Whitaker v. 3M Co.

OPINION HUDSON, Judge. Appellant 3M Company challenges the district court’s certification of a class of more than 4,900 current and former 3M employees in this age-discrimination action brought under the Minnesota Human Rights Act (MHRA), Minn.Stat. §§ 363A.01-.41 (2008). Because we conclude that the district court did not properly apply the standards for certification under Minnesota Rule of Civil Procedure 23, we reverse and remand for further proceedings consistent with this opinion. FACTS Respondents, five current and former 3M employees, initiated this action against 3M in 2004, asserting that, since at least as early as 2001, 3M has engaged in a pattern and practice of age discrimination in employment in violation of the MHRA. See Minn.Stat. § 363A.08, subd. 2. Respondents alleged intentional discrimination in five areas: performance appraisals, selection for training programs, promotions, compensation, and terminations. In the alternative to their intentional-discrimination allegations, respondents asserted that 3M maintained neutral policies in these five areas that had a disparate impact on older employees. , See MinmStat. § 363A.28, subd. 10. Respondents asserted all of their claims on behalf of themselves and others similarly situated. In November 2007, respondents moved for class certification pursuant to rule 23 of the Minnesota Rules of Civil Procedure. In support of their motion for certification, respondents relied on the expert opinion of *634 Dr. Janet R. Thornton. Based on a combination of multiple regression, pools (or selection), and logistic regression analyses, Thornton opined that there were statistically significant disparities in the treatment of 3M employees under the age of 46 versus those 46 and older in each of the five areas. In conducting her analyses, Thornton controlled for various organizational and employee characteristics, including employee grade. In the four areas other than performance evaluations, Thornton conducted analyses both with and without performance-related controls, although none of her analyses regressed for all available performance controls. Thornton defended the exclusion of performance measures as a control because she found age disparities in those measures and concluded that they potentially were “tainted” by age discrimination. Thornton’s analyses were “cross-sectional” in that they examined employment events at certain points in time. 3M opposed respondents’ motion for class certification with its own experts. Dr. Bernard Siskin opined generally that age discrimination is difficult to identify statistically because of negative correlations within grade levels between age and “speed of success,” performance, and potential. In other words, Siskin opined that one would expect performance, potential, rate of promotion, and the rate of pay increases to decline as an employee ages, even absent any age discrimination. Sis-kin conducted an alternative, longitudinal analysis, which assessed trends in compensation and promotion rates, and concluded that those trends were consistent over time and thus inconsistent with the initiation of a pattern and practice of age discrimination beginning in 2001, as respondents alleged. Siskin further criticized Thornton’s cross-sectional analyses, opining that they fail to account for performance and related relevant variables. Siskin asserted that the disparities in performance measures observed by Thornton do not demonstrate age discrimination; that Thornton essentially assumed that they did; and that, based on that flawed assumption, Thornton excluded performance controls from most of her analyses and did not use all available measures in any of her analyses. Sis-kin conducted his own cross-sectional anal-yses controlling for available performance measures and found that disparities either ran in favor of older employees, or that there was no statistically significant disparity. 3M also submitted a report from labor economist Robert Topel, who opined that Thornton’s analyses revealed no more than expected and documented trends in the labor market and did not provide a basis for concluding that observed disparities were the result of age discrimination. Thornton responded to Siskin’s report, opining that his longitudinal analysis failed to control for factors that he had identified as key to the regression analyses and that his regression analyses included factors not supported by 3M documentation, and generally reinforcing her original report. Respondents also submitted a rebuttal report from David Neumark, another labor economist, who opined that Topel and Sis-kin overstated the extent to which established labor trends explained the statistical disparities observed by Thornton. The district court granted respondents’ class-certification motion. In its discussion regarding the commonality requirement of Minn. R. Civ. P. 23.01(b), the district court acknowledged the dispute among the experts regarding the significance of the various statistical analyses, but declined to resolve the dispute. The court explained that: [Respondents] have presented statistical evidence that strongly suggests a consistent pattern across 3M’s business units *635 of disparities suffered by older employees in each of the human resource practices challenged. Although 3M disputes the analysis conducted by [respondents’] expert, the court finds that sufficient statistical evidence has been presented to suggest that the data presents common questions for a class-wide pattern or practice trial. Finding that each of the requirements of rule 23.01 was met and that class treatment was appropriate under Minn. R. Civ. P. 23.02(b) and (c), the district court certified a class of “[a]ll persons who were 46 or older when employed by 3M in Minnesota in a salaried exempt position below PS grade 180 at any time on or after May 10, 2003, and who did not sign a [release].” 