Equal Employment Opportunity Commission et al v. DHL Express (USA), Inc., et al, No. 1:2010cv06139 - Document 254 (N.D. Ill. 2016)
Court Description: MEMORANDUM Opinion and Order Signed by the Honorable John Z. Lee on 9/30/16. Mailed notice(ca, ).
Download PDF
<![CDATA[
Equal Employment Opportunity Commission et al v. DHL Express (USA), Inc., et al Doc. 254 IN THE UNITED STATES DISTRICT COURT FOR THE NORTHERN DISTRICT OF ILLINOIS EASTERN DIVISION EQUAL EMPLOYMENT OPPORTUNITY COMMISSION, Plaintiff, and REGINALD BAILEY, KENNETH BRISCO, OLIVER DEAN, MELVIN EDWARDS, JOHN ELLIS, RONNIE FORD, BENITA GREEN-RILEY, MICHAEL JOHNSON, ANTHONY JORDAN, MIRANDA LESTER, SANDRA McNEELY, EDGAR MEDLEY, TIMOTHY PRICE, ALONZO STUDSTILL, PAUL THOMAS, RANDY THOMPSON, SHREE WASHINGTON, GEORGE WHITE, and SANDRA WILLIAMS, Intervening-Plaintiffs, v. DHL EXPRESS (USA), INC., Defendant. ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) No. 10 C 6139 Judge John Z. Lee MEMORANDUM OPINION AND ORDER The United States Equal Employment Opportunity Commission (EEOC) has sued international shipping company, DHL Express USA, Inc., on behalf of ninety-four African American drivers, for discriminating against them based on race in violation of Title VII of the Civil Rights Act of 1964, as amended, 42 U.S.C. § 2000e et seq. and Title I of the Civil Rights Act of 1991, 42 U.S.C. § 1981a. In short, the EEOC asserts that DHL used race to assign less desirable delivery routes to black drivers. DHL denies that this is so. Dockets.Justia.com During the course of this litigation, both sides have presented experts to analyze the route assignment data maintained by DHL, and now both sides seek to bar the opponent’s expert on numerous grounds. DHL moves to bar the EEOC’s expert, Dr. Thomas DiPrete. In turn, the EEOC moves to bar DHL’s experts, Dr. James Langenfeld and D. Jan Duffy. For the reasons provided herein, the Court denies DHL’s motion to bar DiPrete and denies the EEOC’s motions to bar Langenfeld and Duffy. Legal Standard District courts have broad discretion to rule on evidentiary issues prior to trial. See United States v. Chambers, 642 F.3d 588, 594 (7th Cir. 2011). The admissibility of expert testimony is governed by Federal Rule of Evidence 702 and the Supreme Court’s seminal case, Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993). See United States v. Parra, 402 F.3d 752, 758 (7th Cir. 2005) (“At this point, Rule 702 has superseded Daubert, but the standard of review that was established for Daubert challenges is still appropriate.”). By its terms, Rule 702 allows the admission of testimony by an expert, someone with the requisite “knowledge, skill, experience, training, or education[,]” to help the trier of fact “understand the evidence or to determine a fact in issue.” Fed. R. Evid. 702. Experts are permitted to testify when their testimony is: (1) “based upon sufficient facts or data”; (2) “the testimony is the product of reliable principles and methods”; and (3) “the expert has reliably applied the principles and methods to the facts of the case.” Id. Daubert requires the district court to act as the evidentiary gatekeeper, ensuring that Rule 702’s requirements of reliability and relevance are satisfied before allowing the finder of fact to hear the testimony of a proffered expert. See Daubert, 509 U.S. at 589; see also Kumho Tire Co. v. Carmichael, 526 U.S. 137, 147–49 (1999). 2 District courts have broad discretion in determining the admissibility of expert testimony. See Gen. Elec. Co. v. Joiner, 522 U.S. 136, 142 (1997); Lapsley v. Xtek, Inc., 689 F.3d 802, 810 (7th Cir. 2012) (“[W]e ‘give the district court wide latitude in performing its gatekeeping function and determining both how to measure the reliability of expert testimony and whether the testimony itself is reliable.’”) (quoting Bielskis v. Louisville Ladder, Inc., 663 F.3d 887, 894 (7th Cir. 2011)). And the proponent of the expert bears the burden of demonstrating that the expert’s testimony would satisfy the Daubert standard by a preponderance of the evidence. Lewis v. CITGO Petroleum Corp., 561 F.3d 698, 705 (7th Cir. 2009). Analysis I. Dr. Thomas DiPrete The EEOC’s expert, Dr. Thomas DiPrete, is a sociology professor at Columbia University. DiPrete’s task was “to determine whether black DHL drivers were more likely than white drivers to drive routes in predominantly black neighborhoods and to drive routes that were ‘less desirable, more difficult, and/or more dangerous.’” Def.’s Mem. Supp. Mot. Strike, Def.’s Ex. 10, DiPrete Rep. at 1. After using regression analysis to study staffing data and pick up delivery data for the delivery areas covered by the DHL stations located in Lisle, Alsip, and Franklin Park, DiPrete concluded that, in general, “black drivers from the [stations] were more likely than white drivers to pick up or deliver packages in neighborhoods that were more black, more non-white, and with higher rates of violent and property crime.” Id. at 2. DHL attacks DiPrete’s opinions on three fronts. First, DHL argues that DiPrete’s opinion is irrelevant to the issues in the case. Second, DHL contends that the regression methodology that DiPrete employed was not reliable or probative. Third, DHL asserts that DiPrete’s opinion improperly relies on the opinions of other experts, whom the EEOC has failed to disclose as Fed. R. Civ. P. 26(a)(2) requires. 