Chen-Oster v. Goldman, Sachs & Co.

Full title: H. Cristina CHEN-OSTER, Lisa Parisi, Shanna Orlich, Allison Gamba, and…

Court: United States District Court, S.D. New York

Date published: Mar 30, 2018

325 F.R.D. 55 (S.D.N.Y. 2018)

Opinion

[Copyrighted Material Omitted] [Copyrighted Material Omitted] [Copyrighted Material Omitted] [Copyrighted Material Omitted] [Copyrighted Material Omitted]

          Adam T. Klein, Christopher McNerney, Cara Elizabeth Greene, Carmelyn Pingol Malalis, Daniel S. Stromberg, Justin Mitchell Swartz, Ossai Miazad, Melissa Lardo Stewart, Outten & Golden, LLP, Rachel Geman, Lieff Cabraser Heimann & Bernstein, LLP, New York, NY, Alison M. Stocking, Anne B. Shaver, Heather H. Wong, Kelly Dermody, Michelle Lamy, Tiseme Gabriella Zegeye, Lieff Cabraser Heimann & Bernstein, LLP, San Francisco, CA, Dana Gale Sussman, Safe Horizon Anti-Trafficking Program, Brooklyn, NY, Paul W. Mollica, Outten & Golden LLP, Chicago, IL, for Plaintiffs.

         Neal D. Mollen, Paul Hastings LLP, Washingon, DC, Theodore Otto Rogers, Jr., Beth Deborah Newton, Margaret Elizabeth Bartlett, Robert Joseph Giuffra, Jr., Robin D. Fessel, Suhana S. Han, Ann-Elizabeth Ostrager, Michael Peter Reis, Joshua Seth Levy, Sullivan and Cromwell, LLP, New York, NY, Barbara B. Brown, Carson H. Sullivan, Paul Hastings LLP, Jeffrey B. Wall, Sullivan & Cromwell LLP, Washington, DC, for Defendants.

          OPINION AND ORDER

         ANALISA TORRES, United States District Judge

          On September 16, 2010, Plaintiffs, a group of women who worked at Goldman, Sachs & Co. and The Goldman Sachs Group, Inc. (collectively, "Goldman Sachs" or "Defendants"), filed this putative class action alleging gender discrimination under Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e et seq., and the New York City Human Rights Law ("NYCHRL"), N.Y.C. Admin. Code § 8-101 et seq. On March 10, 2015, the Honorable James C. Francis issued a Memorandum and Order ("M & O") excluding some, but not all, expert testimony submitted in connection with Plaintiffs’ motion for class certification. Francis M & O, ECF No. 363. On the same day, Judge Francis issued a Report and Recommendation ("R & R") recommending that the Court deny Plaintiffs’ motion. Francis R & R, ECF No. 364. The Court held in abeyance briefing on objections to Judge Francis’ M & O and R & R as to Rule 23(b)(3) class certification while it addressed other motions. See ECF Nos. 370, 437, 454. On April 28, 2017, the Court set a briefing schedule for objections to the M & O and R & R, see ECF No. 488, and briefs were fully submitted on August 1, 2017, see ECF Nos. 522, 531-35. For the reasons stated below, the parties’ objections on expert testimony are OVERRULED. Objections on class certification are SUSTAINED in part and OVERRULED in part. Accordingly, Plaintiffs’ motion to certify the class pursuant to Rule 23(b)(3) is GRANTED in part and DENIED in part.

          BACKGROUND

          I. Individual Plaintiffs

          A. H. Christina Chen-Oster

          Plaintiff H. Chen-Oster joined Goldman Sachs in March 1997. She worked as a salesperson in the firm’s Securities Division— first in the Convertible Sales Business Unit and later in U.S. Research Sales. Francis R & R, at 2. She held the title of Vice President from June 1997 until she resigned in March 2005. Chen-Oster Decl. ¶ 3, ECF No. 253. Chen-Oster claims she that was constructively discharged as a result of systemic discrimination at the firm, id. ¶¶ 3, 13, and that, because of her gender, she was denied a fair performance review, denied fair pay, and denied promotion to Managing Director, id. ¶ 4. At the time of the R & R, Chen-Oster still worked in the financial services industry as a Managing Director for Deutsche Bank. Id. ¶ 16; Class Cert. Hr’g Tr. I, at 113, ECF No. 343; Francis R & R, at 3.

          B. Shanna Orlich

          Plaintiff Shanna Orlich began working at Goldman Sachs in 2006 as a summer associate, where she rotated through various trading departments. Orlich Decl. ¶ 2, ECF No. 255. Orlich returned to the firm in July 2007, working as an Associate in the Securities Division. Id. ¶ 3. In November 2008, she was laid off. Id. at ¶ 4; Francis R & R, at 3. Orlich contends that during her time at the firm she was subjected to an unfair performance review and lower pay because of her gender. Orlich Decl. ¶¶ 4, 6-8. She also reports a culture that is "hostile to women." Id. ¶ 9. She cites examples such as her male Managing Director hiring scantily clad female escorts to attend a holiday party, and a Vice President inviting every man in her department to attend a golf game, but not Orlich, despite Orlich having played varsity golf in high school. Id.           C. Allison Gamba

          Plaintiff Allison Gamba joined Goldman Sachs as a trader in 2001 when it acquired her then-employer Spear, Leeds & Kellogg, L.P. Gamba Decl. ¶ 3, ECF No. 254; Second Am. Compl. ¶ 16, ECF No. 411. In 2003, she began working in the Securities Exchange Equities unit, and in about 2005 or 2006, she was promoted to Vice President of that unit. Id. Gamba avers that, because of her gender, she faced discriminatory performance reviews and compensation, and that she was passed over in 2008 for a promotion to Managing Partner— despite earning the firm significant profits during her tenure as Vice President. Id. ¶¶ 5-9, 11. After she assumed control of her section in 2006, for example, stock earnings quadrupled to $4 million. Id. ¶ 11. Likewise, profits rose to $6 million in 2007 and $9.5 million in 2008. Id. Like Orlich, Gamba also reports experiencing what she calls a "boy’s club" culture at the firm, such as exclusion from events such as barbeques and golf outings organized and attended by her male colleagues. Id. ¶ 18. She left the firm in August 2014, when Goldman Sachs sold the business unit in which she worked. Second Am. Compl. ¶ 16; Answer ¶ 16, ECF No. 426.

          D. Mary De Luis

         Plaintiff Mary De Luis began working for Goldman Sachs in 2010 as a Senior Analyst in the Investment Management Division. Second Am. Compl. ¶ 137.¹ In 2012, she was promoted to the position of Associate, and later to Vice President in December 2014. Id. De Luis believes that her gender influenced unfair performance evaluations and a smaller bonus. Id. ¶ 139. For example, De Luis alleges that when she told her male supervisors that she was undercompensated compared to the market, they brought up the career of her male domestic partner, who is an oncologist. Id. ¶¶ 141-42.

1 Plaintiffs have not submitted a declaration for De Luis and she has not yet been deposed. Accordingly, the Court relies on the Second Amended Complaint’s allegations about her employment history, which do not appear to be in substantial dispute. See Answer ¶¶ 137-141. Goldman Sachs denies that De Luis’ supervisors referenced her domestic partner’s occupation. Id. ¶ 142.

          E. The Class

          Plaintiffs bring this action on behalf of a class of similarly situated current and former female Associates and Vice Presidents employed by Goldman Sachs in revenue-generating divisions: Investment Banking, Investment Management, and Securities. Id. ¶ 10 n.1. The class covers employees at Goldman Sachs and its predecessors in the United States from September 10, 2004 through the resolution of this action, id. ¶ 61, and employees working in New York City from July 7, 2002 through the resolution of this action, id. ¶ 62. Although "the precise number of female financial-services employees at the Associate[ ] and Vice President levels at Goldman Sachs" is not known, id. ¶ 64, the parties estimate the class size is between approximately 1,762 and 2,300 persons.

          Plaintiffs allege that Goldman Sachs has discriminated against class members in evaluation, compensation, and promotion. See generally Second Am. Compl. Specifically, Plaintiffs claim that Goldman Sachs (1) employs multiple facially neutral policies that have a disparate impact on women, (2) continues to employ those policies despite knowledge of their disparate impact on women, and (3) maintains a "boy’s club" culture that discriminates against women. Id.

          II. Goldman Sachs

          Goldman Sachs is a leading financial services company headquartered in New York. The named Plaintiffs have worked in three out of four of Goldman Sachs’ revenue-generating divisions: Investment Banking, Investment Management, and Securities. These divisions are ach split into two subdivisions. The Securities Division is divided into Fixed Income, Currencies and Commodities ("FICC") and Equities. Kirk Decl. ¶ 3, ECF No. 279; Levene Decl. ¶ 2, ECF No. 280. The Investment Banking Division contains Classic Banking and Financing. Francis R & R, at 5; Packer Decl. ¶ 3, ECF No. 284; Benz Decl. ¶ 3, ECF No. 266. And the Investment Management Division is divided into Private Wealth Management and Goldman Sachs Asset Management. Francis R & R, at 4-5; Taylor Decl. ¶ 3, ECF No. 291.

          Within each of these subdivisions are even more specialized business units. See Schematic of Business Units, ECF No. 264-1. For example, within the Financing subdivision are business units organized by product, such as Leveraged Finance, Structured Finance, and Equity Capital Markets. Benz Decl. ¶ 3. Business units in the Classic Banking Subdivision are organized by industry, including Industrials, Consumer Retail, and Technology/Media and Telecommunications. Id.

          Business units are not stable; they move between divisions and subdivisions and can be eliminated or consolidated. The Public Sector & Infrastructure business unit, for instance, which is currently housed in the Investment Banking Division, Benz Decl. ¶ 2, has moved from the FICC subdivision (itself housed within a different overall division, Securities), to the Classic Banking subdivision, to the Financing subdivision, id. ¶ 13. In a similar vein, in 2003, there were as many as 227 business units, but as few as 94 two years later in 2005. Farber Rebuttal ¶ 32, ECF No. 314; Class Cert. Hr’g Tr. II, at 218, ECF No. 345. Across business units, Goldman Sachs employed two systems for evaluating its employees, known as "360 review" and "manager quartiling."

          A. Performance Evaluation

          1. 360 Review

          Each year, every Goldman Sachs professional undergoes the 360 review process. At the outset, each employee performs a self-evaluation and names 8 to 12 evaluators— subordinates, bosses, peers, and internal clients— with whom he or she has recently worked. Kung Dep. at 280-81, ECF Nos. 265-11, 265-12, 265-13. An employee’s direct manager may modify the list of evaluators, typically in consultation with that employee. Id. at 281-85. The evaluators, who are pulled from both the employee’s business unit and other units, Geman Decl. ¶¶ 5-6, ECF No. 311, then assign a numerical value to a range of criteria. In one year, all Goldman Sachs professionals were evaluated on these criteria: "Technical Skills," "Communication Skills," "Judgment," "Teamwork," "Compliance," "Diversity," "Leadership," "Overall Commercial Effectiveness," and "Overall Professional Performance." Francis R & R, at 7; see also Kung Dep. at 305-07. The 360 review process was employed across each of Goldman Sachs’ divisions throughout the class period. After the 360 review process is complete, managers receive the reviews and then add their own evaluations and narrative summaries. Kung Dep. at 321.

