LAWRENCE R. LEONARD, Magistrate Judge.
Before the Court are Defendant Newport News Industrial Corporation's ("Shipyard" or "Defendant") Motion to Strike Plaintiffs' Untimely Expert Report ("Motion to Strike"), ECF No. 194, and Motion to Preclude Testimony of Dr. Mark R. Killingsworth ("Motion to Preclude Killingsworth"), ECF No. 229, as well as Plaintiffs' Motion to Extend and to Compel, or in the Alternative, to Strike ("Motion in the Alternative"), ECF No. 272, and Motion to Exclude Anderson Testimony ("Motion to Exclude Anderson"), ECF No. 285. A hearing was held on these matters on April 13, 2017, at which Rebecca Houlding appeared on behalf of the Plaintiffs and Nicole Howell and Steven Moore appeared on behalf of the Defendant. ECF No. 289. The matter therefore is now ripe for disposition. For the following reasons, Defendant's Motion to Preclude, ECF No. 229, is
This case involves thirty-seven plaintiff Shipyard workers — welders, fitters and laborers — alleging various individual employment-related claims including discrimination and retaliation against Defendant Newport News Industrial Corporation. Included among the Plaintiffs are permanent employees of the Shipyard, temporary employees who performed work at the Shipyard through temporary staffing agencies, and employees who at one time were temporary but later became permanent. The Shipyard maintained two data systems, denominated "Legacy" and "Navision," to keep track of, inter alia, the job titles and pay of the employees. Employment and pay information for permanent employees was tracked though Legacy, and such information for temporary employees was tracked through Navision. Importantly, each of these systems had shortcomings in being able to reflect precisely which employees held what job titles at what specific point in time and at what specific pay rate.
The parties' four motions
At the hearing, Ms. Howell argued on behalf of the Shipyard that it experienced prejudice from Plaintiffs' late filing of the report by disrupting the discovery schedule laid out by the Court and the briefing for the summary judgment motions, although she stated that Defendant was able to address the issues pertinent to summary judgment without the report. Ms. Houlding responded that Defendant was not prejudiced because the delay did not impact the trial date and Defendant could have moved for an extension to file its summary judgment motions until it received the report, but chose not to.
Defendant filed a Motion to Preclude the Testimony of Dr. Killingsworth on December 12, 2016, and an accompanying memorandum. ECF Nos. 229-230. The Shipyard's motion was predicated on the contention that Dr. Killingsworth's analysis was irrelevant and unreliable under Federal Rule of Evidence 702 and Daubert v. Merrill Dow Pharmaceuticals, 509 U.S. 579 (1993) and Kumho Tire Co. v. Carmichael, 526 U.S. 137 (1999) and their progeny. As a part of its motion, Defendant attached the expert report of G. Edward Anderson, Ph.D., a labor economist who performed a statistical analysis of his own. In his report Dr. Anderson criticized Dr. Killingsworth's analysis because it did not limit the data analyzed to the pay of workers in the same trades as Plaintiffs. Then, relying on adjustments to the data
At the hearing, Mr. Moore argued on behalf of Defendant that broad statistical evidence of the type offered by Dr. Killingsworth's testimony can only be used in a case alleging "pattern or practice" discrimination, which is not the case here. This is especially true, Defendant argued, when analysis which is limited to the relevant comparable trades demonstrates no statistically significant difference between pay rates of black and white employees in those specific relevant trades. In response, Ms. Houlding claimed that Dr. Killingsworth's analysis could be used to show the Shipyard's intent to discriminate in its pay of Plaintiffs because it demonstrates that managers and supervisors, who alone decide whether employees received a raise or are promoted, act in a discriminatory manner towards all employees.
On December 16, 2016, Plaintiffs filed a Motion to Extend and to Compel, or in the Alternative, to Strike, ECF No. 272, and an accompanying memorandum, ECF No. 271. Plaintiffs sought an extension of time or permission to serve their rebuttal report as well as to compel FRCP 26 disclosures and complete responses to their requests for production Nos. 59 and 60 regarding the data reviewed and relied upon by Defendant's experts. Id. Plaintiffs also sought permission to depose Scott Jones and other individuals mentioned in Dr. Anderson's expert report, or in the alternative to strike Defendant's report because all of the Scott Jones data on which Dr. Anderson relied was not provided to Plaintiffs in discovery. Defendant filed its Memorandum in Opposition on January 5, 2017, ECF No. 276, to which Plaintiffs replied on January 19, 2017, ECF No. 283.