3 A. Relevance DHL argues that DiPrete’s opinion cannot assist the jury because his regression analysis does not prove that any individual driver was disadvantaged in their route assignments or was intentionally discriminated against by any individual station supervisor. In response, the EEOC counters that DiPrete’s multiple regression analyses is relevant because he concludes that for the three DHL stations at issue, there is a correlation between the drivers’ race and their assignment to “less desirable” delivery routes. The EEOC’s theory is that DHL intentionally shunted black drivers to neighborhoods that were more dangerous, poorer, and predominantly black in comparison to areas to which white drivers were assigned. According to the EEOC, this practice caused black drivers to work in conditions that objectively created hardship by subjecting them to an environment that was humiliating, degrading, and unsafe. See Tart v. Ill. Power Co., 366 F.3d 461, 475 (7th Cir. 2004). To help prove this, DiPrete has analyzed the crime rates of the delivery neighborhoods by zip code. He also has analyzed the rates of poverty and percentage of black residents for each zip code area. The EEOC concedes that additional evidence beyond DiPrete’s expert report will be necessary to prove that the assignment of drivers to these neighborhoods constituted materially adverse employment actions. But a brick is not a wall, as they say, 1 and the EEOC is not required to rely solely on DiPrete to prove their entire case. DHL’s objection to the relevance of DiPrete’s testimony boils down to this. DHL argues that studying the aggregate effect of its policies does not prove discriminatory intent, because it says nothing about whether a particular driver experienced discriminatory route assignments 1 See United States v. Porter, 881 F.2d 878, 887 (10th Cir. 1989) (quoting McCormick on Evidence § 285, at 542–43 (E. Cleary 3d ed. 1984) (“‘An item of evidence, being but a single link in the chain of proof, need not prove conclusively the proposition for which it is offered. . . . It is enough if the item could reasonably show that a fact is slightly more probable than it would appear without that evidence. . . . A brick is not a wall.’”). 4 from a particular supervisor. The EEOC responds that the use of regression analysis to help prove intentional discrimination is well-accepted in disparate treatment cases and is especially useful where, as here, the various factors used by supervisors to determine the working conditions of a group of employees are unknown. The Seventh Circuit, in Adams v. Ameritech Services, Inc., 231 F.3d 414, 425 (7th Cir. 2000), considered this recurrent debate about the probity of statistical evidence in discrimination cases: [W]hat is the proper level of aggregation or disaggregation at which [defendant’s] actions should be assessed? At one extreme, one could perhaps look at the [defendant’s] entire workforces, management and non-management alike; at the other extreme, one could take a highly individualistic view of humanity and conclude that no two people are exactly alike and statistics are therefore worthless. Neither approach has much to recommend it, of course, but the thought experiment suggests the outer possibilities. Id. Nonetheless, the Seventh Circuit has repeatedly held that statistical evidence, including regression analysis, may be used to demonstrate discrimination in disparate treatment cases. See id. at 417 (reversing district court’s bar of plaintiffs’ statistical expert on summary judgment in disparate treatment case); EEOC v. Sears, Roebuck & Co., 839 F.2d 302, 324 n.22 (7th Cir. 1988) (“Multiple regression analyses, designed to determine the effect of several independent variables on a dependent variable, . . . are an accepted and common method of proving disparate treatment claims.”); Mister v. Ill. Cent. Gulf R.R. Co., 832 F.2d 1427, 1430–31 (7th Cir. 1987) (reversing grant of summary judgment in a disparate treatment case where defendant failed to rebut the plaintiffs’ statistical showing that the defendant hired a much larger proportion of white than black applicants). For example, in Adams, the Seventh Circuit held that an expert’s statistical analysis was helpful even when the expert merely concluded that the correlation between an employee’s age 5 and the employer’s decision to terminate was unlikely to have occurred by chance. 