3M petitioned this court for discretionary review of the district court’s certification decision. We granted that petition based on our conclusion that the district court’s rule 23 analysis was questionable, in light of federal precedent; that review would serve judicial economy; and that the case presented issues of statewide significance such that a decision in this case could provide important guidance for the Minnesota bench and bar. ISSUES I. What standard of review applies to a district court decision on a motion for class certification? II. Did the district court err by failing to require respondents to prove the certification requirements of rule 23 by a preponderance of the evidence? III. Did the district court err by failing to resolve factual disputes among the parties’ experts? ANALYSIS I Class certification decisions by Minnesota district courts are governed by rule 23 of the Minnesota Rules of Civil Procedure. A certifiable class must meet all of the elements of rule 23.01 — numerosity, commonality, typicality, and adequacy of representation — and one of the sections of rule 23.02. Lewy 1990 Trust ex rel. Lewy v. Inv. Advisors, Inc., 650 N.W.2d 445, 451 (Minn.App.2002), review denied (Minn. Nov. 19, 2002). Because of the substantial similarity between Minnesota’s rule 23 and Fed.R.Civ.P. 23, “federal precedent is instructive in interpreting our rule.” Id. at 452; see also Minn. R. Civ. P. 23 2006 & 1968 advisory comm. cmt. (noting that Minn. R. Civ. P. 23 closely tracks Fed. R.Civ.P. 23 and is intended to produce consistent results). Although Minnesota courts have not expressly articulated a standard of review, it is clear that our review of the district court’s certification decision should be for abuse of discretion. See Gordon v. Microsoft Corp., 645 N.W.2d 393, 402 (Minn.2002) (holding that this court’s decision to grant discretionary review of class certification decisions should “take into consideration the district court’s discretion regarding class actions as well as the deferential appellate standard of review”); Peterson v. BASF Corp., 618 N.W.2d 821, 825 (Minn.App.2000) (acknowledging that “[district courts have considerable discretionary power to determine whether a class action is maintainable”), review denied (Minn. Jan. 26, 2001); see also In re St. Jude Med., Inc., 425 F.3d 1116, 1119 (8th Cir.2005) (expressly holding that certification decisions are reviewed for abuse of discretion). 3M resists application of the abuse-of-discretion standard, arguing that de novo review should apply because the district court’s certification order is based on *636 an erroneous interpretation of the law. However, as 3M acknowledges, it is an abuse of discretion when the district court improperly applies the law. See St. Jude, 425 F.3d at 1119; Shepard v. City of St. Paul, 380 N.W.2d 140, 143 (Minn.App.1985) (explaining that “[w]hen a district court errs as a matter of law in applying improper standards ... an abuse of discretion occurs” (quotation omitted)). We find guidance in one federal court’s description of the ways in which a district court can abuse its discretion in the class-certification context: An abuse occurs when a court, in making a discretionary ruling, relies upon an improper factor, omits consideration of a factor entitled to substantial weight, or mulls the correct mix of factors but makes a clear error of judgment in assaying them. An abuse of discretion also occurs if the court adopts an incorrect legal rule. In re New Motor Vehicles Canadian Exp. Antitrust Litig., 522 F.3d 6, 17 (1st Cir.2008) (quotation omitted); see also Shook v. El Paso County, 386 F.3d 963, 968 (10th Cir.2004) (“An abuse of discretion occurs where the district court misapplies the Rule 23 factors in deciding whether class certification is appropriate”). Thus, we proceed to examine the district court’s certification decision under an abuse-of-discretion standard, mindful of the various forms that such an abuse may take. II 3M asserts that the district court erred by failing to require proof of rule 23 certification requirements by a preponderance of the evidence and by failing to resolve factual disputes relevant to the class-certification requirements. While neither this court nor the Minnesota Supreme Court has addressed the standard of proof to be applied to rule 23 requirements, 3M relies on federal caselaw applying a preponderance standard. See, e.g., Teamsters Local 445 Freight Div. Pension Fund v. Bombardier, Inc., 546 F.3d 196, 