At the hearing, Ms. Houlding advised that Plaintiffs served a preliminary rebuttal report upon Defendant after Defendant withdrew its objection, but that she wanted to reserve the right to serve a more complete rebuttal report at a later date if necessary. Ms. Houlding also noted that the Shipyard did not supplement its discovery responses with the data Scott Jones relied upon in preparing his chart which he provided to Dr. Anderson, and Plaintiffs were denied the opportunity to depose Jones regarding this information. Ms. Howell advised no data exists on which Jones relied to prepare his chart, and that instead he relied on his own memory to fill in gaps in Defendant's employment data to determine which employees served in which specific jobs and at which specific times. Ms. Howell argued on behalf of Defendant that the Jones data was provided to Dr. Anderson to "fill in the gaps" of the Legacy and Navision data, and that Plaintiffs had the opportunity to depose Jones as a corporate designee witness but failed to ask him questions in this regard.
Lastly, on January 20, 2017, Plaintiffs filed a Motion to Exclude Dr. Anderson's Testimony, and an accompanying memorandum, ECF Nos. 285-286, in which they argued his testimony should be precluded under FRE 702(b) and Daubert. Relying on FRE 703, Plaintiffs contended that no expert in the field of labor economics would reasonably rely on the kinds of facts or data that Jones provided via his chart simply from memory. Consequently, they argued, Dr. Anderson's reliance on the Jones data rendered his opinions unreliable and thus inadmissible. Moreover, Plaintiffs contended that the Shipyard "refused to produce Jones for deposition, and effectively the mystery employees who helped him and the documents on which they relied (if any)." ECF No. 286 at 7. Plaintiffs also claimed that "[t]he data represent-allegedly-job titles for 350 temporary and permanent employees (in 794 data points) during the period 2008 to 2016" and that Defendant "never produced evidence for majority of the job titles during discovery, in any form", id. at 3, despite the fact that Plaintiffs in their First Set of Interrogatories asked Defendant "[f]or each Production worker working at Employer, directly or through a staffing agency, from 2001 to the present, list: Name . . . position at hire", "wage at hire", "wage changes including dates of changes", "promotion(s)/demotion(s) including date", "termination(s)", "rehire(s)", "Regular hours worked per year" and "Overtime hours worked per year." Id., attach. 6 at 3-5.
Defendant responded on February 3, 2017 that Plaintiffs were aware of the shortcomings in the electronic data systems but chose not to "fill in the blanks" left by it while discovery was still open. ECF No. 287 at 2. Defendant explained that it maintained two databases, Legacy and Navision, to store a record of its employees' job titles. Id. at 4. Defendant advised that while Navision was more detailed than Legacy, "historical changes in `actual job titles' were not stored in Navision and, consequently, were not maintained in any electronic format." Id. Defendant stated that "Dr. Anderson determined that Scott Jones, NNI Operations Manager, would be a reliable source to answer some of the questions presented in the data, including what positions were held by the workers at various points" so Dr. Anderson used the Jones data to respond to Dr. Killingsworth's report. Id. at 8. Plaintiffs timely filed a Reply on February 9, 2017, claiming that if Dr. Anderson's testimony hinges on the Jones data at the least the Defendant should be compelled to turn over the material relied upon by Jones to compile it. ECF No. 288.
At the hearing, Ms. Houlding argued that Dr. Anderson should have verified the Jones data to ensure it was not fabricated or did not contradict evidence in the record, and that the data itself should be excluded because it cannot be verified or replicated and was created late in the litigation. Additionally, Ms. Houlding claimed that Plaintiffs asked for the Jones data for Dr. Killingsworth's expert opinion, but that Defendant did not provide it until after Dr. Killingsworth's opinion was submitted. Mr. Moore said in response that it is customary for Dr. Anderson to contact the company when he has questions about the data and that the company officials themselves are often a better source then their records for information such as this. He also stated that Dr. Anderson's report is merely a rebuttal report to Dr. Killingsworth, so Plaintiffs' motion to exclude Dr. Anderson would become moot if the Court grants either of Defendant's instant motions and strikes Dr. Killingsworth's report on either basis sought.