231 F.3d at 425 (holding that, to be relevant, the statistical analysis “need only make the existence of ‘any fact that is of consequence’ more or less probable”). As in Adams, DiPrete offers his opinion that it is highly unlikely that the correlation between a driver’s race and assignment to a driving route in a predominantly black, higher-poverty, higher-crime neighborhood occurred by chance. As a result, the Court concludes that DiPrete’s opinion will aid the jury in its task of determining whether DHL intentionally discriminated against black drivers and denies DHL’s motion to bar DiPrete’s testimony on this ground. B. Reliability Next, DHL argues that the Court should bar DiPrete from testifying because his analysis is not reliable for three reasons. First, according to DHL, DiPrete’s regression analysis fails to control for significant variables and utilizes control variables that in fact have no effect on route assignments. Second, DHL contends that the analysis has extraordinarily low explanatory power, as indicated by the low R-squared (R2) values. And, third, DHL argues that DiPrete’s conclusions are unreliable and not sufficiently robust, because they are overly sensitive to even small changes in the input data. DHL first contends that DiPrete’s regression analysis failed to account for significant variables that affected the assignment of routes, such as a driver’s preference and route familiarity. In addition, DHL claims that the control variables that DiPrete used, such as driver seniority, had no effect on the outcome and, thus, did not control for other potential influences on route assignments, other than race. “Regression analysis permits the comparison between an outcome (called the dependent variable) and one or more factors (called independent variables) that may be related to that outcome.” Manpower, Inc. v. Ins. Co. of Pa., 732 F.3d 796, 806 (7th 6 Cir. 2013). Accordingly, DHL is correct to point out that “the choice of independent variables to include in any regression analysis is critical to the probative value of that analysis.” Id. at 808. But, “the Supreme Court and this Circuit have confirmed on a number of occasions that the selection of the variables to include in a regression analysis is normally a question that goes to the probative weight of the analysis rather than to its admissibility.” Id. at 808. The Court finds that such is the situation here. Second, DHL posits that DiPrete’s analysis has extraordinarily low explanatory power, as indicated by the R-squared values. R-squared is “a statistic that measures the percentage of variation in the dependent variable that is accounted for by all the explanatory variables.” Daniel Rubinfeld, Reference Guide on Multiple Regression, in Reference Manual on Scientific Evidence 345 (Federal Judiciary Center, 3d ed. 2011). Thus, the R-squared “provides a measure of the overall goodness of fit of the multiple regression equation.” Id. The R-squared value ranges from 0 to 1. A value of 0 means that “the explanatory variables [for example, a driver’s race] explain none of the variation of the dependent variable [for example, the route to which a driver is assigned], while a R-square of 1 means that “the explanatory variables explain all of the variation.” Id.; see Sanner v. Bd. of Trade of City of Chi., No. 89 C 8467, 2001 WL 1155277, at *4 (N.D. Ill. Sept. 28, 2001) (“[I]f a dependent variable perfectly explained what was being analyzed, the R-squared would be 1. On the other hand, if a dependent variable only explained what was being measured 25% of the time, the R-squared would be .25.”). The degree to which an R-squared value reflects the reliability of the overall regression analysis, however, is unclear. This is because “the magnitude of R-squared depends on the characteristics of the data being studied and, in particular, whether the data vary over time or over individuals.” Rubinfeld at 345. “Typically, an [R-squared] is low in cross-sectional studies 7 in which differences in individual behavior are explained.” Id. For this reason, although “the explanatory power of a regression model is clearly relevant to the validity of the model,” Griffin v. Bd. of Regents of Regency Univs., 795 F.2d 1281, 1292 n.23 (7th Cir. 1986), “courts should be reluctant to rely solely on a statistic such as R-squared.” Rubinfeld at 345. This would appear to be the case here, where DiPrete performed a cross-sectional study and the regression analysis performed by DHL’s own expert, Langenfeld, also produced similarly low R2 values.2 The case on which DHL relies, Griffin, 795 F.2d 1281, is distinguishable. In Griffin, the Seventh Circuit rejected the regression analysis in question for two primary reasons. First, the court was “reluctant to rely on a single regression when other regressions could have been presented” to aid their determination. Griffin, 795 F.2d at 1292 n.23. The court also was concerned with the lack of proven instances of discrimination to support the disparate treatment claim, stating that “the lack of such proof reinforces the doubt arising from the questions about the validity of the statistical evidence.” Id. In contrast, here, the EEOC intends to present testimony from individual employees to prove specific instances of race discrimination at the three DHL stations. Furthermore, rather than relying on a single regression analysis, DiPrete performed multiple regression analyses to support his conclusions. Next, DHL argues that DiPrete’s analysis is unreliable because removing just a few black drivers from his model would produce results that are not statistically significant. Plaintiffs counter that DHL’s argument is based on the removal of critical data points, not simply outliers. 