202 (2nd Cir.2008). While few of the federal courts have expressly articulated a preponderance standard, recent federal decisions make clear that the prohibition against merits-related inquiries does not apply when class-certification issues overlap with the merits. See, e.g., In re Live Concert Antitrust Litig., 247 F.R.D. 98, 113-15 (C.D.Cal.2007) (summarizing case-law from Second, Third, Fourth, Fifth, Seventh, and Eighth Circuits requiring resolution of class issues that overlap with merits). Rather, the federal cases hold that district courts are required to resolve factual disputes relevant to rule 23 requirements — even if the same factual disputes are relevant to the merits of the case. See In re Initial Pub. Offerings Sec. Litig. (IPOS), 471 F.3d 24, 41 (2d Cir.2006) (requiring a “definitive assessment” of class-certification requirements, even if they coincide with merits issues); Blades v. Monsanto Co., 400 F.3d 562, 567 (8th Cir.2005) (explaining that “[t]he preliminary inquiry at the class certification stage may require the court to resolve disputes going to the factual setting of the case, and such disputes may overlap the merits of the case”); Gariety v. Grant Thornton, LLP, 368 F.3d 356, 366 (4th Cir.2004) (“[WJhile an evaluation of the merits to determine the strength of plaintiffs’ case is not part of a Rule 23 analysis, the factors spelled out in Rule 23 must be addressed through findings, even if they overlap with issues on the merits.”); Szabo v. Bridgeport Machs., Inc., 249 F.3d 672, 676 (7th Cir.2001) (“Before deciding whether to allow a case to proceed as a class action, therefore, a [district court] should make whatever factual and legal inquiries are necessary under Rule 23.”). *637 The federal courts have also held that the factual disputes requiring resolution by the district court at the class-certification stage can include disputes that are the subject of expert testimony. As the Third Circuit recently explained, “[e]xpert opinion with respect to class certification, like any matter relevant to a Rule 23 requirement, calls for rigorous analysis.” In re Hydrogen Peroxide Antitrust Litig. (Hydrogen Peroxide), 552 F.3d 305, 323, 2008 WL 5411562, at *12 (3d Cir. Dec.30, 2008) (citations omitted); see also id. (concluding that “[wjeighing conflicting expert testimony at the certification stage is not only permissible; it may be integral to the rigorous analysis Rule 23 demands”). Thus, the federal courts have increasingly held that a district court must go beyond a finding that each side’s experts presented credible evidence, requiring that disputes regarding expert evidence be addressed at the time of certification, rather than being deferred for trial. See id.; Blades, 400 F.3d at 575 (explaining that district court’s obligation to resolve factual disputes “extends to the resolution of expert disputes”); West v. Prudential Sec., Inc., 282 F.3d 935, 938 (7th Cir.2002). The Seventh Circuit has bluntly observed that failing to resolve expert disputes at the time of a certification application “amounts to a delegation of judicial power to the plaintiffs, who can obtain class certification just by hiring a competent expert.” West, 282 F.3d at 938. The court admonished that [a] district [court] may not duck hard questions by observing that each side has some support, or that considerations relevant to class certification also may affect the decision on the merits. Tough questions must be faced and squarely decided, if necessary by holding eviden-tiary hearings and choosing between competing perspectives. Id. These federal cases resolve a longstanding tension between class certification and the merits stemming from seemingly contradictory language in two U.S. Supreme Court decisions. In Eisen v. Carlisle & Jacquelin, the Court explained that “nothing in either the language or history of Rule 23 ... gives a court any authority to conduct a preliminary inquiry into the merits of a suit in order to determine whether it may be maintained as a class action.” 417 U.S. 156, 177, 94 S.Ct. 2140, 2152, 40 L.Ed.2d 732 (1974). But in General Telephone Co. v. Falcon, the Court explained that the class determination generally involves considerations that are enmeshed in the factual and legal issues comprising the plaintiffs cause of action. Sometimes the issues are plain enough from the pleadings to determine whether the interests of absent parties are fairly encompassed with the named plaintiffs claims, and sometimes it may be necessary for the court to probe behind the pleadings before coming to rest on the certification question.... [A]dual, not presumed, conformance with Rule 23(a) remains, however, indispensable. 