          2. Quartiling

          The second evaluation procedure Goldman Sachs uses is "quartiling," or what Plaintiffs refer to as "forced ranking." Managers assign each professional to one of five "quartiles." The top 25% of employees go to quartile 1, the next 25% to quartile 2, the next 25% to quartile 3, the next 15% to quartile 4, and the bottom 10% to quartile 5. Kung Dep. 316-18; Francis R & R, at 8.

          Goldman Sachs instructed its managers to place employees in one of the five quartiles based on seven factors: (1) 360 review data, (2) quality of performance, (3) long-term commercial impact or contribution, (4) technical and functional expertise, (5) potential to assume increasing responsibility, (6) leadership and management skills, and (7) diversity and citizenship-related activities. Manager Performance Rating Toolkit, ECF No. 265-16; see also Guidelines for the Manager Performance Rank, ECF No. 265-17 (instructing managers to rank employees based on their "performance, contribution, and potential relative to: (1) expectations for that individual’s level of experience and position; and (2) that individual’s peers"). Generally, quartile ranking "relies heavily on the results of the 360 review process and on production numbers, where present." Ward Report, at 32, ECF Nos. 298, 298-1; see also Guidelines for the Manager Performance Rank ("Comments and ratings from 360 reviewers and any relevant mid-year comments should be a key input for your ranking decisions."). Goldman Sachs guidelines also advise that rankings "should approximate a forced distribution over the entire division, but need not be exact." Guidelines for the Manager Performance Rank; see also Heller Dep. at 322, ECF No. 248-21 (noting that quartiles have "soft edges").

          At the conclusion of the quartiling process, managers ensure the units have complied with distribution goals. See Heller Dep. 84:15-85:2, 90:17-91:1, 320-21; Landman Dep. at 15:3-11, ECF Nos. 248-13, 265-14; Larson Dep. at 191:23-192:10, ECF Nos. 248-24, 265-19, 358-5; Kung Dep. at 311:16-312:8. That is, managers may move employees to different quartiles in order to enforce the distribution and avoid "grade inflation." Francis R & R, at 9; Heller Dep. at 320-21; Kung Dep. at 322, 332; Larson Dep. at 161-62, 187-88; see also Campion Report ¶¶ 50, 59, ECF No. 294 (forced ranking "mitigates the tendency to homogenize evaluation scores" and to award everyone "an above-average rating"). In a similar vein, Human Capital Management— Goldman Sachs’ human resources department— will inquire and adjust rankings if there are great discrepancies between the 360 reviews and the quartile. Kung Dep. at 325-27.

          B. Compensation

          Professionals at Goldman Sachs are typically compensated with a base salary and a year-end bonus, with the bonus representing the largest portion of a professional’s annual compensation. Francis R & R, at 10; Mehling Dep. at 226-27, ECF Nos. 248-25, 265-22, 358-6. Some professionals, such as Private Wealth Advisors, receive a cut of commissions rather than a year-end bonus.

          The process of determining compensation begins with Goldman Sachs allocating a budget to each division in the company. Francis R & R, at 10; Kung Dep. at 56-57; Larson Dep. at 60. Managers at the division level, typically the Chief Operating Officer and Chief Financial Officer, then further divide the budget among the various business units. Francis R & R, at 10; Kung Dep. at 56-57; Larson Dep. at 61. Managers then recommend compensation for specific employees. Francis R & R, at 10; Kung Dep. at 58-59; Larson Dep. at 61. Compensation recommendations then wind their way back up the chain, Francis R & R, at 10, with compensation proposals reviewed by the business unit head, then reviewed by a division-wide compensation committee, and finally, by the firm-wide compensation committee, id. at 10-11; Kung Dep. at 59-60, 64-67, 68-70; Heller Dep. at 24; Larson Dep. at 62-63. Quartile scores informed but were not determinative of compensation decisions. See, e.g., Casturo Decl. ¶ 17, ECF No. 268 ("[W]e consider manager quartile in making compensation decisions for both sales representatives and traders.... [but] it is just one input into compensation determinations."); McNamara Decl. ¶ 18 n.5, ECF No. 282.

          C. Promotion

          At Goldman Sachs, the vetting process for promoting Vice Presidents to Managing Directors is known as "cross-ruffing." Francis R & R, at 11; Kung Dep. at 398; Heller Dep. at 80; Larson Dep. at 227. Cross-ruffing does not involve an application process; rather, the firm selects Vice Presidents to promote through cross-ruffing. Vice Presidents generally need to rank in the first or second quartile to be promoted. Kung Dep. at 429 ("[O]ur strongest performers are promoted and typically they would be ranked Q1s and Q2s.").

          At the outset, human resources staff work with business unit heads and other managers to develop a division-wide list of candidates for promotion. Heller Dep. at 205:18-23, 234:1-3; Larson Dep. at 227:15-21; Kung Dep. at 414-15, 419:5-24; see also Ward Report, at 73. The list is often ranked to emphasize priority candidates. There is significant overlap between the persons on the compensation committee and the persons on the cross-ruffing committee that determines eligibility for promotion. Larson Dep. at 228.

          This list of candidates then goes to division heads. Larson Dep. at 232. Division heads may remove or add a candidate, and they may modify their relative rank. Kung Dep. at 417-21; Larson Dep. at 232-35. Following approval by the division head, the list goes onto the firm-wide talent assessment group. Larson Dep. at 240.

          In addition to compiling a list of candidates for promotion, the firm develops a list of "cross-ruffers." Cross-ruffers are Managing Directors at the firm who evaluate the candidates. Heller Dep. at 218, 224; Larson Dep. at 224; Kung Dep. at 440-42. Human resources initially generates a list of proposed cross-ruffers in consultation with the division heads, which is then submitted to the firmwide talent assessment group. Larson Dep. at 235, 240. After the cross-ruffers are chosen, they are trained by the firm-wide talent assessment group. Id. at 241.

          The process of evaluating the candidates primarily involves interviewing about a dozen persons familiar with the candidate’s work, including managers, peers, and internal clients. Id. at 242-44. Cross-ruffers will be assigned a handful of candidates for evaluation who are outside the cross-ruffer’s business unit. In the course of their interviews, cross-ruffers complete standardized forms and criteria to be discussed in the interviews. Id. at 245. After the interviews are complete, the cross-ruffers meet for a day-long meeting to develop a ranked list of candidates, which is ultimately submitted to the firm-wide talent assessment group and division heads. Larson Dep. at 245-46; Heller Dep. at 226; Kung Dep. at 448-49. Ultimately, Goldman Sachs’ management committee decides who is promoted. Larson Dep. at 248-50; Francis R & R, at 13-14.

          III. Procedural History

         On September 16, 2010, Plaintiffs H. Christina Chen-Oster and Shanna Orlich² filed this putative class action alleging intentional discrimination, disparate impact discrimination, retaliation, and pregnancy discrimination under Title VII and the NYCHRL. ECF No. 5. The Honorable Leonard B. Sand originally presided over this case, before it was reassigned to the undersigned on May 24, 2013. ECF No. 181.

2 Originally, Lisa Parisi was also a named plaintiff. Because Parisi’s claims are subject to a binding arbitration agreement, however, ECF No. 171; see Parisi v. Goldman, Sachs & Co., 710 F.3d 483 (2d Cir. 2013), she is no longer a named Plaintiff, see generally Second Am. Compl.

          A. Standing

         This case’s long and circuitous history is largely attributable to a dispute over whether former Goldman Sachs employees can obtain injunctive and declaratory relief against the firm. In 2012, Judge Sand ruled that they could not, granting Goldman Sachs’ motion to strike Plaintiffs’ Rule 23(b)(2) class allegations for want of standing. ECF No. 158 at 14; Chen-Oster v. Goldman, Sachs & Co., 877 F.Supp.2d 113, 122 (S.D.N.Y. 2012). That is, Judge Sand held that Plaintiffs who no longer worked at Goldman Sachs lacked standing as a class to enjoin or declare unlawful the firm’s practices that allegedly discriminated against women. Chen-Oster, 877 F.Supp.2d at 122.

          On April 13, 2015, Plaintiffs filed a motion to add two additional named Plaintiffs, Allison Gamba and Mary De Luis, ECF No. 377, which was granted on August 3, 2015, ECF No. 410; see also ECF No. 450 (affirming magistrate order at ECF No. 410). That same day, Plaintiffs filed a second amended complaint that included Gamba’s and De Luis’ claims. ECF No. 411.

          Relying on Judge Sand’s 2012 order, ECF No. 158, Defendants moved to dismiss Gamba’s claims for injunctive and declaratory relief on September 28, 2015, ECF No. 441. In a similar vein, by letter dated May 9, 2016, Defendants notified the Court that De Luis no longer worked for Goldman Sachs, ECF No. 446, and subsequently moved to dismiss her claims on the same ground, ECF No. 457.

          On April 12, 2017, the Court denied Goldman Sachs’ motion to dismiss Gamba and De Luis’ claims for injunctive and declaratory relief, and permitted Plaintiffs to file a supplemental complaint, ECF No. 479, which they did on April 12, 2017, ECF No. 486. Because the rule of law announced in Judge Sand’s 2012 order relied on dicta and was in conflict with several other decisions considering the issue, the Court abrogated it and revived Plaintiffs’ claims for declaratory and injunctive relief. ECF No. 479, at 5-11. In effect, the Court reasoned that because a former employee may face the same allegedly discriminatory policy after she is reinstated, she has standing to sue for injunctive and declaratory relief. See id.           Goldman Sachs sought interlocutory review of this ruling on April 25, 2017. ECF No. 484. On June 14, 2017, the Court granted Goldman Sachs’ request and certified two questions for appeal: (1) whether former employees lack standing to seek injunctive or declaratory relief against their former employer, and (2) whether a former employee who has not alleged unlawful discharge can seek the remedy of reinstatement under Title VII. ECF No. 500. On August 29, 2017, the Second Circuit denied Goldman Sachs’ petition for interlocutory review. ECF No. 539.

          B. Class Certification and Daubert Motions

          Plaintiffs filed their motion for class certification on May 19, 2014, ECF No. 246, and the Court referred it to the Honorable James C. Francis for a Report and Recommendation, ECF No. 240. Judge Francis held a two-day hearing on the admissibility of expert testimony and class certification. See generally Class Cert. Hr’g Tr. I; Class Cert. Hr’g Tr. II. The M & O on expert testimony excluded some, but not all, expert testimony submitted in connection with Plaintiffs’ motion for class certification. See generally Francis M & O. The R & R determined that all factors of Rule 23(a)— numerosity, commonality, typicality, and adequacy— had been satisfied. Yet, Judge Francis ultimately recommended against certifying a class pursuant to Rule 23(b)(3) because individual causation issues would predominate over class-wide issues. Francis R & R, at 6-14, 24-35. And in reliance on Judge Sand’s 2012 order, the R & R also recommended that former employees could not seek injunctive or declaratory relief.

          On August 30, 2017, the Court revived Plaintiffs’ injunctive and declaratory class claims after the Second Circuit’s denial of interlocutory review, and vacated those parts of the R & R that, in reliance upon Judge Sand’s earlier ruling, recommended that Rule 23(b)(2) and (c)(4) class certification were unavailable to Plaintiffs. ECF No. 540.

