotion decisions. See Bradley Dep., ECF No. 331-3 at 27-29. For example, when asked whether his analysis took into consideration a particular individual's work experience in assessing whether the selection process had an adverse impact, Dr. Bradly admitted that it did not. Accordingly, he acknowledged that it was "possible" that his findings of disparate impact could be explained by a wholly "legitimate factor" that played "a decisive role" in the decision-maker's employment selection. Id. at 28. For this reason, too, plaintiffs' statistical evidence does not demonstrate commonality. See Garcia v. Johanns , 444 F.3d 625, 635 (D.C. Cir. 2006) (district court acted within its discretion in rejecting statistical analysis where the expert "failed to account for variables that affected the analyses" and therefore did not connect any alleged disparate impact to defendant's policy or practice);

*321Gonzalez v. Brady , 136 F.R.D. 329, 333 (D.D.C. 1991) (because "plaintiffs' statistics merely compared the relative number of Hispanics and non-Hispanics at the various grade levels" and did not "show the comparison between similarly situated Hispanic and non-Hispanic employees (i.e., employees with similar qualifications and experience)," they "offer[ed] little assistance in establishing the existence of the aggrieved class").

3. Plaintiffs' Anecdotal Evidence Shows Variability, Not Commonality

Plaintiffs also offer anecdotal evidence-in the form of declarations from 101 putative class members-in support of their contention that a common mode of discretionary decisionmaking resulted in racial discrimination across each alleged subclass. Pls.' Class Cert. Mem., ECF No. 303 at 19-22, 31; id. Ex. 8, ECF No. 304-8.

To be sure, these declarations provide far too many examples of very serious racial discrimination. For instance:

• Bryant Cox states that a white manager called him a "nigger" on multiple occasions, including in front of an Amtrak EEO Officer and a supervisor. Decl. of Bryant Cox, ECF No. 304-8 at 39, ¶ 11.

• Windell Greene explains that, in addition to being subjected to racial epithets, he once found "a rope noose about 6-7 inches in diameter hanging from a beam" in a common work area. Decl. of Windell Greene, ECF No. 304-8 at 80, ¶ 27.

• Betty Haymer states that, when she and other African-American employees objected to being assigned maintenance work outside on a rainy day, the supervisor yelled at them and called them a "[b]unch of niggers." Decl. of Betty Haymer, ECF No. 304-8 at 81, ¶ 17.

• Lena Johnson recounts that, "[b]etween 1980 and 1995, [her] white Supervisor, Bill Lake, would use the word 'nigger' on a daily base. For example, he would say to an African-American employee, 'Nigger, get that machine started.' " Decl. of Lena Johnson, ECF No. 304-8 at 126, ¶ 6.

• Alfred Felton recalls that, "[i]n about 1998, white employees hung a black doll from a noose in the locker room. Later, the doll was taken down and attached to the back of a golf cart. [He] also observed a white employee dragging a black doll behind his scooter. On several occasions, [he] heard white employees threaten to drag "Niggers" behind the trains when they departed." Decl. of Alfred Felton, ECF No. 304-8 at 451, ¶ 16.⁹

The use of deeply offensive racial epithets by Amtrak supervisors almost certainly created a hostile work environment for the individual employees subject to the abuse. See Ayissi-Etoh v. Fannie Mae , 712 F.3d 572, 577 (D.C. Cir. 2013) (suggesting that "the use of an unambiguously racial epithet such as 'nigger' by a supervisor" could alone be sufficient to establish a hostile work environment) (citation omitted).

The declarations do not, however, offer evidence that Amtrak's supervisors exercised their discretion in a uniform manner. For one, the declarations demonstrate that, although some supervisors exercised discretion in a discriminatory manner, others did not. See, e.g , Decl. of Garner Willis, ECF No. 304-8 at 297, ¶¶ 7-8 (attesting that one white manager "screamed at [him] .... using racially charged language"