457 U.S. 147, 160, 102 S.Ct. 2364, 2372, 72 L.Ed.2d 740 (1982) (quotation omitted) (emphasis added). The federal courts of appeal have reconciled Eisen and Falcon by reasoning that Eisen forbids merits inquiries that are unrelated to rule 23 requirements. See Hydrogen Peroxide, 552 F.3d at 315 & n. 17, 2008 WL 5411562, at *8 & n. 17 (stating distinction and citing supportive cases from the First, Second, Fourth, Fifth, and Seventh Circuits); see also IPOS, 471 F.3d at 33 (explaining that Eisen arose from district court’s determination that defendants should not bear cost of notifying class members unless plaintiff could show a probability of success on the merits). *638 The federal courts of appeal have addressed expert testimony related to class certification primarily in two contexts: securities litigation and antitrust litigation. In the securities context, the modern trend is to preclude certification of claims based on fraud on the market unless the district court concludes that the market is efficient — an issue that is typically the subject of competing expert opinions — and thus that the plaintiffs will be entitled to a presumption of reliance. See, e.g., IPOS, 471 F.3d at 42 (reversing certification order, rejecting district court’s application of “some showing” standard to determine whether market was efficient, and concluding that evidence could not support finding of efficient market); West, 282 F.3d at 938 (reversing certification order because evidence did not support application of fraud-on-the-market theory of reliance). The Second Circuit has taken this analysis one step further, holding that district courts must also determine, at the class-certification stage, whether the presumption of reliance has been rebutted. See In re Salomon Analyst Metromedia Litig., 544 F.3d 474, 485 (2d Cir.2008) (vacating certification order and remanding for district court’s consideration of defendant’s rebuttal evidence). In the antitrust context, the modern trend precludes certification unless the court concludes that the antitrust-impact element can be proven by common evidence, an issue that is usually the subject of expert testimony. See Hydrogen Peroxide, 552 F.3d at 327, 2008 WL 5411562, at *16 (vacating and remanding for district court to resolve evidentiary disputes related to whether antitrust impact could be proved by common evidence); In re New Motor Vehicles Canadian Export Antitrust Litig., 522 F.3d at 29-30 (same); cf. Blades, 400 F.3d at 572-75 (affirming district court’s denial of class certification based on evaluation of evidence and conclusion that class-wide antitrust impact could not be shown through common evidence). We find the federal caselaw persuasive and conclude that the certification requirements of Minnesota’s rule 23 — like those of its federal counterpart — must be established by a preponderance of the evidence. We further hold that the preponderance standard requires the district court to resolve factual disputes relevant to rule 23 certification requirements, including relevant expert disputes. We note that the factual findings at the class-certification stage are not binding on the ultimate trier of fact. Gariety, 368 F.3d at 366; IPOS, 471 F.3d at 41. Having identified the appropriate standards to be applied under rule 23, we next determine whether the district court’s decision complied with these standards in this ease. Ill Because the “nature of the evidence that will suffice to resolve a question determines whether the question is common or individual,” courts must analyze rule 23 certification requirements with specific reference to the cause of action asserted in a particular ease. Blades, 400 F.3d at 566; see also Hydrogen Peroxide, 552 F.3d at 311, 2008 WL 5411562, at *4 (stating that “we examine the elements of plaintiffs’ claim through the prism of Rule 23”) (quotation omitted). Respondents assert two theories of discrimination, each of which is proved using statistics. To prove their pattern-or-practice claim, a claim of intentional discrimination, respondents must show that age discrimination was 3M’s “standard operating procedure — the regular rather than the unusual practice.” Intern. Broth. of Teamsters v. United States, 431 U.S. 324, *639 336, 97 S.Ct. 1843, 1855, 52 L.Ed.2d 396 (1977). Such a claim is typically proved through a combination of statistics and anecdotal evidence. See, e.g., EEOC v. Dial Corp., 469 F.3d 735, 741 (8th Cir.2006). In contrast, to prove their disparate-impact claim, respondents need not prove intent, but rather must identify the existence of a neutral practice that has “a statistically significant adverse impact” on older employees and, if the employer demonstrates that the practice “is manifestly related to the job or significantly furthers an important business purpose,” that an alternative practice would significantly lessen the disparate impact. See Minn.Stat. § 363A.28, subd. 10 (emphasis added); see also Schlemmer v. Farmers Union Cent. Exch., Inc., 397 N.W.2d 903, 908 (Minn.App.1986) (outlining elements of claim). Statistics are relevant not only to the merits of respondents’ claims, but to certification as well. This is because, in order to proceed as a class, respondents must bridge the gap between their individual claims of discrimination and the existence of a class of similarly situated employees. See Falcon, 457 U.S. at 157-58, 102 S.Ct. at 2370-71. An appropriate statistical analysis may provide the bridge required by Falcon by showing that employees were subject to a common discriminatory practice. We conclude that the district court erred in its certification analysis, both by failing to require proof of rule 23 certification requirements by a preponderance