         As a result, this opinion addresses only the portion of Plaintiffs’ motion that seeks class certification under Rule 23(b)(3). Class claims for injunctive and declaratory relief under Rule 23(b)(2) and (c)(4) are now being pursued on a separate track— currently scheduled to be fully submitted in late 2018.³ See ECF No. 568.

3 In their objections, Plaintiffs alternatively argue for the Court to certify a class pursuant to Rule 23(c)(4). The Court declines to consider (c)(4) arguments in the context of this opinion on Rule 23(b)(3).

          DISCUSSION

          I. Expert Testimony

          A. Legal Standards

          1. Rule 702

         Under Rule 702 of the Federal Rules of Evidence, expert testimony may be admitted if it (1) "help[s] the trier of fact to understand the evidence or to determine a fact in issue," (2) "is based on sufficient facts or data," (3) "is the product of reliable principles and methods," and (4) "the expert has reliably applied the principles and methods to the facts of the case." Fed.R.Evid. 702. In other words, Rule 702 requires expert testimony to "be not only relevant, but reliable." United States v. Romano, 794 F.3d 317, 330 (2d Cir. 2015).

         The Supreme Court’s decision in Daubert detailed factors for district courts to consider in determining reliability, "includ[ing] the theory’s testability, the extent to which it ‘has been subjected to peer review and publication,’ the extent to which a technique is subject to ‘standards controlling the technique’s operation,’ the ‘known or potential rate of error,’ and the ‘degree of acceptance’ within the ‘relevant scientific community.’ " Id. (quoting Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 593-94, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993) ). "Under Daubert and Rule 702, expert testimony should be excluded if the witness is not actually applying expert methodology." United States v. Dukagjini, 326 F.3d 45, 54 (2d Cir. 2003).           2. Standard of Review

         A district court reviews a magistrate judge’s decision to admit or exclude expert testimony under the "clearly erroneous or contrary to law" standard. 28 U.S.C. § 636(b)(1)(A); see also Sansalone v. Bon Secours Charity Health Sys., Inc., No. 05 Civ. 8606, 2009 WL 1649597, at *1 (S.D.N.Y. June 11, 2009) (explaining that "the decision to admit or exclude expert testimony is considered nondispositive of an action" and, therefore, the " ‘clearly erroneous or contrary to law’ standard of review ... [of] Federal Rule of Civil Procedure 72" applies). "An order is clearly erroneous if the reviewing court is left with the definite and firm conviction that a mistake has been committed." Lifeguard Licensing Corp. v. Ann Arbor T-Shirt Co., LLC, No. 15 Civ. 8459, 2017 WL 3142072, at *1 (S.D.N.Y. July 24, 2017) (internal quotation marks omitted). "An order is contrary to law when it fails to apply or misapplies relevant statutes, case law or rules of procedure." Id. Here, the Court concludes that the M & O does not contain clear error and is not contrary to law.

          B. Michael A. Campion

         Judge Francis denied Plaintiffs’ motion to strike portions of an expert report prepared for Defendants by Michael A. Campion, Ph.D., a Professor of Management at Purdue University. See Francis M & O, at 33-34. Judge Francis reasoned that Plaintiffs’ argument went "exclusively to Dr. Campion’s opinions on the merits of their claims and not to issues of class certification," and, accordingly, denied their motion as premature. Id. at 34. In their objections, Plaintiffs do not argue that Dr. Campion’s opinions do, in fact, go to class certification. Rather, they argue only that "another Court may determine otherwise" and Judge Francis’ ruling thereby "opens the possibility that starkly unreliable opinions ... will inform the record at a later stage of the case, or even at certification." Pls. Objs. Campion, at 3-4, ECF No. 501.

         Even if Plaintiffs are correct, the Court cannot infer that Judge Francis’ ruling on prematurity was clear error or contrary to law. The Daubert inquiry is limited by the stage of the litigation. See, e.g., In re NYSE Specialists Sec. Litig., 260 F.R.D. 55, 66 (S.D.N.Y. 2009) (explaining that, at the certification stage, "the Court may only examine the expert reports as far as they bear on the Rule 23 determination"). And even if, as Plaintiffs argue, "the Court has the authority to exclude unreliable opinions that do not go to disputed issues," Pls. Objs. Campion, at 5 (emphasis added), Judge Francis’ decision not to exercise that authority is not clearly erroneous.

          C. Michael P. Curran

         In his M & O, Judge Francis excluded the expert report that Michael P. Curran, a director of Towers Watson, a business consulting firm, prepared on behalf of Defendants (the "Curran Report"). Francis M & O, at 34-37. In his report, Curran "expressed opinions about the structure of compensation in the financial services industry" at large. Id. at 35. In their motion to strike the Curran Report, ECF No. 306, Plaintiffs explained, first and foremost, that, by his own admission, Curran "did not consider, much less rely on, any Goldman-specific documents in connection with any of his opinions," id. at 2. As a result, Plaintiffs argued, the Curran Report was irrelevant at this stage of the litigation. Id. Judge Francis agreed, ruling that the fact that the report was not Goldman-specific was "fully dispositive" because industry practice was "not pertinent to class certification." Francis M & O, at 36-37.

          Defendants largely argue that Judge Francis committed "clear error" because the Curran Report’s opinions on industry practice (1) buttress the other testimony Defendants submitted, especially the testimony of their statistical expert, Michael P. Ward, Ph.D., on the impact of gender at Goldman, and (2) undermine the testimony of Plaintiffs’ rival statistical expert, Henry S. Farber, Ph.D. See Defs. Objs. Curran, at 3-6, ECF No. 503. Defendants are incorrect.

         Because the Curran Report discusses only industry practice, the Court finds it unlikely that it can either support or undermine Dr. Ward’s or Dr. Farber’s statistical analyses of Goldman Sachs, specifically. General knowledge about the financial services industry (or, indeed, uninformed speculation about Goldman Sachs) could hardly be relevant or reliable on the question of whether a statistician’s methodology is sound or supported. See SEC v. Tourre, 950 F.Supp.2d 666, 677-78 (S.D.N.Y. 2013) (excluding testimony because the expert’s knowledge and expertise was in "so broad a category as to become meaningless when particularized" to the issues of the case). To be sure, the Court agrees with Judge Francis and Defendants that, as a general matter, expert testimony on industry practice may be admissible at the class certification stage. See Defs. Objs. Curran, at 6-8. But as Judge Francis explained, the Curran Report does not address the question at hand: "whether there are common issues relating to Goldman Sachs and the claims against it." Francis M & O, at 37. Accordingly, excluding the report was not clear error.

          D. Michael P. Ward

         Judge Francis also granted Plaintiffs’ motion to strike a portion of the expert report of Dr. Ward, Defendants’ statistical expert on gender. Id. at 25-29. Specifically, Judge Francis excluded two studies consisting of "matching pair" analyses. Id. In both exercises, Dr. Ward examined pairs of similarly situated Vice Presidents at Goldman Sachs who, despite their similarity, had ended up with "very different" compensation. Id. at 26-27. Dr. Ward then spoke with the Vice Presidents’ managers, and, ultimately, concluded that there were factors contributing to the difference in compensation that the parties’ statistical models did not capture. For the first study, Dr. Ward looked to Dr. Farber’s factors to match "similarly situated" pairs. Id. at 26-27. For the second study, Dr. Ward matched pairs based on his own factors. Id.

          Judge Francis ruled that Dr. Ward’s matching pair studies were unreliable because (1) Dr. Ward made no findings on the statistical significance of the studies, (2) his sample size of only 2 pairs in the first study could not be statistically significant, (3) his analysis was "subject to bias" because it was Defendants’ counsel who hand-picked the pairs Dr. Ward discussed in his report, and (4) his analysis suffered from "a significant danger of reporting bias," given Dr. Ward’s reliance on interviews with the Vice Presidents’ managers. Id. at 27-28. Finally, in response to Defendants’ argument that the studies were not proffered as statistical evidence, Judge Francis concluded that, if true, the studies were not properly "expert" testimony and should be excluded. Id. at 29 ("It is inappropriate to present that evidence with the imprimatur of an expert witness.").

          Defendants argue that Judge Francis’ ruling was "clear error" because (1) each study "was not itself a ‘statistical analysis,’ " but was a "a real world check" on the reliability of his and Dr. Farber’s statistical analyses, and that such a "real world check" permitted Dr. Ward to "render an expert opinion on the competency of those models," (2) Dr. Ward was acting as an expert when conducting the studies, (3) Dr. Ward was permitted to rely on historical facts that had already come into the case, and (4) Dr. Ward was permitted to rely on biased witnesses. Defs. Objs. Ward, at 8-12, ECF No. 504.

         As an initial matter, as far as the Court is aware, it is undisputed that matching is a type of statistical technique. See id. at 10 ("The use of matched pairs also is routine in the field."); id. at 13 (arguing that matching is "a valid mechanism commonly used by statisticians"). Defendants cannot circumvent the requirements of Rule 702 and Daubert by labeling a statistical expert’s statistical exercises a "real world check." See id. at 9. A wolf in sheep’s clothing is still a wolf. Under Daubert, the Court must exercise its gatekeeping function accordingly, and "exclude unreliable expert testimony and junk science from the courtroom." See Almeciga v. Ctr. for Investigative Reporting, Inc., 185 F.Supp.3d 401, 415 (S.D.N.Y. 2016); see also F.D.I.C. v. Suna Assoc., Inc., 80 F.3d 681, 686 (2d Cir. 1996) ("[E]xpert testimony must be based upon reliable theories or principles."). For all the reasons Judge Francis outlined, therefore, the matching pair studies are unreliable. Excluding the studies was not clear error.⁴           E. Henry S. Farber

4 Even if Dr. Ward were permitted to testify that the matching pairs are simply a "meaningful lens" through which to consider his and Dr. Farber’s statistical analysis, Defs. Objs. Ward, at 9, Defendants would still be caught in an untenable position. In effect, they would be using junk science to argue that Plaintiffs’ expert analysis is based on junk science. Id. at 9-10. This argument is akin to pointing to a flying bird to refute the law of gravity— it is grounded in neither logic nor science.

         Defendants moved to exclude all the testimony of Dr. Farber, the Plaintiffs’ statistical expert on the impact of gender at Goldman Sachs,⁵ for two primary reasons: (1) "Dr. Farber impermissibly aggregated across three distinct Divisions in building his model," thereby failing both to pass the "Chow"⁶ test for aggregate analysis and to provide proof of Rule 23(a) commonality, and (2) "Dr. Farber’s regression analysis omitted the key variables that drove evaluation, compensation, and promotion decisions at Goldman Sachs," including the effect of the 360 review and quartiling processes on compensation and promotion. Defs. Objs. Farber, at 8, ECF No. 505.

5 In his report, Dr. Farber tested:

6 As the M & O explains, the "Chow" test "is designed to determine the extent to which statistical relationships are identical across different populations or within the same population over different time periods." Francis M & O, at 21; see also Ward Report, at 3 (The Chow test "tells statisticians whether or not the models for such groups are sufficiently homogenous to be analyzed together."). In layman’s terms, "the Chow test determines whether data can be pooled." Francis M & O, at 22.