2 Furthermore, another widely-used (and some courts would say, more accurate) measure of the reliability of a regression model is the “t-statistic.” See Lyman v. Cardiostat Med. LLC v. St. Jude Med. S.C., Inc., 580 F. Supp. 2d 719, 725 (E.D. Wis. 2008) (finding that “use of the t-statistic is a better measure than R2 to determine the reliability of a regression model”). A t-statistic is a measure of standard deviation in which any departure of two or more standard deviations is viewed as significant. See Dicker v. Allstate Life Ins. Co., No. 89 C 4982, 1997 WL 182290, at *8 (N.D. Ill. Apr. 9, 1997). According to DiPrete’s deposition testimony (which DHL does not contest), a regression analysis may show significant t-statistics with low R2 values. See Pl.’s Ex E, Diprete Dep. at 173–77. 8 The Court agrees. It seems only logical that removing from DiPrete’s analysis a number of black drivers, who were assigned predominantly to the particular neighborhoods in question, would impact the statistical significance of the model. In essence, DHL tries to demonstrate the unreliability of DiPrete’s model by removing from it some of the central data points that his analysis is intended to study. It is difficult to understand how these data could be outliers, and the Court will not bar DiPrete’s testimony on this basis. C. Reliance on Undisclosed Experts Lastly, DHL argues that DiPrete should be barred from testifying because his models rely on the geocoding performed by James Quinn and Andrew Rundle of Geographic Information Systems (collectively “GIS”). In support, DHL relies on Dura Automotive Systems of Indiana, Inc. v. CTS Corp. 285 F.3d 609 (7th Cir. 2002). In Dura, the plaintiff presented an expert to testify about historical groundwater patterns at a particular site. The expert was a recognized hydrogeologist, but he admitted that he had no expertise in mathematical modeling of groundwater flow and that he had relied on the results of such a model in arriving at his opinions. The defendant sought to bar the hydrogeologist’s testimony, arguing that the plaintiff was required to disclose the individuals who created and performed the groundwater model. The district court agreed, and the Seventh Circuit affirmed. The Seventh Circuit started by acknowledging the general rule that “[a]n expert witness is permitted to use assistants in formulating his expert opinion, and normally they need not themselves testify.” Id. at 613. “Analysis becomes more complicated,” continued the court, “if the assistants aren’t merely gofers or data gatherers but exercise professional judgment that is beyond the expert’s ken.” Id. The court recognized that “it is common in technical fields for an expert to base an opinion in part on what a different expert believes on the basis of expert 9 knowledge not possessed by the first expert; and it is apparent from the wording of Rule 703 that there is no general requirement that the other expert testify as well.” Id. Along these lines, the court agreed that it is not the case that “the leader of a clinical medical team must be qualified as an expert in every individual discipline encompassed by the team in order to testify as to the team's conclusions.” Id. But, circumstances are different where “the soundness of the underlying expert judgment is in issue.” Id. This is more of a continuum, than a bright line, and the key is whether the disclosed expert, using inputs that are generally relied upon by other experts in the field, is offering opinions within his or her expertise, or merely serving as “the mouthpiece of a scientist in a different specialty.” Id. In Dura, the court found that the individuals who constructed and performed the groundwater modeling “did not merely collect data for [the hydrogeologist] to massage or apply concededly appropriate techniques in a concededly appropriate manner, or otherwise perform routine procedures,” id. at 615, but exercised a substantial degree of technical judgment critically relevant to the central contested issues in the case. Based upon the record presented here, the Court does not believe that the geocoding work performed by GIS crosses this line. GIS was tasked with converting the street addresses of the delivery points provided by DHL into X-Y coordinates. It did so in three ways. See DiPrete’s Report, Ex. 8-A, at 6. First, GIS used an “address point locator” (a commonly used locator program) to match each address to a database containing the X-Y coordinates of more than 54 million residential and commercial addresses. Id. Second, where the address locator did not produce an X-Y coordinate, GIS utilized a “street ranges locator,” which tried to interpolate an X-Y coordinate for an address on a particular block between two streets. Id. Third, where no information was available other than a zip code, GIS used a zip code locator to determine the X- 10 Y coordinate at the center of the zip code area. Id. And, once this was accomplished, GIS employed routine procedures to ensure the accuracy of the X-Y coordinate results. 