*322but that his white supervisor removed the manager from the crew and sent him to training upon learning of the behavior). Moreover, the declarations establish that members of plaintiffs' putative subclasses experienced discrimination based on a number of different policies or practices. For example, plaintiffs submit four declarations from Amtrak employees in the engineering craft groups. As evidence of racial discrimination at Amtrak, one of those employees, Marcus Brunswick, points to the fact that he and another African-American applicant failed the "subjective visual" portion of a test required for an electrical traction position, whereas several white applicants took the test and passed. Decl. of Marcus Brunswick, ECF No. 304-8 at 282, ¶ 6.¹⁰ Mr. Brunswick also claims that it took him longer to be promoted than his white peers, who "were promoted faster because of their race and because they received training and mentorship from white supervisors." Id. ¶ 12. Another employee in the engineering craft group, Alfred Jones, states in his declaration that a supervisor retaliated against him after he complained to his union representative about the supervisor's racist comments. Decl. of Albert Jones, ECF No. 304-8 at 290, ¶¶ 6-8.¹¹ Specifically, Mr. Jones avers that his supervisor denied him higher pay rates, asked him to complete additional work, and eventually terminated him based on an insubordination charge. Id. ¶¶ 8-11. Although Mr. Jones was eventually reinstated after filing a claim for race discrimination, he continued to experience retaliatory conduct in the form of lower overtime pay, being forced to bid into lower-paying positions, and unfavorable job assignments that should have been given to less-senior white employees. Id. ¶¶ 14-18. As the declarations of Mr. Brunswick and Mr. Jones demonstrate, even employees in the same craft groups experienced discrimination in different ways at the hands of different individuals. Accordingly, the declarations suffer from the same defects as plaintiffs' statistical evidence-i.e., they do not establish that the putative class members are "victim[s] of one common discriminatory practice." Dukes , 564 U.S. at 345, 131 S.Ct. 2541.

4. Ms. Hightower's Testimony Is Insufficient To Tie Together The Claims Of Classes Spanning Sixteen Years

Finally, plaintiffs argue that the testimony of Wanda Hightower-Amtrak's Vice President for Diversity from 1999 to 2001-provides compelling evidence that supervisors received "signals from top management that they did not have to submit to investigations of discrimination and harassment or implement recommendations remedial action." Pls.' Class Cert. Mem., ECF No. 303 at 29. Ms. Hightower testified that her efforts to ensure appropriate consequences were meted out for egregious instances of racial discrimination were met with resistance, that Amtrak's Chief Executive Officer George Warrington asked her to "slow the pace down" and suggested she stop making "aggressive *323recommendations" with respect to discipline, and that she was abruptly fired when she refused to comply with those suggestions. See Hightower Dep., ECF No. 309-9 at 11-15, 18, 34-35.

Ms. Hightower's testimony provides forceful evidence that, at least for the period during which Ms. Hightower was employed at Amtrak, Amtrak's leadership-and, in particular, Mr. Warrington-did not support Amtrak's corporate policy prohibiting racial discrimination. As compelling as Ms. Hightower's testimony is in this regard, it nonetheless fails to tie together the many discretionary employment decisions to which the putative class members were subject over the class period. For one, the testimony of one former employee who worked at Amtrak for less than two years does not reasonably raise an inference that Amtrak "operated under a general policy of discrimination" over the entire sixteen-year class period. Moreover, even for the period during which Ms. Hightower was employed at Amtrak, Ms. Hightower's testimony does not show that the disciplinary decisions of most supervisors at Amtrak were discriminatory. According to plaintiffs' experts, there were 24,136 charges for disciplinary charges over the 152-month period that was analyzed. See Bradley/Fox Rep., ECF No. 304-1 at 29. Accordingly, during Ms. Hightower's 22-month tenure, approximately 3,500 disciplinary charges would have been brought on average. Ms. Hightower testified, however, that there were only twenty-five cases during her time at Amtrak where she "felt that supervisors or managers had engaged in some kind of discriminatory or retaliatory activity" that "was not adequately addressed." Hightower Dep., ECF No. 309-9 at 33. Thus, Ms. Hightower's testimony does not provide "significant proof" that Amtrak operated under a general policy of discrimination. Dukes , 564 U.S. at 355, 131 S.Ct. 2541.