          Judge Francis rejected both arguments. On Defendants’ aggregation argument, Judge Francis reasoned that (1) aggregate analysis was appropriate because "Goldman Sachs utilizes 360 review and quartiling in each division and every business unit," (2) disaggregation "tends to mask common mechanisms because the sample size in each unit is so small," (3) Dr. Farber’s analysis, under the law, was not required to pass the "Chow test," especially because there was a rational basis for aggregation, and (4) the large "standard error of estimate" of Dr. Farber’s analysis was irrelevant because he had demonstrated the statistical significance of his analysis. Francis M & O, at 17-25.

          On Defendants’ "key variable" argument, Judge Francis reasoned (1) Dr. Farber reasonably excluded the 360 review and quartile placement from his analysis because those factors, themselves, were tainted by gender disparities, (2) Dr. Farber nonetheless incorporated these two factors "into one iteration of his analysis, and doing so did not substantively alter his conclusions," (3) Dr. Farber reasonably excluded certain productivity measures from his analysis because of "practical barriers," (4) "the productivity that Goldman Sachs considers a crucial variable is already part of Dr. Farber’s model in the form of the performance measures," and (5) the fact that Defendants would have used a different proxy for job function than the job codes assigned pursuant to U.S. Department of Labor regulations "does not render Dr. Farber’s report inadmissible." Id. at 12-17.

         In their objections, Defendants reiterate their claim that Dr. Farber’s aggregate analysis does not follow the structure of Goldman Sachs’ business— specifically, its business units— and Defendants ask the Court to draw two distinct yet related conclusions from this fact. First, Dr. Farber’s failure to control for Goldman Sachs’ business units and his use of aggregation, Defendants argue, does not comport with Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 350, 131 S.Ct. 2541, 180 L.Ed.2d 374 (2011), because the Supreme Court rejected such analysis. Defs. Objs. Farber, at 11-15. But, as the Court details more fully below, Dukes involved the use of anecdotal and statistical evidence in the absence of common policies, which is not the case here. See infra Part II. Defendants provide no authority for their objection that "the Rule 23 question on commonality is not answered by the fact that the Firm used the same high-level personnel tools" across business units. Id. at 17. Accordingly, any argument that Dr. Farber’s aggregation violates Dukes is incorrect.

         Second, Defendants ask the Court to conclude that Dr. Farber’s aggregate analysis does not satisfy the reliability standards under Daubert . As Judge Francis ruled, however, "Dr. Farber’s decision to aggregate data ... was reasonable both in light of the evidence that Goldman Sachs applies common performance measures that influence pay and promotion across business units and in light of the statistical pitfalls of disaggregation." Francis M & O, at 21. Even if it were true that Dr. Farber’s analysis would be improved if it controlled for business units, the Court cannot conclude that it is so unreliable as to justify preclusion. Fed.R.Evid. 702 advisory committee’s note to 2000 amendment ("[T]he rejection of expert testimony is the exception rather than the rule."). Nor do these arguments persuade the Court that Judge Francis committed clear error.⁷

7 On the question of the Chow test, Defendants’ only answer to Judge Francis’ determination that passing the Chow test was not definitively required was to argue that Judge Francis’ "point is irrelevant because, in this case, the Chow test results merely provide an emphatic confirmation of what the undisputed evidence already establishes." Defs. Objs. Farber, at 19-20. This unsupported argument does not demonstrate clear error.

         Defendants next repeat that Dr. Farber’s analysis violates Daubert because it did not properly account for "key" variables: business units, job function, and employee productivity. Defs. Objs. Farber, at 20-29. The Court disagrees. At bottom, none of Defendants’ arguments go to admissibility. "[T]o be admissible, a regression analysis must control for the ‘major factors’ that might influence the dependent variable. But, ‘[n]ormally, failure to include variables will affect the analysis’ probativeness, not its admissibility.’ " Reed Const. Data Inc. v. McGraw-Hill Cos., Inc., 49 F.Supp.3d 385, 400-01 (S.D.N.Y. 2014) (internal citation omitted) (quoting Bazemore v. Friday, 478 U.S. 385, 400, 106 S.Ct. 3000, 92 L.Ed.2d 315 (1986) ), aff’d, 638 Fed.Appx. 43 (2d Cir. 2016). Especially in light of the fact that Dr. Farber accounted for divisions (rather than the less stable business units, as discussed in Part II) and job function via Department of Labor job codes, Judge Francis’ ruling on the disputed factors was not clear error.⁸ "Whether [Dr. Farber’s] omissions impermissibly biased the results [is] a matter for the jury to decide." Id. at 401.

8 Additionally, at least as to the Investment Banking Division, Dr. Ward also recognized the analytical issues of using the "productivity" variable. Ward Report, at 66 ("I found the statistical impact of production in IBD to be erratic, sometimes showing negative (nonsense) patterns, i.e. more production meaning lower [total compensation].").

          F. Wayne F. Cascio

         Finally, Judge Francis denied Defendants’ motion to strike the expert report prepared for Plaintiffs by Wayne F. Cascio, Ph.D. ("the Cascio Report"). The Cascio Report "evaluate[s] the performance assessment and compensation recommendation processes at Goldman Sachs," ultimately concluding that the processes were deficient for a number of reasons. Francis M & O, at 30. Judge Francis ruled that because the Cascio Report is relevant to "rebutting any claim by Goldman Sachs that its evaluation processes must be allowed even if they have a disparate impact on women," the report went "to the business necessity defense and is relevant at the class certification stage to the question of commonality." Id. at 31-32.

          Defendants first object that the Cascio Report does not address all the required elements of a business necessity defense. Defs. Objs. Cascio, at 8-9, ECF No. 506. To that end, Defendants, once again, argue that the Cascio Report does not address the processes on a job-by-job basis. Id. at 8. Additionally, Defendants argue that, even if the Cascio Report bears on business necessity, it does not bear on the question of commonality. Id. at 10-12. These arguments are without merit.

          Defendants’ contention that the Dr. Cascio’s report does not adequately address the specific elements of business necessity is, at best, a matter for the jury to decide, and, at worst, wrong on the law of business necessity. The same holds true for Defendants’ argument that the report does not bear on commonality, and, in fact, undermines a finding of commonality. Accordingly, the Court concludes that Judge Francis did not commit clear error.

          II. Class Certification

          A. Legal Standard

         A district court "may accept, reject, or modify, in whole or in part, the findings or recommendations made by the magistrate judge." 28 U.S.C. § 636(b)(1). "Courts review de novo those parts of a report and recommendation to which objections are made, and the remainder for clear error on the face of the record." Mulosmanaj v. Colvin, No. 14 Civ. 6122, 2016 WL 4775613, at *2 (S.D.N.Y. Sept. 14, 2016) (internal quotation marks omitted).

          B. Rule 23(a) Prerequisites

         A class may only be certified under Rule 23(b)(3) if it meets the requirements of Rule 23(a): numerosity, commonality, typicality, and adequacy. See Fed.R.Civ.P. 23(a)-(b). Although Judge Francis concluded that the proposed class satisfies all four of Rule 23(a)’s requirements, see Francis R & R, at 24-38, Defendants object to Judge Francis’ recommendation on the grounds of commonality, typicality, and adequacy.⁹ The Court disagrees. Defendants’ Rule 23(a) objections are, therefore, overruled.

9 Because no party objected to Judge Francis’ finding on numerosity, the Court reviews it for clear error. The Court concludes that Judge Francis did not commit clear error in determining that the putative class is sufficiently numerous for the purposes of Rule 23(a)(1).

          1. Commonality

          The crux of the commonality prerequisite is simpler than the parties’ briefing make it out to be. As the Supreme Court explained in Dukes, a putative class action against Wal-Mart, commonality is not merely a matter of common questions, "but, rather[,] the capacity of a classwide proceeding to generate common answers apt to drive the resolution of the litigation." Dukes, 564 U.S. at 350, 131 S.Ct. 2541 (quoting Richard A. Nagareda, Class Certification in the Age of Aggregate Proof, 84 N.Y.U. L.Rev. 97, 132 (2009) ). The Dukes court identified two ways that a class may, as relevant here, demonstrate commonality: (1) the employer’s use of "biased testing procedure[s]," and (2) "significant proof" that the employer "operated under a general policy of discrimination." Dukes, 564 U.S. at 353, 131 S.Ct. 2541 (internal quotation marks omitted). Here, Plaintiffs argue that they have met the requirements of this second standard— significant proof of a general policy of discrimination— for the purposes their disparate impact and disparate treatment claims.¹⁰

10 Disparate treatment claims concern "allegations of widespread acts of intentional discrimination." Robinson v. Metro-N. Commuter R.R. Co., 267 F.3d 147, 158 (2d Cir. 2001). Disparate impact claims, by comparison, involve "employment policies ... that are neutral on their face and were not intended to discriminate" but that "have nevertheless had a disparate effect on the protected group." Id. at 160.

         First, as to Plaintiffs’ disparate impact claims, the central question is whether a "general policy of discrimination" can exist where Goldman Sachs’ managers use discretion in making employment decisions. Defs. Objs. at 25-27, ECF No. 502. Dukes holds that "a policy of allowing discretion by local supervisors over employment matters" is, "[o]n its face, of course, ... just the opposite of a uniform employment practice that would provide the commonality needed for a class action; it is a policy against having uniform employment practices." Dukes, 564 U.S. at 355, 131 S.Ct. 2541. But beyond the "face" of a discretion-based system of supervision, "giving discretion to lower-level supervisors can be the basis of Title VII liability under a disparate-impact theory." Id. That is, "a common mode of exercising discretion that pervades the entire company" can constitute a general policy. Id. at 356, 131 S.Ct. 2541. Identifying this "common mode," id., is where Plaintiffs "must begin," id. at 357, 131 S.Ct. 2541 ("[T]he plaintiff must begin by identifying the specific employment practice that is challenged." (quoting Watson v. Fort Worth Bank & Tr., 487 U.S. 977, 994, 108 S.Ct. 2777, 101 L.Ed.2d 827 (1988) ) ). Once Plaintiffs have identified a common mode, they must offer "significant proof" of disparate impact to demonstrate commonality.

          The question of a "common mode" is irrelevant, however, for the purposes of pattern-or-practice disparate treatment claims. Davis v. District of Columbia, 246 F.Supp.3d 367, 393 (D.D.C. 2017) ("[D]isparate impact plaintiffs identify particular employment practices that are allegedly responsible for those disparities, while ‘pattern or practice’ plaintiffs do not."). This is because, if Plaintiffs can prove at the merits stage that Defendants had a common intention to discriminate, it is of no moment whether Defendants carried out their intention through a "common mode." Accord Hill v. City of New York, 136 F.Supp.3d 304, 354 (E.D.N.Y. 2015) ("[I]f Defendants are correct that Plaintiffs cannot show discriminatory intent, the class racial discrimination claims will be resolved in a single stroke."). Notably, however, "evidence of intent can be circumstantial, including evidence that is entirely statistical in nature." Davis, 246 F.Supp.3d at 393 (internal quotation marks omitted).

          a. Disparate Impact

          i. Common Mode

         Here, Defendants argue that Dukes requires Plaintiffs to show that Goldman Sachs managers, in making compensation and promotion decisions, applied the 360 review, quartiling, and cross-ruffing processes uniformly. Defs. Objs. at 26-30. They further argue that Plaintiffs cannot make that showing because the challenged processes "contained general criteria" that "individual managers applied ... in highly individualized ways." Id. at 25 (emphasis omitted). Defendants’ objections stretch Dukes far beyond its meaning.