3 What is more, all of the materials related to the geocoding work were produced to DHL as part of expert discovery, and lengthy reports describing the work were provided as attachments to DiPrete’s expert report. 4 After reviewing the present record, the Court finds that the geocoding work performed by GIS does not raise the concerns voiced by the court in Dura. Additionally, Federal Rule of Evidence 702 provides that “[a]n expert may base an opinion on facts or data in the case that the expert has been made aware of or personally observed. If experts in the particular field would reasonably rely on those kinds of facts or data in forming an opinion on the subject, they need not be admissible for the opinion to be admitted.” Geocoding data is the kind of data upon which statistical analysts can reasonably rely. See, e.g., Am. Honda Motor Co., Inc. v. Bernardi’s Inc., 188 F. Supp. 2d 27, 32 (D. Mass. 2002) (rejecting argument that geocoding is insufficiently developed and an unacceptable method of determining market area); Ohio Organizing Collaborative v. Husted, No. 2:15-CV1802, 2016 WL 3248030, at *7 (S.D. Ohio May 24, 2016) (analyst provided with a geocoded voter file). This is particularly true in this case, where DHL has not identified any irregularities or deficiencies in GIS’s work and the parties have jointly retained the services of another geocoding company, Location, Inc., to supply data to their experts. 3 For example, GIS discarded any X-Y coordinate produced solely by the zip code locator because it could not pinpoint the geographic location of an address. See id. at 7. GIS also examined how well the address found in the reference data matched the actual address data being searched. And GIS disregarded any address that matched two X-Y coordinates. 4 Despite receiving extensive information about the geocoding project during expert discovery, DHL has not articulated in its briefs any objections to or deficiencies in the data or methodology employed by GIS. For its part, DHL suggests that it needs additional information from Rundle in order to assess the validity of GIS’s work. Given the extensive materials already produced to DHL, this argument is unpersuasive. 11 For these reasons, the Court denies DHL’s motion to exclude the testimony of Dr. Thomas DiPrete. II. Dr. James Langenfeld After reviewing DiPrete’s regression analysis and conducting his own, Dr. James Langenfeld, a trained economist, offered his own opinions attesting to the unreliability, inaccuracy, and limitations of DiPrete’s regression model. The EEOC challenges various portions of Langenfeld’s expected testimony on the grounds that they are not based on reliable principles and methods, will only confuse or mislead the jury, and are not relevant to the issues in this case. A. Reliable Principles and Methods First, the EEOC attacks Langenfeld’s conclusion that some of DiPrete’s results are not “economically significant.” See Langenfeld Report at 15, 18. The EEOC asserts that this conclusion is not based on reliable principles or methodologies because Langenfeld himself concedes that there is no bright line rule to measure economic significance and he has not performed any of his own tests as to this issue. See Langenfeld Dep. at 188, 199. As applied to regression models, economic significance assesses whether the association between an independent variable and the dependent variable “is causal, follows theoretical expectations in terms of the direction (sign) and size of the association, and is large enough to matter in its real world context.” See Jane E. Miller & Yana van der Meulen Rodgers, Economic Importance and Statistical Significance: Guidelines for Communicating Empirical Research, 14 Feminist Econ. 117, 120 (2008) (cited in Langenfeld Report at 15 n.48). Economic significance is a recognized tool to evaluate the results of multiple regression analyses. See id.; Arthur S. Goldberger, A Course in Econometrics 122–23 (Harvard Univ. Press 2003) (discussing statistical versus economic significance of coefficient estimates); see also Def.’s Opp’n Mot. Exclude 12 Langenfeld, Ex. 2, Deirdre N. McCloskey & Stephen T. Ziliak, The Standard Error of Regression, Journal of Economic Literature, March 1996 (criticizing econometricians for failing to distinguish between economic significance and statistical significance). Here, Langenfeld questions whether the association found in DiPrete’s analysis between certain independent variables and the dependent variable is large enough to matter in its real world context. The EEOC is correct to point out that there are different ways to evaluate economic significance and Langenfeld has not conducted his own tests, but this does not mean that Langenfeld’s testimony is the product of unreliable principles and