The cases cited by plaintiffs in which courts have certified employment-discrimination classes confirm the Court's conclusions. Each of those cases involved more tightly-knit classes and concrete theories of discrimination based on common employment practices or the decisions of a common supervisor. For example, plaintiffs point to McReynolds v. Merrill Lynch, Pierce, Fenner & Smith, Inc. , 672 F.3d 482 (7th Cir. 2012), in which the Seventh Circuit affirmed certification of a class of 700 financial brokers who alleged that their employer's policies had a disparate impact on African-American employees. See Pls.' Reply in Supp. of Mot. for Class Cert., ECF No. 344 at 9. Although the plaintiffs in that case also accused the defendant of "delegate[ing] discretion over decisions that influence the compensation of all the company's 15,000 brokers," plaintiffs pointed to two specific employment policies that they claimed led to the discriminatory impact. McReynolds , 672 F.3d at 488. These company-wide policies provided the "glue" that held the plaintiffs' claims together because they purportedly explained how the directors exercised their discretion in a common way that had a discriminatory impact. Id. at 488-89.

Likewise, in Moore v. Napolitano , 926 F.Supp.2d 8 (D.D.C. 2013), the court certified a class of current and former special agents in the United States Secret Service alleging discrimination in the Secret Service's promotion practices. See Pls.' Suppl. Mem. in Supp. of Class Cert., ECF No. 370 at 11. The plaintiffs there challenged the Secret Service's use of the Merit Promotion Program, which produced a numerical score for each candidate that was used to make promotion decisions. Moore , 926 F.Supp.2d at 12-13. Here, by contrast, plaintiffs have not identified any selection policy that applies to all putative class members across various job functions.

*324In short, plaintiffs have not identified a specific employment practice applicable to all putative class members that purportedly caused the alleged discrimination about which plaintiffs complain. Moreover, plaintiffs' evidence makes clear that many putative class members suffered discrimination in a variety of ways through the decisions of different individuals in a wide range of contexts. "Such potential breadth of experiences and claims among the putative class members is not the mark of a class that meets the commonality requirement of Rule 23(a)." Ross v. Lockheed Martin Corp. , 267 F.Supp.3d 174, 201 (D.D.C. 2017).

VI. SUMMARY JUDGMENT

Although plaintiffs' claims are not amenable to class treatment, the individual claims of the named plaintiffs survive. The Court therefore proceeds to consider Amtrak's motion for partial summary judgment on plaintiffs' disparate-impact claims.

Summary judgment is appropriate when the moving party has shown that there are no genuine issues of material fact and that the moving party is entitled to judgment as a matter of law. Fed. R. Civ. P. 56 ; Celotex Corp. v. Catrett , 477 U.S. 317, 322, 106 S.Ct. 2548, 91 L.Ed.2d 265 (1986). A material fact is one that is capable of affecting the outcome of the litigation. Anderson v. Liberty Lobby, Inc. , 477 U.S. 242, 248, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986). A genuine issue exists where the "evidence is such that a reasonable jury could return a verdict for the nonmoving party." Id. A court considering a motion for summary judgment must draw all "justifiable inferences" from the evidence in favor of the non-movant. Id. at 255, 106 S.Ct. 2505.

To survive a motion for summary judgment, however, the non-movant "must do more than simply show that there is some metaphysical doubt as to the material facts"; instead, the nonmoving party must come forward with " 'specific facts showing that there is a genuine issue for trial.' " Matsushita Elec. Indus. Co. v. Zenith Radio Corp. , 475 U.S. 574, 586-87, 106 S.Ct. 1348, 89 L.Ed.2d 538 (1986) (quoting Fed. R. Civ. P. 56(e) ). Moreover, "although summary judgment must be approached with special caution in discrimination cases, a plaintiff is not relieved of his obligation to support his allegations by affidavits or other competent evidence showing that there is a genuine issue for trial." Adair v. Solis , 742 F.Supp.2d 40, 50 (D.D.C. 2010) (quotation marks and alterations omitted).