         For the purposes of a disparate impact claim, Plaintiffs must first show that Defendants used "a common mode of exercising discretion. " Dukes, 564 U.S. at 356, 131 S.Ct. 2541 (emphasis added). In other words, a common mode need not strip managers of all flexibility in compensation and promotion decisions. Contra Defs. Objs. at 26. That is because, if "common mode" were so defined, it would divest lower-level managers of all discretion, and, as a result, would render the phrase "common mode of exercising discretion" oxymoronic. The Court declines to indulge Defendants’ semantic somersaults.

         In Dukes, despite Defendants’ arguments to the contrary, Defs. Objs. Reply, at 7-9, ECF No. 522, it was "the absence of a common job evaluation procedure" that precluded class certification on the plaintiffs’ disparate impact claim, Brown v. Nucor Corp., 785 F.3d 895, 909 (4th Cir. 2015) (discussing Dukes ). As the Dukes court explained, the Wal-Mart employees in that case were "subject to a variety of regional policies that all differed," rather than common policies or processes. Dukes, 564 U.S. at 359-60, 131 S.Ct. 2541 (quoting Dukes v. Wal-Mart Stores, Inc., 603 F.3d 571, 652 (9th Cir. 2010) (Kozinski, J., dissenting) ); see also id. at 356, 131 S.Ct. 2541 (explaining that plaintiffs had "not identified a common mode of exercising discretion," but only "statistical and anecdotal evidence"). As a result, because "respondents [in Dukes ] ... identified no ‘specific employment practice,’ " the "mere[ ] showing" of "an overall sex-based disparity d[id] not suffice." Id. at 357, 131 S.Ct. 2541.

         The Supreme Court has since reiterated this holding in Dukes . See, e.g., Tyson Foods, Inc. v. Bouaphakeo, __ U.S. __, 136 S.Ct. 1036, 1048, 194 L.Ed.2d 124 (2016) ("The plaintiffs in [Dukes ] did not provide significant proof of a common policy of discrimination to which each employee was subject. The only corporate policy that the plaintiffs’ evidence convincingly establishe[d was] Wal-Mart’s policy of allowing discretion by local supervisors over employment matters; and even then, the plaintiffs could not identify a common mode of exercising discretion that pervade[d] the entire company." (internal quotation marks omitted) ).

         Other courts have likewise interpreted Dukes to stand for the proposition that the absence of a company-wide policy was fatal to establishing commonality. Brown, 785 F.3d at 909; McReynolds v. Merrill Lynch, Pierce, Fenner & Smith, Inc., 672 F.3d 482, 488 (7th Cir. 2012) ("[B]ecause there was no company-wide policy to challenge in [Dukes ] ... there was no common issue to justify class treatment."); Houser v. Pritzker, 28 F.Supp.3d 222, 242 (S.D.N.Y. 2014) ("The Court ultimately determined that the plaintiffs had not satisfied the commonality requirement because Wal-Mart did not use a uniform ‘testing procedure or other companywide evaluation’ in making pay and promotion decisions, but rather left those decisions to regional managers’ discretion." (quoting Dukes, 564 U.S. at 353, 131 S.Ct. 2541) ); Calibuso v. Bank of Am. Corp., 893 F.Supp.2d 374, 377 (E.D.N.Y. 2012) ("The fatal flaw identified by the Supreme Court in Dukes was that the class claims alleged discrimination by local managers exercising their broad, subjective discretion in the absence of any policies, which by its very nature could not satisfy the commonality requirement of Rule 23(a)(2)." (emphasis added) ).

          Here, however, "[t]here is no dispute that each of the class members was subject to [the 360 review and quartiling] processes," Francis R & R, at 27, or that all promotions from Vice President were awarded through the cross-ruffing process, Kung Dep. 398:11-21. Nor is it disputed that 360 reviews inform quartiling decisions, which, in turn, inform promotion decisions.

         These critical facts distinguish the instant case from Dukes, as well as from the other cases upon which Defendants rely. Cf. Breen v. Chao, 253 F.Supp.3d 244, 266 (D.D.C. 2017) (explaining that the "specific employment practice" plaintiffs identify for the purposes of a disparate impact claim had to be an "employment policy or practice ... that defendants ... repeat over and over with respect to different employees").¹¹ For example, Defendants rely on Ruiz v. Citibank, N.A. without explaining that the district court in that case understood the plaintiffs to be attempting to prove commonality via "anecdotal evidence that reaches a certain critical mass" in light of "the absence of an express company policy." 93 F.Supp.3d 279, 289 (S.D.N.Y. 2015). The instant case is not merely an "anecdotal evidence" case. Plaintiffs’ disparate impact claims are predicated on express company policies. See Menocal v. GEO Grp., Inc., 882 F.3d 905, 916 (10th Cir. 2018) (finding commonality "[b]ecause all members of the ... class base their claims on the [same] [p]olicy"). These common policies include: (1) the job evaluation processes developed by a central Goldman Sachs committee and applied to every single member of the class, Francis R & R, at 6-11, and (2) the cross-ruffing process applied to every single possible promotion from Vice President, id. at 11-14.

11 Dukes is inapplicable to this case in yet another way. The Dukes court repeatedly returned to the massive scale of the class. Indeed, the first sentence of the opinion begins by calling the case "one of the most expansive class actions ever." Dukes, 564 U.S. at 342, 131 S.Ct. 2541. The class in Dukes, the Court emphasized, comprised 1.5 million plaintiffs. Id. Here, however, there are not millions of class members, only about 2,000— an average number for class actions. See Employment Class and Collective Actions: Proceedings of the New York University 56th Annual Conference on Labor, at 122-23 ("For a race claim, the average class size is 1,285, for a sex claim it is 6,432 and for an age claim it is only 761.").

         These processes permitted managers to exercise discretion, but, as explained above, Dukes specifically anticipated a case where a discretion-based policy that "pervades the entire company" would support a disparate impact claim. Dukes, 564 U.S. at 356, 131 S.Ct. 2541. Managerial discretion does not destroy Plaintiffs’ claims. Whether Goldman Sachs discriminated against the class via its use of 360 reviews, quartiling, or cross-ruffing raises yes-or-no questions that can each be answered "in one stroke." Id. at 350, 131 S.Ct. 2541. Plaintiffs have, therefore, demonstrated that Defendants employed "common mode[s] of exercising discretion" for the purposes of Dukes .

          ii. Significant Proof

         Defendants next challenge Plaintiffs’ "significant proof" that these "common modes" were discriminatory. The significant proof, offered by Plaintiffs, is contained in the expert report of Dr. Farber, which analyzed the impact of gender on evaluation, compensation, and promotion at Goldman Sachs and found statistically significant gender disparities. Farber Report ¶¶ 7(a) ("Women are paid less than otherwise similar men, on average, and the difference in pay is statistically significant."), 7(b) ("Women are evaluated lower on the 360-degree review and are less likely to be ranked in the top quartile than otherwise similar men[.]"), 7(c) ("Women are paid less than similar men even when they receive the same quartile score."), 7(d) ("[D]ifferences between women’s and men’s quartile and 360-degree review ratings explain" some of the pay difference.), 7(f) ("[W]omen were promoted from Vice President to Managing Director at a lower rate than one would expect if they were promoted in the same pattern as men were promoted.").

          As with their motion to preclude Dr. Farber’s testimony, see supra Part I, Defendants argue here that Dr. Farber should have tested whether the gender disparities "were the result of anomalies specific to individual Business Units." Defs. Objs. at 31. Defendants take issue with Dr. Farber’s aggregate analysis, which controlled for divisions, but not business units. Id. Additionally, Defendants maintain that Dr. Farber "failed to account for the factors that individual Goldman Sachs managers actually relied on in making evaluation and compensation decisions across the putative class." Id. The Court disagrees, concluding that Plaintiffs have established "significant proof" of disparate impact.

          As an initial matter, Defendants’ emphasis on business units is misguided. First, the business units are not stable or permanent. Over the years, Goldman Sachs has transferred business units from division to division, e.g., Benz Decl. ¶ 13, or eliminated units altogether. For example, 20% of units disappeared from one year to the next. Farber Rebuttal ¶ 32. And "there were as many as 227 business units in 2003 and as few as 94 in 2005." Id. Even if the units had been stable, the record shows Goldman Sachs employees were not. "Approximately 11 percent of Goldman Associates and Vice Presidents change business units in each year. " Id. (emphasis added). The volatility of the business units and their personnel counsel against analysis at the unit level.

          Additionally, the challenged performance evaluation processes do not occur on a unit-by-unit basis. The record shows that, for the 360 review process, 81.6% of class members received at least one review from a reviewer outside her business unit. Geman Decl. ¶ 5. In total, 37.4% of all 360 reviews of class members were conducted by reviewers from other business units. Id. ¶ 6. Meanwhile, for quartiling, a business unit manager’s "tentative [quartile] ranking is subject to adjustment at the division level and by Human Capital Management, Goldman Sachs’ human resources department." Francis R & R, at 9. "These modifications may be triggered by discrepancies between 360 review scores and the initial quartiling, by the desire to enforce the quartile distribution to prevent ‘grade inflation,’ or by other considerations." Id. In other words, 360 reviews were the product of cross-unit collaboration, and division heads enforced quality control of ranking to be consistent with those reviews.

          Similarly, cross-ruffing is an inter-business unit undertaking. Indeed, cross-ruffers are not permitted to evaluate candidates in their own business unit. See, e.g., Kung Dep. at 398 ("[C]ross-ruffer[s] will be assigned to interview candidates in a business area outside of their own."). In addition, no conclusive promotion decisions are made at the business unit level— division heads, the firm-wide talent assessment group, and the firm’s senior management wield a great deal more influence. Managers across business units develop the first iteration of a list of promotion candidates with human resources, which division heads then have free reign to review or modify. Kung Dep. at 417-21; Larson Dep. at 232-35. The division-approved list then goes onto the firm-wide talent assessment group, and ultimately, decisions are made by the firm’s management committee. Larson Dep. at 240, 248-50. Tellingly, Defendants’ own expert did not control for business units on the question of promotion. Ward Report, at 74-75.

          Still, Defendants’ briefs emphasize the role of business unit managers in these processes. See generally Defs. Objs. But the question is not whether business unit managers were involved in the evaluation, compensation, and promotion processes. Obviously, managers were involved. The question is whether Dr. Farber’s analysis controlling for divisions provides "significant proof" of discrimination. Given the instability of the business unit designation, the cross-unit evaluation of class members, and the funneling of virtually all final decision-making to division heads and either division-wide or firm-wide committees, the Court finds that Dr. Farber’s aggregate analysis is more than adequate proof.

          Indeed, Dr. Farber controlled for the following variables:

division, year, office, education, affirmative action ("AA") job group, experience at Goldman, experience at Goldman squared (both from the most recent hire date), relevant experience prior to most recent date at Goldman (including experience in prior employment spells at Goldman), relevant experience squared, whether a direct hire into the Associate or Vice President position and whether a direct hire into the Associate or Vice President position in the current year.