methods. Langenfeld may properly describe the concept of economic significance to the jury based on the common understanding of that term in the field of economics and statistics. In addition, Langenfeld may offer his critique of DiPrete’s analysis for not having considered whether the results have economic significance and why it should have done so. However, because Langenfeld has not conducted his own study of the data in terms of economic significance or offered any benchmarks to which he assessed the economic significance of DiPrete’s results, Langenfeld may not offer an opinion that DiPrete’s results, in fact, lack economic significance. Second, the EEOC urges the Court to jettison Langenfeld’s sensitivity analysis because it is based upon unreliable and cherry-picked data. However, sensitivity analysis is a well- accepted method of determining the reliability of a regression model. See Mohan P. Rao & Christian D. Tregillis, Econometric Analysis, Litigation Services Handbook 6.11 (Roman L. Weil et al. eds., 4th ed. 2007). To the extent that the EEOC takes issue with the inputs that Langenfeld employed in his sensitivity analysis, the EEOC can cross-examine him at trial. 13 B. Confusing or Misleading Testimony The EEOC also argues that portions of Langenfeld’s testimony would only confuse or mislead the jury. Specifically, it takes issue with Langenfeld’s criticism of DiPrete for not offering any opinions as to causation, for failing to control for factors, other than race, and for not accounting for pick-up times. But, such opinions are suitable rebuttal topics, and the Court finds little danger that the jury will be confused or misled by them. To the extent that the EEOC finds them to be so, it can address these issues upon cross-examination and seek to clarify them using their own evidence and the testimony of DiPrete. C. Relevance In addition, the EEOC argues that two of Langenfeld’s conclusions have no relevance to the issues the jury must decide. The EEOC first asserts that Langenfeld’s comparison of the routes of claimant black drivers to the routes of non-claimant black drivers is irrelevant. The comparison shows that some non-claimant black drivers were not assigned to delivery routes in primarily black, higher-poverty, higher-crime neighborhoods. But this portion of Langenfeld’s analysis is probative to show that DHL did not assign routes on the basis of race and, therefore, is relevant to the issue at hand. Next, the EEOC contends that Langenfeld’s criticism of DiPrete’s failure to compare the crime rates in neighborhoods on the assigned routes to those in large cities is irrelevant. But Langenfeld’s opinions are relevant to DHL’s efforts to show that DiPrete should have considered the degree to which the crime rates differed between the various routes (assuming that they did). As such, the testimony in question is relevant to the jury’s inquiry as to whether the assignment of black drivers to routes with higher crime rates presented an objectively more dangerous working condition. 14 For these reasons, the Court denies the EEOC’s motion to exclude the testimony of Dr. James Langenfeld. III. D. Jan Duffy In response to the EEOC’s claim for punitive damages, DHL asserts that, assuming for the sake of argument that the EEOC prevails in this case, DHL is not liable for punitive damages, because it made a good-faith effort to implement its antidiscrimination policies, citing Kolstad v. American Dental Association, 527 U.S. 526, 529–30 (1999) (“[I]n the punitive damages context, an employer may not be vicariously liable for the discriminatory employment decisions of managerial agents where these decisions are contrary to the employer’s good-faith efforts to comply with Title VII.”). 5 See Bruso v. United Airlines, Inc., 239 F.3d 848, 860–61 (7th Cir. 2001); Hertzberg v. SRAM Corp., 261 F.3d 651, 663–64 (7th Cir. 2001). 5 As explained in Seventh Circuit explained in E.E.O.C. v. Autozone, 707 F.3d 824, 835 (7th Cir. 2013) (citations omitted): In Kolstad v. American Dental Ass’n, the Supreme Court established a three-part framework to determine whether punitive damages are proper . . . . First, the plaintiff must show that the employer acted with “malice” or “reckless indifference” toward the employee's rights under federal law. A plaintiff “may satisfy this element by demonstrating that the relevant individuals knew of or were familiar with the anti-discrimination laws” but nonetheless ignored them or lied about their discriminatory activities. The plaintiff has the burden of proving “malice” or “reckless indifference” by a preponderance of the evidence. Second, the plaintiff must establish a basis for imputing liability to the employer based on agency principles. Employers can be liable for the acts of their agents when the employer authorizes or ratifies a discriminatory act, the employer recklessly employs an unfit agent, or the agent commits a discriminatory act while “employed in a managerial capacity and . . . acting in the scope of employment.” Third, when a plaintiff imputes liability to the employer through an agent working in a “managerial capacity . . . in the scope of employment,” the employer has the opportunity to avoid liability for punitive damages by showing that it engaged in good-faith efforts to implement an anti-discrimination policy. 