Employment discrimination claims under Title VII may proceed under both a "disparate treatment" and a "disparate impact" theory. See Ross v. Lockheed Martin Corp. , 267 F.Supp.3d 174 (D.D.C. 2017) (citation omitted). A plaintiff alleging disparate impact must show that an employer uses "a particular employment practice that causes a disparate impact on the basis of race, color, religion, sex, or national origin." Ricci v. DeStefano , 557 U.S. 557, 578, 129 S.Ct. 2658, 174 L.Ed.2d 490 (2009) ; see also 42 U.S.C. § 2000e-2(k)(1)(A) (plaintiff alleging disparate impact must demonstrate that the employer "uses a particular employment practice that causes a disparate impact on the basis of race," at which point the employer must show "that the challenged practice is job related for the position in question and consistent with business necessity" or adopt an "alternative practice" that has less disparate impact but still meets the employer's needs). Although a plaintiff generally must identify a specific employment practice that is the subject of the challenge, if the plaintiff "can demonstrate to the court that the elements of a[n] [employer's] decisionmaking process are not capable of separation for analysis, the *325decisionmaking process may be analyzed as one employment practice." 42 U.S.C. § 2000e-2(k)(1)(B)(i).

Amtrak first asserts that it should be granted summary judgment on plaintiffs' disparate-impact claims because plaintiffs have not identified any specific employment practice in the fourth amended complaint or motion for class certification that qualifies as a "particular employment practice." Def.'s Mem. in Supp. of Mot. for Summ. J. ("Def.'s Summ. J. Mem."), ECF No. 328-1 at 11-14. In response, plaintiffs state that they "challenge the selection interview process, ratings, rank-orderings, input from other managers, amorphous decision-making, and the disqualifying discipline criterion" as having an adverse impact on African-American employees and applicants. Pls.' Opp. to Mot. for Summ. J. ("Pls.' Summ. J. Opp."), ECF No. 343 at 7. Plaintiffs claim that, aside from the disqualifying discipline criteria, these practices are components of Amtrak's "overall selection process" and cannot be separated for two reasons: (1) the practices are so "interwoven" that no single practice is "determinative" of the disparate outcomes; and (2) Amtrak failed to keep adequate records to permit analysis of particular practices. Pls.' Summ. J. Opp., ECF No. 343 at 8.

Amtrak asserts that both these arguments fail. First, Amtrak contends that the "incapable-of-separation" exception applies only "where common components of a uniform selection process exist, but it is unreasonably difficult to isolate those common components from each other." Def.'s Reply in Supp. of Mot. for Summ. J. Reply ("Def.'s Summ. J. Reply"), ECF No. 356 at 11. Amtrak further argues that plaintiffs have not met their burden to show that the incapable-of-separation exception should apply based on the paucity of data. Amtrak points to plaintiffs' own arguments that Amtrak's selection process is comprised of five distinct steps, which suggests that Amtrak's selection process was, in fact, capable of separation. Def.'s Summ. J. Mem., ECF No. 328-1 at 21-22.

Taking Amtrak's second argument first, the Court agrees that plaintiffs have not sufficiently shown that the various components of Amtrak's process for hiring or promoting employees were not capable of being divided into smaller subsets related to specific employment practices for purposes of a statistical analysis. As an initial matter, although the parties used a joint database to avoid an "intractable debate" over how to merge Amtrak's various sources of employment data, there does not appear to be any serious dispute that the database alone did not contain sufficient information to engage in a statistical analysis of specific employment policies for disparate impact. See Pls.' Summ. J. Opp., ECF No. 343 at 13-14; Def.'s Summ. J. Reply, ECF No. 356 at 16-17.¹² Even so, Amtrak maintained "job files" that contained documents spanning the various stages of the employment selection process for each individual. Pls.' Summ. J. Opp., ECF No. 343 at 15; Deere Rep., ECF No. 331-5 at 21. Some of these files include *326applicant questionnaires, pre-employment tests and surveys, interviewer report forms, and interview questions. Deere Rep., ECF No. 331-5 at 21. As such, the data from these job files could have been used to evaluate the effect of a specific practice-for example, whether the use of a pre-employment test or ratings forms increased the likelihood that an African-American individual would not be selected for the position.