of the evidence and by failing to resolve factual disputes relevant to class-certification requirements. While the district court considered respondents’ statistical analy-ses as relevant to the class-certification decision, it declined to address 3M’s objections to those statistics or 3M’s alternative statistical analyses. The standards that we adopt from the federal caselaw require the district court to resolve the differences among the experts to the extent that they are relevant to, and for the limited purposes of, determining whether class-certification requirements were met. In particular, the district court, using the preponderance-of-the-evidence standard, must address and decide all of the alleged defects in respondents’ expert-witness testimony relating to the rule 23 requirements of numerosity, commonality, typicality, and predominance. This means that the district court must determine the validity and weight of 3M’s concerns with demonstrating a baseline of expected, non-discriminating differences, bridging the gap between individual claims of discrimination and the existence of a class of similarly situated employees, selected problems with doing a “snap-shot” analysis, and the alleged inappropriate use of certain statistical controls for the predicted non-discriminatory correlation between age and employment outcomes. Respondents assert that the holdings in the antitrust and securities cases should not apply in the context of their age-discrimination claims, attempting to distinguish the former as cases involving legal issues rather than factual disputes. See Hnot v. Willis Group Holdings Ltd., 241 F.R.D. 204, 210-11 & n. 3 (S.D.N.Y.2007) (drawing distinction between claims). But we are unable to discern such a distinction. Rather, the determination of an efficient market, while subject to some legal rules, typically revolves around hotly contested factual disputes about the particular market in question. As in employment-discrimination litigation, parties commonly submit expert reports supporting their characterization of the underlying facts. See, e.g., Bombardier, 546 F.3d at 210 (discussing competing expert opinions regarding efficiency of market). The same is true in the antitrust context with respect *640 to the existence of a common impact or injury. See, e.g., Hydrogen Peroxide, 552 F.3d at 312-15, 2008 WL 5411562, at *5-7 (discussing competing expert opinions regarding anti-trust impact). Respondents further urge this court to adopt the analysis of the federal district court in Hnot, that the determination of the existence of common questions can be separated from finding the answer to those questions. See 241 F.R.D. at 211. But Hnot suggests that the district court needs to examine the statistical proof only insofar as necessary to make sure that the plaintiffs will present their claims in a unified manner. See id. Put another way, the plaintiffs must put forth a prima facie statistical case. This is precisely the standard that has been rejected by the federal courts of appeal. See IPOS, 471 F.3d at 42 (rejecting standard allowing certification as long as expert report was not “fatally flawed”); see also Hydrogen Peroxide, 552 F.3d at 321, 2008 WL 5411562, at *11 (“It is incorrect to state that a plaintiff need only demonstrate an ‘intention’ to try the case in a manner that satisfies the predominance requirement”). In sum, we conclude that parties moving for class certification under Minn. R. Civ. P. 23 must prove, by a preponderance of the evidence, that the certification requirements of the rule are met. This means that district courts must address and resolve factual disputes relevant to class-certification requirements, including disputes among expert witnesses. Because the district court failed to apply the appropriate standards, we reverse and remand for the district court’s reconsideration of the class-certification motion. 1 On remand, the district court, in its discretion, may elect to re-open the record and accept additional evidence. See IPOS, 471 F.3d at 41 (explaining that the district court “must receive enough evidence, by affidavits, documents, or testimony, to be satisfied that each Rule 23 requirement has been met”). But the district court also “has ample discretion to circumscribe ... the extent of a hearing to determine whether such requirements are met in order to assure that a class certification motion does not become a pretext for a partial trial of the merits.” Id. DECISION Because the district court failed to require proof of rule 23 certification requirements by a preponderance of the evidence and failed to resolve factual disputes relevant to rule 23 certification requirements, we reverse and remand for further proceedings consistent with this opinion. Reversed and remanded. Footnotes [1] 3M urges this court to hold that certification is not available under rule 23 on the facts of this case. We decline to do so because our role as an error-correcting court does not extend to making factual findings in the first instance. See Sefkow v. Sefkow, 427 N.W.2d 203, 210 (Minn.1988) (holding that court of appeals exceeds its scope of review by "reweighing the evidence and finding its own facts”).