          Farber Report ¶ 7(a). These are reasonable factors when testing for gender discrimination at the three relevant divisions within Goldman Sachs. The Court agrees with Judge Francis that, for the reasons discussed above, "it is perfectly appropriate for the model to include data across business units when it is being used to explore the impact of a firm-wide policy such as 360 review or quartiling." Francis R & R, at 31. Accordingly, Plaintiffs have established commonality for the purposes of their disparate impact claims.

          b. Disparate Treatment

         Defendants spend the majority of their argument on the commonality of Plaintiffs’ disparate treatment claims debating the merits of the claims. Defs. Objs. at 36-46. But "this is not the stage at which to decide whether there is in fact a pattern of pervasive discrimination at Goldman Sachs." Francis R & R, at 33; see also Hill, 136 F.Supp.3d at 352 (explaining that the commonality "analysis will frequently entail overlap with the merits of the plaintiff’s underlying claims.... [but that] FRCP 23 does not grant the Court license to engage in free-ranging merits inquiries; merits questions are to be considered only to the extent that they are relevant to the FRCP 23(a) inquiry" (internal quotation marks and citations omitted) ).

          Here, Plaintiffs have provided "significant proof" on their disparate treatment claims via statistical and anecdotal evidence. As to statistical evidence, Dr. Farber’s report shows that Defendants pay female Vice Presidents 21% less and Associates 8% less, on average, than their male counterparts. Farber Report ¶ 7(a). The report explains that approximately 22% and 50% of the observed pay differences, respectively, can be explained by the differences in female and male 360 review and quartile ratings. Id. ¶ 7(d). These pay differences are statistically significant by several standard deviations, see Farber Report, Tables 10, 14, 15, and are sufficient to meet the "significant proof" standard at the class certification stage of litigation.

          As to anecdotal evidence, Plaintiffs have submitted internal complaints, external complaints, survey answers, emails, articles, business records, and declarations from class members. Pls. Objs. at 13-28. This evidence details how women at Goldman Sachs have been subject to sexual assault, e.g., Chen-Oster Decl. ¶ 10 ("I was sexually assaulted by a married male co-worker."), sexual harassment, e.g., ECF No. 248-6 ("I told [a male co-worker] how it made me uncomfortable how the guys were touching me, and he was really supportive and giving me advice on what to do, and the next thing I know, his hand is on my ass, too!"), sexualization, e.g., id. (managers made comments that women are "no good after [age] 27"), physical endangerment, e.g., id. (male manager followed female employee into her room during a firm retreat, tried to get in her bed, and did not leave until she was able to lock the door), private verbal abuse, e.g., ECF No. 248-7 ("How did the bitch session go?"), public verbal abuse, e.g., ECF No. 248-6 (head of Private Wealth Management, told his female assistant she would be a "trophy wife" once she marries), slander and libel, e.g., ECF No. 248-12 (male employee perpetuated a rumor that a sex tape of him with an unidentified woman was actually of him and a female co-worker), gender favoritism, e.g., ECF No. 248-6 ("I have to compensate the men better. They are heads of households."), pregnancy stereotyping, e.g., Albanese Decl. ¶ 12, ECF No. 251 ("During my pregnancy, Goldman Sachs removed my duties and took away my assistants. I was told that I would be transferred to a department of one person (me) with no advancement opportunities."), motherhood stereotyping, e.g., ECF No. 248-6 ("At one point my boss suggested that I might want to speak with the women’s committee and go into a job that would not be as demanding" because of new role as a mother.), denial of employment opportunities, e.g., Gamba Decl. ¶ 15 ("Since returning from maternity leave for my second child, I have never again been nominated for promotion despite a record of achievements.... [My manager] sa[id] he would be ridiculed if he [supported me]."), e.g., Orlich Decl. ¶ 9 ("[A] Vice President asked every man in my department and neighboring groups to join him on a golf outing, but did not ask me.... [although] I have played golf since childhood and played varsity golf on my high school team."), and retaliation, e.g., ECF No 248-6 ("[A]fter I raised complaints of gender discrimination, [my manager] questioned my ‘work ethic’ and said I was not working hard enough.").

          Plaintiffs have also provided proof that Defendants were aware of gender disparities and gender bias at Goldman Sachs, but did not adjust their policies or practices to account for the discrimination women faced. Plaintiffs provide documents demonstrating that Goldman Sachs recognized the bias of its processes, e.g., ECF No. 248-7 ("The underlying assessment of individuals that feed into the firm’s processes are colored and impacted by gender differences."), that women at Goldman Sachs suffered professionally because of the firm’s policies and processes, e.g., ECF No. 248-8 ("We have limited experience in dealing with the issues surrounding managing a professional on maternity leave, and some of our experience has proven that we are not very good at it."), and that Goldman Sachs did not adequately apply corrective measures, e.g., ECF No. 248-7 ("Our recommendation is that he be issued a strongly worded written warning" for having "perpetuated" rumors that he had recorded a sex tape of him with a female co-worker.).

         Plaintiffs have, therefore, provided "significant proof" of discriminatory disparate treatment. Whether Plaintiffs’ evidence conclusively proves discrimination is a matter for trial. Cf. Hnot v. Willis Grp. Holdings Ltd., 241 F.R.D. 204, 211 (S.D.N.Y. 2007) (observing that even where "the merits and the Rule 23(a) commonality determination significantly overlap" the question of whether defendants actually discriminated is a question for trial).

          2. Typicality and Adequacy

          Typicality "is satisfied when the lead plaintiffs’ claims arise from the same series of events and find support in the same legal theories as the claims of all of the remaining class members." Houser, 28 F.Supp.3d at 245; see also Fed.R.Civ.P. 23(a)(3) (requiring that "the claims or defenses of the representative parties are typical of the claims or defenses of the class"). The lead plaintiffs’ claims do not have to be "identical" to the class members’ claims, but must "have the ‘incentive to prove all the elements of the cause of action which would be presented by the individual members of the class were they initiating individualized actions.’ " Houser, 28 F.Supp.3d at 245 (quoting In re NASDAQ Market-Makers Antitrust Litig., 169 F.R.D. 493, 510 (S.D.N.Y. 1996) ). In other words, typicality "requires that the disputed issue of law or fact ‘occupy essentially the same degree of centrality to the named plaintiff’s claim as to that of other members of the proposed class.’ " Caridad v. Metro-N. Commuter R.R., 191 F.3d 283, 293 (2d Cir. 1999), called into question on other grounds by In re IPO, 471 F.3d 24 (2d Cir. 2006).

         Somewhat similarly, adequacy requires courts to inquire whether (1) "plaintiff’s interests are antagonistic to the interest of other members of the class," and (2) "plaintiff’s attorneys are qualified, experienced and able to conduct the litigation." Floyd v. City of New York, 283 F.R.D. 153, 161 (S.D.N.Y. 2012) (quoting Baffa v. Donaldson, Lufkin & Jenrette Secs. Corp., 222 F.3d 52, 60 (2d Cir. 2000) ). Defendants do not appear to contest the latter point. On the former point, "[i]n order to defeat a motion for certification, any conflicts between the class representative and members of the putative class must be ‘fundamental.’ " Ligon v. City of New York, 288 F.R.D. 72, 80 (S.D.N.Y. 2013) (quoting In re Flag Telecom Holdings, Ltd. Sec. Litig., 574 F.3d 29, 35 (2d Cir. 2009) ).

          Although Rule 23(a) commonality, typicality, and adequacy are technically distinct from one another, they tend to blur in certain contexts:

The commonality and typicality requirements of Rule 23(a) tend to merge. Both serve as guideposts for determining whether under the particular circumstances maintenance of a class action is economical and whether the named plaintiff’s claim and the class claims are so interrelated that the interests of the class members will be fairly and adequately protected in their absence. Those requirements therefore also tend to merge with the adequacy-of-representation requirement, although the latter requirement also raises concerns about the competency of class counsel and conflicts of interest.

Dukes, 564 U.S. at 349 n.5, 131 S.Ct. 2541 (quoting General Telephone Co. of Southwest v. Falcon, 457 U.S. 147, 157-58 n.13, 102 S.Ct. 2364, 72 L.Ed.2d 740 (1982)). Accordingly, and given that Defendants in the instant case argue about typicality and adequacy together, the Court will address both "in tandem." See, e.g., Floyd v. City of New York, 283 F.R.D. 153 (S.D.N.Y. 2012) (discussing typicality and adequacy argument "in tandem" because "Defendants make overlapping objections on the basis of typicality and adequacy"). The Court concludes that each of the named Plaintiffs meet the typicality and adequacy requirements.

          a. Shanna Orlich

         Defendants argue that each of the named Plaintiffs is atypical of the class and, at least some, if not all, are inadequate representatives.¹² First, Defendants argue that Orlich was a first-year associate whose compensation was in "lockstep" with her peers, and, as a result, not based on the 360 review or quartiling process.¹³ Defs. Objs. at 47. Indeed, Defendants state that Plaintiffs’ claim that Orlich was subjected to the challenged processes is "incorrect[ ]." Defs. Objs. Reply, at 36. As Plaintiffs point out, however, just because Orlich’s initial starting salary and sign-on bonus were in lockstep with her peers does not make her atypical of the class. Pls. Opp. Defs. Objs. at 28. Orlich’s declaration states that she participated in the 360-review process and was ranked by her manager through the quartiling system, Orlich Decl. ¶¶ 5-6, before being fired, id. ¶ 11. Her claims, therefore, "arise from the same series of events and find support in the same legal theories as the claims of all of the remaining class members." Houser, 28 F.Supp.3d at 245.           b. H. Christina Chen-Oster

12 In instances where Defendants appear to have challenged a named Plaintiff’s typicality but not adequacy, or vice-versa, the Court deems Defendants to have conceded the unchallenged ground.

13 Defendants further argue that Orlich was not eligible for a promotion and, therefore, she is atypical of the class. Defs. Objs. at 47. But a named plaintiff is not required, individually, to be typical of the class in every respect. Rather, the named plaintiffs, in aggregate, must cover each of the class claims.

         Second, Defendants take issue with Chen-Oster’s claim that Defendants retaliated against her for reporting that a male Goldman Sachs co-worker sexually assaulted her in 1997. Defs. Objs. 47-48. Defendants argue that "Judge Francis brushed aside the idiosyncratic nature of Mr. Chen-Oster’s [sic ] claim with the conclusory statement that it does not render her atypical of the class." Id. at 48. Defendants maintain that Chen-Oster is atypical because her "retaliation allegations are the core of her claim." Id. at 48 n.36. Plaintiffs respond that "[i]t is well-established, however, that a class representative’s individual retaliation claim does not defeat typicality." Pls. Opp. Defs. Objs. at 29.

         Although Plaintiffs are correct that an individual retaliation claim is not disqualifying for a class representative, Spencer v. No Parking Today, Inc., No. 12 Civ. 6323, 2013 WL 1040052, at *19 (S.D.N.Y. Mar. 15, 2013), adopted by 2013 WL 2473039 (S.D.N.Y. June 7, 2013), Plaintiffs gloss over the nuance of Defendants’ argument. The problem, Defendants argue, is not that Chen-Oster has an individual retaliation claim, but that her "retaliation allegations are the core of her claim." Defs. Objs. at 48 n.36. Still, Defendants’ objection is overruled. Chen-Oster participated in the 360 review process and was ranked through the quartiling process. Chen-Oster Decl. ¶¶ 5-6. She claims that she was "treated unfairly in the 360 review process and the forced ranking system," "was also denied compensation [and promotions] because of [her] gender," and that "Goldman Sachs maintains a culture that is hostile to women." Id. ¶¶ 7-10. Her class claims are, therefore, entirely typical.