Plaintiffs respond that these job files could not be used to analyze different employment practices because the "contents of each file were inconsistent and varied." Pls.' Summ. J. Opp., ECF No. 343 at 16 (providing examples of the inconsistency in documents contained in each file). Nonetheless, plaintiffs have not sufficiently shown that this information could not have meaningfully been used to evaluate different employment practices. For example, plaintiffs do not offer any testimony from their statistical experts that the files do not contain adequate data to conduct a reliable analysis. Indeed, plaintiffs' statistical expert acknowledged that the racial identity of thirty-five percent of the applicants in the applicant flow data was "unknown," but argued that, from a methodological perspective, the missing information was not an insurmountable obstacle to his analysis. See Expert Rebuttal Rep. of Edwin L. Bradley, ECF No. 342-6 at 18. Dr. Bradley did not provide any similar analysis or opinion with respect to the purportedly missing job file data. Accordingly, the Court finds that plaintiffs have failed to meet their burden to show that Amtrak's selection procedures are "not capable of separation for analysis."

Likewise, plaintiffs have not identified any specific disciplinary practices that they claim led to a disparate impact. Indeed, Dr. Bradley conceded that he did not attempt to study particular forms of discipline used at Amtrak:

Q. So you can't say anything based on your discipline study about what might have caused the disparate impact in the award of disciplines to African-Americans?

A. That's correct.

Q. And you didn't study disciplinary-any particular disciplinary infraction to see if maybe that type of infraction had a discriminatory impact on African-Americans, did you?

A. What do you mean by that?

Q. Well, you took all charges no matter what kind of charge. You didn't look at, well, this charge involves absenteeism or this charge involves tardiness or this charge involves insubordination? You didn't look at the different types of charges, did you?

A. No, I did not.

Q. You had the data to look at those different types of charges, didn't you?

A. They did show the different types of charges, that's correct.

Q. Why didn't you look at the different types of charges?

A. I was interested in the disciplinary process as a whole.

Bradley Dep., ECF No. 331-3 at 65. To survive summary judgment, plaintiffs are "responsible for isolating and identifying the specific employment practices that are allegedly responsible for any observed statistical disparities." Smith v. City of Jackson, Miss. , 544 U.S. 228, 241, 125 S.Ct. 1536, 161 L.Ed.2d 410 (2005). This is because a "failure to identify the specific practice being challenged is the sort of omission that could result in employers being potentially liable for the myriad of innocent causes that may lead to statistical imbalances." Id. (citation and internal quotation marks omitted). Here, plaintiffs have failed to identify any specific practice, and therefore, their disparate-impact *327claims must fail. For all of these reasons, Amtrak's motion for summary judgment on plaintiffs' disparate-impact claims is granted.¹³

VII. CONCLUSION

For the foregoing reasons, the Court GRANTS Amtrak's motion to exclude Jay Finkelman, DENIES Amtrak's motion to exclude Thomas Roth, DENIES Amtrak's motion to exclude Edwin Bradley and Liesl Fox, GRANTS in part Amtrak's motion to strike portions of plaintiffs' declarations, GRANTS in part and DENIES in part Amtrak's motion to strike plaintiffs' reply brief, DENIES plaintiffs' motion for class certification, and GRANTS defendant's partial motion for summary judgment. Plaintiffs' class claims are dismissed, as are plaintiffs' disparate-impact claims under Title VII and Section 1981. An appropriate Order accompanies this Memorandum Opinion.

SO ORDERED.

1 When citing to the electronic filings in this opinion, the Court cites to the ECF page numbers, not the page number of the filed document.

2 Prior to the dissolution of the DRO, Amtrak's EEO Compliance Unit only handled complaints by employees that were filed with a federal, state, or local agency, along with any internal complaints in which an employee was represented by counsel. Marcelle Decl., ECF No. 322-2 ¶ 10.

3 On September 14, 2017, a petition for writ of certiorari was filed presenting this precise question-to wit, "[w]hether a district court may certify a class action based on information that does not meet the standards of admissibility set forth in the Federal Rules of Evidence and Civil Procedure." See Notice of Filing of Petition for a Writ of Certiorari, Taylor Farms Pac., Inc. v. Pena , No. 15-15965 (9th Cir. Sept. 14, 2017), ECF No. 57 Ex. A at 2. The Supreme Court denied the petition for writ of certiorari on February 20, 2018. See Taylor Farms Pacific, Inc. v. Pena , --- U.S. ----, 138 S.Ct. 976, 200 L.Ed.2d 300 (2018).