15 As an initial matter, based upon DHL’s reliance on the Kolstad defense, the Court finds it appropriate to bifurcate the trial in this case. The jury will first be tasked with determining liability and, if appropriate, compensatory damages. If the jury finds DHL liable and awards compensatory damages, the Court will then allow the parties to present evidence pertinent to the Kolstad defense to enable the same jury to determine whether punitive damages are appropriate. Cf. EEOC v. Global Horizons, Inc., No.: CV-11-3045-EFS, 2014 WL 11429301, at *3 (E.D. Wash. Apr. 9, 2014) (bifurcating liability phase of Title VII bench trial seeking injunctive relief from jury trial on punitive damages where defendant raised Kolstad defense). Bifurcating the trial will be the most straightforward and efficient solution to avoid prejudice and jury confusion. Turning to the EEOC’s motion, to support its Kolstad defense, DHL has presented the testimony of D. Jan Duffy. Duffy is a management practices and compliance consultant, who has thirty-nine years of experience as an attorney, workplace investigator, and consultant for private and public employers regarding labor relations, employment law, employee rights and responsibilities, and managerial practices. She has authored over twenty publications, including peer-reviewed articles, on employment issues such as preventing, investigating, and correcting workplace discrimination, harassment, and retaliation. For this case, Duffy has reviewed DHL’s employee handbook, anti-discrimination training materials, declarations by human resource managers, employee complaints, depositions of managers taken in this case, and depositions of roughly a quarter of the claimants. EEOC Mot. Exclude Duffy, Ex. A, 3/13/15 Materials for J. Duffy, at 46. Based upon her evaluation, Duffy opines that DHL has “fully complied with its obligations to establish, maintain, and enforce appropriate anti-discrimination policies, procedures, and programs. It also met, and in certain respects even exceeded[,] the usual and reasonable management practice or appropriate 16 standard of care.” Duffy Report at 4. The EEOC seeks to exclude Duffy’s testimony under Daubert. As discussed, the Daubert inquiry involves a “three-step analysis,” which asks “whether the witness is qualified; whether the expert's methodology is scientifically reliable; and whether the testimony will ‘assist the trier of fact to understand the evidence or to determine a fact in issue.’” Myers v. Ill. Cent. R.R. Co., 629 F.3d 639, 644 (7th Cir. 2010) (quoting Ervin v. Johnson & Johnson, 492 F.3d 901, 904 (7th Cir. 2007)). “Daubert sets forth a non-exhaustive list of guideposts to consult in assessing the reliability of expert testimony: (1) whether the scientific theory can be or has been tested; (2) whether the theory has been subjected to peer review and publication; and (3) whether the theory has been generally accepted in the relevant scientific, technical, or professional community.” Am. Honda Motor Co. v. Allen, 600 F.3d 813, 817 (7th Cir. 2010). Having reviewed Duffy’s professional qualifications, the Court first concludes that she has the necessary knowledge, skill, experience, and training to testify about the methods that companies use to try to comply with their obligations under Title VII. 6 In addition to her other qualifications, as an employment lawyer and consultant, Duffy has counseled numerous clients to help them create anti-discrimination policies and procedures. As such, the Court finds that Duffy is qualified to evaluate DHL’s policies and practices to determine whether they are in line 6 According to Duffy, in order to determine whether a company has made sufficient efforts to do so, one considers whether a company has: (1) adopted clear and comprehensive policies and procedures; (2) invested in programs, systems, and personnel to enforce its policies; (3) educated its employees, managers and supervisors as to their rights and responsibilities under the policies; (4) held its managers accountable for reporting discrimination and potential discrimination; (5) created a system to report, investigate, and evaluate discrimination complaints as well as to make recommendations to decision makers to correct any discrimination; and (6) established an archival recordkeeping system that enables the employer to track similar complaints against similar actors. Duffy Dep. at 68–118. 