         Defendants also seek to disqualify Chen-Oster on adequacy grounds; namely, that she "testified that Goldman Sachs’ processes were marked by unbounded manager discretion." Defs. Objs. at 48. Defendants argue that this testimony puts Chen-Oster "in conflict with the claims that she seeks to litigate on behalf of the class." Id. Defendants, however, mischaracterize Chen-Oster’s testimony. Chen-Oster was discussing the "boy’s club" culture at Goldman Sachs that concentrates power in men, who then give that power to other men, when Defendants asked her if "then the power is given to the managers to exercise unchecked discretion." Chen-Oster Dep. at 436-37. The Court agrees with Judge Francis that Chen-Oster’s affirmative answer "is a matter of semantic shading." Francis R & R, at 35. This stray comment does not place Chen-Oster at "fundamental" odds with the interests of the class. Ligon, 288 F.R.D. at 80.¹⁴ She is an adequate class representative.

14 In a footnote, Defendants also argue that "Chen-Oster’s credibility is in serious question" because she, at some point, recorded the conversations of Goldman Sachs employees. Defs. Objs. at 48 n.37. Given that Chen-Oster does not rely on these recordings, this argument is without merit.

          c. Allison Gamba

         Third, Defendants challenge Gamba as atypical because "Gamba conceded that her claims are limited to 2010, and in that year ‘nobody from [her] business unit was promoted to managing director,’ male or female." Defs. Objs. at 49 (quoting Gamba Dep. 103:18-20, ECF No. 265-10). Again, Defendants distort the actual testimony. Gamba did not concede that all her claims were limited to 2010, Pls. Opp. Defs. Objs. at 29-30, but only that her claims related to the change in her 360 reviewer list was limited to 2010, see Gamba Dep. 102-20. But it is undisputed that Gamba participated in the 360 review and quartiling processes in years prior to 2010, and, accordingly, her claims of gender discrimination are not limited to that time period. See generally Gamba Decl. The fact that she also argues gender discrimination based on a change to her 360 reviewer list in 2010 does not make her atypical of the class.

          d. Mary De Luis

         Finally, Defendants argue that De Luis’ claims "are entirely individualized and in no way typical" because her "compensation allegations are unrelated to the 360 review and manager quartile processes." Defs. Objs. at 50. The Court is not persuaded. De Luis participated in the 360 review and quartile processes and alleges being compensated less than her male peers. Second Am. Compl. ¶¶ 137-45. The fact that De Luis attempted to address the alleged pay discrepancy with her superiors, id., does not render her claims atypical.

          Additionally, somewhat similar to their arguments against Chen-Oster, Defendants argue that De Luis’ "retaliation claims are not brought on behalf of a putative class." Defs. Obj. at 50. Although Plaintiffs concede that De Luis brings an individual retaliation claim, Pls. Opp. Defs. Objs. at 30, they do not, as Defendants suggest, Defs. Objs. Reply, at 38, concede that De Luis’ claims all "arise[ ]" from her individual retaliation claim, id. As discussed above, a unique retaliation claim does not defeat the typicality of De Luis’ other class claims.

          C. Rule 23(b) Requirements

         Under Rule 23(b)(3), a district court may certify a class that meets the Rule 23(a) requirements if (1) "questions of law or fact common to class members predominate over any questions affecting only individual members," and (2) "a class action is superior to other available methods for fairly and efficiently adjudicating the controversy." Fed.R.Civ.P. 23(b)(3). Judge Francis recommended denying certification on the ground that Plaintiffs had not met the predominance requirement. He, therefore, did not reach the question of superiority. This Court concludes that Plaintiffs have met both the predominance and superiority requirements for some, but not all, of their claims.

          1. Predominance

          The predominance requirement "ensures that the class will be certified only when it would ‘achieve economies of time, effort, and expense, and promote uniformity of decision as to persons similarly situated, without sacrificing procedural fairness or bringing about other undesirable results.’ " Cordes & Co. Fin. Servs., Inc. v. A.G. Edwards & Sons, Inc., 502 F.3d 91, 104 (2d Cir. 2007) (internal alteration omitted) (quoting Amchem Prod., Inc. v. Windsor, 521 U.S. 591, 615, 117 S.Ct. 2231, 138 L.Ed.2d 689 (1997) ). Importantly, Plaintiffs meet this requirement "if resolution of some of the legal or factual questions that qualify each class member’s case as a genuine controversy can be achieved through generalized proof, and if these particular issues are more substantial than the issues subject only to individualized proof." Moore v. PaineWebber, Inc., 306 F.3d 1247, 1252 (2d Cir. 2002) (emphasis added). In other words, "[t]he predominance requirement calls only for predominance, not exclusivity, of common questions." In re LIBOR-Based Fin. Instruments Antitrust Litig., 299 F.Supp.3d 430, No. 11 Civ. 5450, 2018 WL 1229761, at *5 (S.D.N.Y. Feb. 28, 2018) (quoting In re Visa Check/MasterMoney Antitrust Litig., 280 F.3d 124, 140 (2d Cir. 2001) ). Still, it is "more demanding" than the commonality requirement under Rule 23(a). Id. (quoting Johnson v. Nextel Commc’ns Inc., 780 F.3d 128, 138 (2d Cir. 2015) ).

          To that end, on the question of predominance, courts "must assess (1) the ‘elements of the claims and defenses to be litigated’; and (2) ‘whether generalized evidence could be offered to prove those elements on a class-wide basis or whether individualized proof will be needed to establish each class member’s entitlement to relief.’ " Nextel, 780 F.3d at 138. This assessment is "more qualitative than quantitative, and must account for the nature and significance of the material common and individual issues in the case." In re LIBOR, 2018 WL 1229761, at *5 (internal quotation marks and alterations omitted).

         Ultimately, "[w]hen ‘one or more of the central issues in the action are common to the class and can be said to predominate, the action may be considered proper under Rule 23(b)(3) even though other important matters will have to be tried separately, such as damages or some affirmative defenses peculiar to some individual class members.’ " Tyson Foods, 136 S.Ct. at 1045 (quoting 7AA C. Wright, A. Miller, & M. Kane, Federal Practice and Procedure § 1778, pp. 123-24 (3d ed. 2005) ).

          Here, Plaintiffs have met the predominance requirements as to some, but not all of their claims. For Plaintiffs’ (1) disparate impact claims, and (2) disparate treatment claim predicated on the statistical evidence of disparate impact, the Court concludes that common questions predominate. By contrast, the Court concludes that Plaintiffs have failed to prove predominance on their disparate treatment claim predicated on their "boy’s club" theory.

          a. Elements and Defenses

         As an initial step in the predominance analysis, as explained above, the Court considers "the elements of the claims and defenses to be litigated." In re LIBOR, 2018 WL 1229761, at *5. Under Title VII, the three-step McDonnell Douglas burden-shifting framework governs disparate impact and disparate treatment claims. See United States v. City of New York, 731 F.Supp.2d 291, 299 (E.D.N.Y. 2010) (disparate impact); United States v. City of New York, 717 F.3d 72, 83 (2d Cir. 2013) (disparate treatment).

         At the first step of McDonnell Douglas, Plaintiffs bear the burden of making a prima facie showing of discrimination. City of New York, 717 F.3d at 84. At the second step, if Plaintiffs meet their initial burden, "the burden then shifts to the employer ‘to rebut the presumption of discrimination.’ " Id. (quoting Texas Dep’t of Cmty. Affairs v. Burdine, 450 U.S. 248, 254, 101 S.Ct. 1089, 67 L.Ed.2d 207 (1981) ). At the third and final step, the burden shifts back to Plaintiffs. As discussed below, however, each of these steps requires a somewhat different showing depending on whether Plaintiffs are bringing a disparate impact or disparate treatment pattern-or-practice claim. Id. ; see also City of New York, 731 F.Supp.2d at 299.

          b. Proof

          i. Disparate Impact

          For the purposes of a disparate impact claim, Plaintiffs make out a prima facie claim if they "(1) ‘identify a specific employment practice’ or policy; ‘(2) demonstrate that a disparity exists; and (3) establish a causal relationship between the two.’ " Chin v. Port Auth. of New York & New Jersey, 685 F.3d 135, 151 (2d Cir. 2012) (quoting Robinson, 267 F.3d at 160) (internal citation omitted). "Statistics alone can make out a prima facie case," City of New York, 731 F.Supp.2d at 300, but " ‘[t]he statistics must reveal that the disparity is substantial or significant,’ and ‘must be of a kind and degree sufficient to reveal a causal relationship between the challenged practice and the disparity,’ " Chin, 685 F.3d at 151 (quoting Robinson, 267 F.3d at 160).

         Reviewing de novo, as the Court must, the question of predominance at the first step of McDonnell Douglas, the Court concludes that Plaintiffs have proven predominance.¹⁵ As an initial matter, Plaintiffs have provided statistical evidence on the 360 review and quartiling processes to demonstrate a gender disparity at the statistically significant levels of 2.7 to 5.15 standard deviations. See Farber Report, Tables 10, 14, 15; see also City of New York, 731 F.Supp.2d at 301 ("The Second Circuit has repeatedly recognized that standard deviations of more than 2 or 3 units can give rise to a prima facie case of disparate impact because of the low likelihood that such disparities have resulted from chance." (quoting United States v. City of New York, 637 F.Supp.2d 77, 93 (E.D.N.Y. 2009) (alteration omitted) ). Whether Plaintiffs’ evidence actually makes out a prima facie case is a question left for trial, but, at this first step of McDonnell Douglas, it is clear that the generalized proof of statistical evidence is sufficient for the predominance inquiry.

15 Plaintiffs argue that Judge Francis misunderstood the first step of the McDonnell Douglas analysis and "missed that causation at the Rule 23 stage of a disparate impact claim may be demonstrated, as here, through expert statistical analysis." Pls. Objs. at 42. But the Court concludes that Plaintiffs misread the R & R. Contrary to Plaintiffs’ contentions, Judge Francis implicitly recognized that Plaintiffs might be able to make out a prima facie claim of disparate impact through statistical evidence. Francis R & R, at 40.

         At the second step, under McDonnell Douglas, the burden would "shift[ ] to the defendant[s] to demonstrate that the challenged practice or policy is ‘job related for the position in question and consistent with business necessity.’ " Id. at 299 (quoting 42 U.S.C. § 2000e-2(k)(1)(A)(i) ). Judge Francis found that, at this step, "Goldman Sachs would retain the right to demonstrate that there were other, legitimate explanations" for the disparate impact. Francis R & R, at 40. Specifically, Judge Francis reasoned that Defendants could successfully rebut the causation prong of a prima facie claim because "there are countless individualized factors that influence whether [the challenged processes] cause legally cognizable injury." Id. at 43. As a result, according to Judge Francis, "any causal relationship between those policies and apparent gender discrepancies is a complex issue, subject to individualized proof." Id. at 41.

          He concluded that Defendants’ introduction of individualized proof into the case "would effectively swamp the common question of whether the evaluative policies have, on average[,] a discriminatory impact." Id. at 40-41. In other words, under Judge Francis’ analysis, Defendants’ attack on Plaintiffs’ prima facie claims would shift the balance between common and individual factors. For the reasons explained below, the Court disagrees.