4 Although this holding conflicts with the D.C. Circuit's unpublished decision in In re Rand , No. 02-8007, 2002 WL 1461810, at *1 (D.C. Cir. July 8, 2002), the Court concludes that it is not bound by that decision. For one, unpublished decisions " 'should not strictly bind panels' of the court of appeals and are often not 'suitable for governing future cases' given that they neither reach the merits nor benefit from oral argument." Martin v. Dist. of Columbia , 78 F.Supp.3d 279, 308 n.36 (D.D.C. 2015) (quoting In re Grant , 635 F.3d 1227, 1232 (D.C. Cir. 2011) ). The persuasiveness of such a disposition is further undermined where, as here, subsequent doctrinal changes have significantly altered the landscape. See id. (noting as an additional reason not to be bound by an unpublished disposition the fact that "the tides have changed in the last seventeen years" regarding the legal doctrine at issue).

5 Because Amtrak filed both an opposition to plaintiffs' motion for class certification and its motion for summary judgment on the same day, it bears noting that "the order of disposition of motions for summary judgement and class certification" is "a question of discretion for the trial court." Curtin v. United Airlines, Inc. , 275 F.3d 88, 92 (D.C. Cir. 2001). Rule 23(c)(1)(A) directs the court to determine "[a]t an early practicable time" whether to certify a class action, and "it is often more efficient and fairer to the parties to decide the class question first,"Curtin , 275 F.3d at 92 ; see also Am. Pipe & Const. Co. v. Utah , 414 U.S. 538, 547-49, 94 S.Ct. 756, 38 L.Ed.2d 713 (1974) (noting that the 1966 amendment to Rule 23 requiring a court to decide the class-certification question "as soon as practicable" was designed, in part, to curtail the abusive practice of one-way intervention). The advisory committee notes to Rule 23 recognize, however, that a decision on summary judgment may be appropriate prior to a certification ruling in certain circumstances. See Fed. R. Civ. P. 23 advisory committee notes ("Other considerations may affect the timing of the certification decision. The party opposing the class may prefer to win dismissal or summary judgment as to the individual plaintiffs without certification and without binding the class that might have been certified."). Here, the Court will follow the ordinary course and will first address plaintiffs' motion for a class certification before addressing the legal merits of plaintiffs' claims under section 1981 and Title VII.

6 Plaintiffs point to the testimony of one deponent to argue that the ultimate decision to hire or promote an employee rested exclusively with Amtrak's Chief Executive Officer. See Pls.' Class Cert. Mem., ECF No. 303 at 16 (citing Walker Dep., ECF No. 309-4 at 4). Declarations from other employees, however, suggest that individual hiring decisions are made by local managers based on a host of factors. See Ray Decl., ECF No. 322-5 ¶ 44 ("the Hiring Manager, sometimes with input from his or her department, is responsible for selecting the final candidate for hire"); Allan Decl., ECF No. 321-3 ¶ 17 (explaining that the "ultimate decision making authority" on a hiring decision "lies with the Hiring Manager"). Accordingly, plaintiffs have not established that only upper-level management were involved in many or all the challenged employment decisions. See Scott v. Family Dollar Stores, Inc. , 733 F.3d 105, 114 (4th Cir. 2013) (explaining that Dukes is "limited to the exercise of discretion by lower-level employees, as opposed to upper-level, top-management personnel").

7 Amtrak moves to strike certain new arguments raised by plaintiffs in their reply memorandum in support of their motion for class certification. See Def.'s Mem. in Supp. of Mot. to Strike Reply, ECF No. 353-1 at 4-18. Amtrak further moves to strike portions of the declarations submitted in support of plaintiffs' opposition to Amtrak's motion to exclude Drs. Bradley and Fox as inadmissible under the Federal Rules of Evidence. See id. at 18-31. Taking the second argument first, the Court agrees with Amtrak that the Court cannot rely on plaintiffs' declarations without resolving Amtrak's evidentiary objections. See supra Part IV . Because the Court does not rely on the portions of the declarations to which Amtrak objects, the Court does not address Amtrak's objections at this time. As for Amtrak's first argument, the Court finds that, to the extent plaintiffs' raised any purportedly new arguments in their reply, Amtrak sufficiently addressed those arguments in its motion to strike. Accordingly, because Amtrak will not be prejudiced by the Court's consideration of plaintiffs' reply memorandum, it denies Amtrak's motion to strike any "new" arguments.