17 with industry standards. See Jimenez v. City of Chi., 732 F.3d 710, 721–22 (7th Cir. 2013) (“Expert testimony regarding relevant professional standards can give a jury a baseline to help evaluate whether a defendant’s deviations from those standards were merely negligent or were so severe or persistent as to support an inference of intentional or reckless conduct that violated a plaintiff’s constitutional rights.”); Peone v. Mary Walker Sch. Dist. No. 207, No. CS-02-135RHW, 2003 WL 25689969, at *2 (E.D. Wash. May 27, 2003) (“It will be helpful to the jury to hear testimony of the types of steps that employers generally take to prevent and respond to sexual harassment in the work place.”). The Plaintiff-Intervenors also object to Duffy’s testimony, arguing that she is usurping the role of the Court. The Court disagrees. Duffy certainly cannot testify that DHL complied with Title VII during the period in question; this would be a legal conclusion, inappropriate for expert testimony and unhelpful to the jury. See Panter v. Marshall Field & Co., 646 F.2d 271, 293 n.6 (7th Cir. 1981); In re Fedex Ground Package Sys., Inc. Employment Practices Litig., No. 3:05–MD–527 RM, 2010 WL 1838400, at *4 (N.D. Ind. May 4, 2010). But Duffy stops well short of that. Duffy concludes that DHL’s policies, procedures, and programs met or exceeded the standard anti-discrimination management practices of similarly situated employers. She also explains that the purpose of these industry standards is to provide guidance to employers to help them comply with federal legal requirements. Duffy concludes, based on her review of the record, that DHL’s managers took certain steps that were consistent with, and went above and beyond, typical anti-discrimination practices employed in similar companies. This testimony will aid the jury in its determination of whether DHL made a good-faith effort to comply with Title VII. 18 To this, the EEOC argues that, in the process of forming her opinions, Duffy has given more weight to certain witness testimony than others and, in doing so, usurped the jury’s role to make credibility determinations and weigh the evidence. But all experts make certain assumptions (factual and theoretical) in rendering opinions, and factual assumptions by their very nature credit one version of the facts over another. For example, among other things, Duffy assumed that employees were generally aware of ways to report their discrimination complaints based on the fact that many had done so. Duffy also assumed that DHL’s employee manual was distributed to all employees and that DHL’s policies were posted. Duffy also assumed that DHL managers forwarded discrimination complaints to their superiors. Rather than excluding Duffy’s testimony, the proper way for the EEOC to challenge these assumptions is to cross-examine her with contrary facts in the record or to present evidence that undermines Duffy’s factual assumptions. See Manpower, 732 F.3d at 808 (stating that typically the “reliability of data and assumptions used in applying a methodology is tested by the adversarial process”); Lapsley, 689 F.3d at 805 (stating that the appropriate way to attack shaky but admissible evidence is through “vigorous cross-examination” and “presentation of contrary evidence”). Finally, the EEOC argues that Duffy’s methodology is unreliable because Duffy has presented no benchmark with which to measure DHL’s policies or any objective sources of comparison. Boiled to its essence, this argument questions Duffy’s qualifications to explain industry standards and challenges the weight of her testimony rather than the reliability of her methodology. As discussed above, based on her thirty-nine years as an employment lawyer and management consultant, Duffy has accrued sufficient experience to assess whether DHL’s employment policies were in keeping with general industry standards. Despite this, the EEOC contends that Duffy has not identified any comparable employers. This is incorrect. In fact, 19 during her deposition, Duffy identified a number of employers in the industry for which she has worked, including Federal Express, United Parcel Service, and CRST. Duffy Dep. at 102-130. In the event that the EEOC does not believe that these companies are comparable to DHL, or that Duffy compared them using incorrect parameters, it may explore that on cross-examination at trial. Conclusion For the reasons stated herein, the Court denies Defendant DHL Express USA, Inc.’s motion to exclude the testimony of Dr. Thomas DiPrete [229] and denies the Equal Employment Opportunity Commission’s motions to exclude the testimony of Dr. James Langenfeld [222] and D. Jan Duffy [224]. Additionally, the Court bifurcates the trial in this case. The jury will first be tasked with determining liability and, if necessary, compensatory damages. If the jury finds DHL liable and awards compensatory damages, the Court will then allow the parties to present evidence pertinent to the Kolstad defense to enable the same jury to determine whether punitive damages are appropriate. IT IS SO ORDERED. ENTERED 9/30/16 __________________________________ John Z. Lee United States District Judge 20
]]>
Some case metadata and case summaries were written with the help of AI, which can produce inaccuracies. You
should read the full case before relying on it for legal research purposes.
This site is protected by reCAPTCHA and the Google
Privacy Policy and
Terms of Service apply.