          Whether the challenged processes are job related or consistent with business necessity is a question of generalized proof. Defendants applied the 360 review and quartiling processes to each and every member of the class, and, as discussed above, Dr. Farber’s generalized statistical proof may well prove that the processes produce statistically significant gender disparities. Defendants’ rebuttal of that proof would require Defendants to explain the business necessity of the processes, and, therefore, require proof on a general level, rather than individualized proof on a claim-by-claim basis to support its company-wide pay, promotion, and performance evaluation systems. This was the fundamental error in Judge Francis’ recommendation on the question of predominance. If Plaintiffs’ statistical evidence makes out a prima facie claim of disparate impact, Defendants are tasked not with rebutting the causal link between the challenged processes and every single class member’s claim, but rather, establishing that the challenged processes were a business necessity. This is an issue capable of generalized proof.

         At the third step of a disparate impact case, "[t]he burden then shifts back to the plaintiff ‘to establish the availability of an alternative policy or practice that would also satisfy the asserted business necessity, but would do so without producing the disparate effect.’ " Id. The Court sees no reason, and the Defendants have presented none, why this step would not be subject to generalized proof as well. Accordingly, the Court concludes that, on the question of predominance for Plaintiffs’ disparate impact claims, certification would "achieve economies of time, effort, and expense, and promote uniformity of decision as to persons similarly situated, without sacrificing procedural fairness or bringing about other undesirable results." Cordes, 502 F.3d at 104 (internal alterations and quotation marks omitted). Plaintiffs’ objections to the R & R’s recommendation on predominance are, therefore, sustained as to its disparate impact claims.

          ii. Disparate Treatment

         Plaintiffs make two claims based on disparate treatment: First, they claim that their statistical evidence of disparate impact also demonstrates disparate treatment. Pls. Objs. at 52. Second, they claim that their anecdotal evidence of a "boy’s club" culture at Goldman Sachs demonstrates disparate treatment. Id. The Court will address each claim separately.

          1. Statistical Theory

          For the purposes of a disparate treatment pattern-or-practice claim, Plaintiffs make out a prima facie claim at the first step of McDonnell Douglas if they can show that (1) "discrimination was the company’s standard operating procedure[,] the regular rather than the unusual practice," and (2) "the discrimination was directed at a class of victims." City of New York, 717 F.3d at 83 (internal quotation marks omitted). For this showing, "instances of discrimination against particular employees are relevant to show a policy of intentional discrimination, [but] they are not required; a statistical showing of disparate impact might suffice." Id. at 84.

         At this first step, Plaintiffs’ statistical evidence may well be able to make out a prima facie showing of disparate treatment.¹⁶ As explained above, statistical evidence of disparate impact may also be used to prove disparate treatment. See City of New York, 717 F.3d at 88 ("The statistical disparities supporting the unchallenged finding ... [of] disparate impact also served to establish a prima facie case on the ... claim of a pervasive pattern of discriminatory treatment."). And as the Court already concluded, Plaintiffs’ statistical evidence of gender disparities is the type of generalized proof that satisfies the predominance requirement.

16 Plaintiffs’ style this claim as Goldman Sachs’ "aware[ness] of the discriminatory impact of its employment practices and continue[d] ... use" of the practices. Pls. Objs. at 52. Defendants oppose their claim by quoting the Supreme Court’s case in Feeney to argue only that " ‘awareness of consequences’ is insufficient to establish ‘discriminatory purpose’ " as a matter of law. Defs. Opp. Pls. Objs. 33 (quoting Pers. Adm’r of Mass. v. Feeney, 442 U.S. 256, 279, 99 S.Ct. 2282, 60 L.Ed.2d 870 (1979) ). Even if Feeney were applicable in the instant case on the question of whether awareness would prove "discrimination was the company’s standard operating procedure," which it is not, Feeney itself acknowledges that its statement on discriminatory purpose "is not to say that the inevitability or foreseeability of consequences of a neutral rule has no bearing upon the existence of discriminatory intent." Feeney, 442 U.S. at 279 n.25, 99 S.Ct. 2282. The only other case Defendants cite similarly acknowledges that "the failure to act [with awareness of the consequences] ... can support an inference of discriminatory intent in some circumstances." City of New York, 717 F.3d at 94. At this stage, it is not for the Court to say whether this is, indeed, such a case. All that matters is that Plaintiffs’ proof present common questions of law or fact that predominate over individual questions.

          At the second step of a pattern-or-practice claim, the burden "shifts to the employer ‘to rebut the presumption of discrimination.’ " City of New York, 717 F.3d at 84 (quoting Burdine, 450 U.S. at 254, 101 S.Ct. 1089). "The employer need only ‘articulate some legitimate, nondiscriminatory reason for’ " its presumptively discriminatory actions. Id. (quoting Burdine, 450 U.S. at 253, 101 S.Ct. 1089) (emphasis omitted). As the Second Circuit has recently explained, Defendants may rebut the presumption by attacking the accuracy or adequacy of Plaintiffs’ statistics, or, alternatively, "by accepting a plaintiff’s statistics and producing non-statistical evidence to show that it lacked such an intent." Id. at 85. Here, as with disparate impact, Defendants’ rebuttal evidence with respect to Plaintiffs’ statistics would be the subject of generalized proof.

         At the third step, if the Defendants do not successfully rebut Plaintiffs’ prima facie case, "the presumption arising from an unrebutted prima facie case entitles the plaintiff to prevail on the issue of liability and proceed directly to the issue of appropriate relief." Id. at 87. If Defendants do rebut, the presumption in favor of Plaintiffs "drops out" and the case proceeds to trial. Id. The Court sees no reason why trial (or the lack thereof) would alter the balance of the evidence any further. Accordingly, the Court finds that Plaintiffs have established predominance on their first claim of disparate treatment.

          2. "Boy’s Club" Theory

         At the first step of McDonnell Douglas, Plaintiffs’ "boy’s club" claim presents a closer question. Regardless of whether "generalized evidence [on the second claim] could be offered to prove th[e pattern-or-practice] elements" that make out a prima facie case, however, Nextel, 780 F.3d at 138 (internal quotation marks omitted), Defendants’ individualized proof on rebuttal at the second step would overwhelm the common issues.

         On rebuttal at the second step, unlike the disparate impact claim, the considerations and intentions of individual managers or business units may well be relevant to the question of intentional discrimination. City of New York, 717 F.3d at 85 (explaining defendants may "produc[e] non-statistical evidence to show that it lacked such an intent"). As such, the Court would likely need to make individualized inquiries into each incident of sexual assault, sexual harassment, stereotyping, impunity for male misconduct, and retaliation, to properly consider Goldman Sachs’ defenses. As a result, the balance between common and individual issues would shift. Individual issues would predominate.¹⁷ And the third step of the analysis would not alter that conclusion. The Court concludes, therefore, that Plaintiffs have not satisfied the predominance requirements on their "boy’s club" disparate treatment claim.

17 In their objections to Judge Francis’ R & R, Plaintiffs acknowledge that the burden-shifting standard applies to their disparate treatment claims, but fail to address whether, at the second step of the analysis, Defendants’ opportunity to rebut would shift the balance between generalized proof and individualized proof. Pls. Objs. at 52.

          3. Superiority

         Because Judge Francis concluded that Plaintiffs did not prove predominance, he did not reach the question of superiority. As the Court has concluded otherwise, it must now consider whether "a class action is superior to other available methods for fairly and efficiently adjudicating" Plaintiffs’ surviving disparate impact and disparate treatment claims. Fed.R.Civ.P. 23(b)(3). For the purposes of the superiority inquiry, courts analyze:

(1) the interest of the class members in maintaining separate actions; (2) "the extent and nature of any litigation concerning the controversy already commenced by or against members of the class"; (3) "the desirability or undesirability of concentrating the litigation of the claims in the particular forum"; and (4) "the difficulties likely to be encountered in the management of a class action."

In re Nassau Cty. Strip Search Cases, 461 F.3d 219, 230 (2d Cir. 2006) (quoting Fed.R.Civ.P. 23(b)(3) ). Each of these factors counsel in favor of class certification.

         For the first factor, it would be nonsensical to disaggregate the claims into hundreds or thousands of individual proceedings. Doing so would only waste "time, effort, and expense" and increase the likelihood of conflicting outcomes for Plaintiffs. Amchem, 521 U.S. at 634, 117 S.Ct. 2231. Second, the current litigation has already been ongoing for 8 years and the parties have already litigated numerous issues that would arise in any new case. In re Nassau Cty., 461 F.3d at 230 (explaining that the "action already has progressed substantially and ... offers the benefit of a liability phase that can be resolved quickly and conclusively"). Third, "because substantive Title VII law suggests that aggregate assessment and pro rata distribution is a fairer method of affording individual relief ..., it is clearly desirable to concentrate the litigation of the claims in this forum." Easterling v. Conn. Dep’t of Correction, 278 F.R.D. 41, 50 (D. Conn. 2011) (internal quotation marks omitted). Fourth, "failure to certify an action under Rule 23(b)(3) on the sole ground that it would be unmanageable is disfavored and should be the exception rather than the rule." In re Visa Check, 280 F.3d at 140 (internal quotation marks omitted). This is especially true given the "number of management tools available to a district court to address any individualized damages issues, such as bifurcation, the use of a magistrate or special master, alteration of the class definition, the creation of subclasses, or even decertification after a finding of liability." In re Nassau Cty., 461 F.3d at 231 (internal quotation marks omitted).

         Accordingly, the Court concludes that Plaintiffs have met the Rule 23(a) and Rule 23(b)(3) requirements for class certification on all their disparate impact and disparate treatment claims except their "boy’s club" claim.

          CONCLUSION

          For the reasons stated above, the parties’ objections to the M & O on expert testimony are OVERRULED. The parties’ objections to the R & R are SUSTAINED in part and OVERRULED in part. Plaintiffs’ motion to certify the class is GRANTED in part and DENIED in part. The Clerk of Court is directed to terminate the motions at ECF Nos. 501, 502, 503, 504, 505, 506, and 511.

          SO ORDERED.

a. Whether there is statistical evidence of discrimination in pay against women who are members of this class.

b. To the extent that I find statistical evidence of discrimination in pay against women who are members of the class, the extent to which these differences in pay can be explained by differences between how men and women far in Goldman’s performance review system.

c. Whether there is statistical evidence of a difference in how men and women fare in Goldman’s performance review system.

d. Whether there is statistical evidence of disparities between men and women in promotion rates from Vice President positions to positions as Managing Directors through 2008.

Farber Report, ¶ 6, ECF No. 259.

The Supreme Court also noted that the class members "held a multitude of different jobs[ ] at different levels of Wal-Mart’s hierarchy [and] for variable lengths of time" such that they had "little in common but their sex and this lawsuit." Id. at 359-60, 131 S.Ct. 2541 (quoting Dukes, 603 F.3d at 652 (Kozinski, J., dissenting) ). Not so here. Plaintiffs hold only one of two jobs: Associate or Vice President, and worked in one of the three revenue-generating divisions largely centered in one office in New York City: Investment Banking, Investment Management, or Securities, from 2002 to the present. These factual distinctions are critical, and place this case outside Dukes ’ ambit.