8 In any event, Dr. Finkelman's report and testimony do not show that managers or others at Amtrak applied their discretion in any common way that caused racial disparities. Indeed, he does not opine that Amtrak's practices actually resulted in any biased employment decisions; rather, he opines that Amtrak's processes created a "potential" for bias. Finkelman Dep., ECF No. 319-3 at 4-5. As Amtrak correctly notes, this sort of testimony is similar to the testimony offered by the plaintiffs' expert in Dukes , who opined that Wal-Mart's corporate culture made it "vulnerable" to bias but "could not calculate whether 0.5 percent or 95 percent of the employment decisions ... might be determined by stereotyped thinking." Dukes , 564 U.S. at 354, 131 S.Ct. 2541. The Supreme Court found that this testimony could be "safely disregard[ed]" and was "worlds away from 'significant proof' that Wal-Mart 'operated under a general policy of discrimination.' " Id.

9 Although Amtrak objects to portions of Mr. Felton's declaration, it does not contest the paragraphs cited by the Court. See Def.'s Mem. in Supp. of Mot. to Strike Decls. of Pls. ("Def.'s Strike Decls. Mem."), ECF No. 330-1 at 18-20 (objecting to paragraphs 7, 10, and 17).

10 Although Amtrak objects to portions of Mr. Brunswick's declaration, it does not contest the paragraphs cited by the Court. See Def.'s Strike Decls. Mem., ECF No. 330-1 at 11-12 (objecting to paragraphs 9, 10, 20, and 23).

11 Amtrak objects that Mr. Jones' declaration, which states that he was terminated as a result of an insubordination charges, is directly contradicted by his deposition testimony in which he explains that he was "taken out of service" as opposed to terminated. See Def.'s Strike Decls. Mem., ECF No. 330-1 at 29. The Court finds that this inconsistency does not relate to a material fact in the case and goes to the weight to be afforded Mr. Jones' testimony. Accordingly, the Court will not strike the testimony at issue. See Ascom Hasler Mailing Sys. v. U.S. Postal Serv. , 815 F.Supp.2d 148, 163 (D.D.C. 2011).

12 Amtrak argues that that the joint database contained sufficient data for plaintiffs to conduct a statistical analysis with respect to a "particular job in a particular location where the selection criteria and procedures were more likely to be similar." Def.'s Summ. J. Reply. ECF No. 356 at 20-23. The fact that the data in the joint database could be sliced to analyze particular jobs by particular cities does not, however, mean that plaintiffs cannot proceed under the incapable-of-separation exception. The key question "is not whether the massive data can be divided up into piles," but rather, "whether the plaintiffs demonstrated any resulting piles that might be formed do not reveal particular employment practices that are capable of separation for statistical analysis." Pippen v. State , 854 N.W.2d 1, 25 (Iowa 2014).

13 Amtrak also moves for summary judgment on the ground that plaintiffs' disparate-impact claims are not cognizable under 42 U.S.C. § 1981. See Def.'s Summ. J. Mem., ECF No. 328-1 at 10; see also Gen. Bldg. Contractors Ass'n v. Penn. , 458 U.S. 375, 391, 102 S.Ct. 3141, 73 L.Ed.2d 835 (1982) (section 1981"can be violated only by purposeful discrimination"); Frazier v. Consol. Rail. Corp. , 851 F.2d 1447, 1449 n.3 (D.C. Cir. 1988) (similar); McReynolds v. Sodexho Marriott Servs., Inc. , 349 F.Supp.2d 1, 7 n.3 (D.D.C. 2004) ("Defendant correctly argues that plaintiffs cannot bring a disparate impact claim under 42 U.S.C. § 1981, since purposeful discrimination is required under § 1981."). Plaintiffs do not offer any response to this argument, and thus concede it. In any event, plaintiffs' fourth amended complaint does not allege disparate impact with respect to plaintiffs' section 1981 claims. See Fourth Am. Compl., ECF No. 145 ¶ 619 (alleging that Amtrak's conduct has been "intentional, deliberate, willful, and conducted in callous disregard of the rights of the named Plaintiffs").