All thirty-seven Plaintiffs have asserted pay disparity claims. ECF No. 15 at ¶¶ 1242-1245. To establish a pay disparity claim, a plaintiff must show the following elements: "(1) that she is a member of a protected class; (2) she was paid less than an employee outside the class; and (3) the higher paid employee was performing a substantially similar job." Davis v. S.C. Dep't of Health & Envtl. Control, No. 3:13-CV-02612-JMC, 2015 WL 5616237, at *3 (D.S.C. Sept. 24, 2015); see also, Brinkley-Obu v. Hughes Training, Inc., 36 F.3d 336, 343 (4th Cir. 1994). Necessarily then, each of the thirty-seven Plaintiffs here must show that they were paid less than their white comparators to prevail on their pay disparity claims.
Plaintiffs seek to prove discriminatory pay disparity by showing, through Dr. Killingworth's regression analysis, that black Shipyard employees generally are paid less than white Shipyard employees. Dr. Killingsworth utilized regression analysis to reach the conclusion that such pay disparity existed at the Shipyard. "Multiple regression analysis . . . compares many characteristics within a particular set of data and enables the determination of how one set of factors is related to another, single factor." Smith v. Virginia Commonwealth Univ., 84 F.3d 672, 674 (4th Cir. 1996). Here, Dr. Killingsworth controlled for multiple variables to reach his conclusions, but did not control for the specific job classification data.
In the first instance, Plaintiffs are correct that Defendants did not produce in discovery the type of specific job classification data that would have permitted analysis of the pay rates of black and white Shipyard employees with the same specific job titles or trades. Plaintiffs' interrogatory 13 quite explicitly asked for the specific job classification data at issue. ECF No. 286, attach. 6 at 3-5. The Shipyard provided the Legacy and Navision data, but advised that to provide any more specific information would require "reviewing hundreds of personnel files" and thus would be "burdensome" and so refused to do so. Id. However, after declining to produce any further specific job classification data, the Shipyard called on Scott Jones, its Operations Manager, to construct that very data for its own expert. This was accomplished by means of a chart he produced for Dr. Anderson — the Jones data — which then enabled Dr. Anderson to perform the very type of analysis Defendant blames Dr. Killingsworth for failing to perform. Dr. Anderson believes Jones relied largely on his memory for the information used in the chart. ECF No. 283, attach. 2 at 12-13. However, the admissibility of Dr. Killingsworth's opinion relies on the relevance and reliability of the opinion itself, and not on the reasons why Dr. Killingsworth did not have the information that would have made that opinion relevant and reliable. In other words, any fault that may be attributed to Defendant because of shortcomings in complying with its discovery obligations does not otherwise make Dr. Killingsworth's opinion admissible.
Dr, Killingsworth did not, in his first report, control for specific job classification data. As discussed supra, to establish a discriminatory pay disparity claim Plaintiffs are required to show individually that they were paid less than similarly situated employees outside their protected class. Brinkley-Obu, 36 F.3d at 343. Certainly, Plaintiffs can utilize regression analysis to show that all black employees with the same jobs as them make less than all white employees with the same job. See, Bazemore v. Friday, 478 U.S. 385 (1986). In Bazemore, a group of black employees and the United States brought a pattern and practice pay disparity discrimination claim against the North Carolina Agricultural Extension Service and various counties. Acknowledging the value and admissibility of regression analysis in employment discrimination cases, the Supreme Court reversed a ruling by the Fourth Circuit that affirmed a district court's rejection of a regression analysis that had omitted certain measurable values. 478 U.S. at 400. The United States' expert had performed a regression analysis that used four independent variables: race, education, tenure, and job title. Id. at 398. The expert had not included other variables that the district court understood was "thought to have an effect on salary level." Id. at 400. In finding that the regression analysis performed by the expert should not have been excluded, the Court held:
Id. (emphasis added). The Court then noted in a footnote: "There may, of course, be some regressions so incomplete as to be inadmissible as irrelevant; but such was clearly not the case here." Id. at n.10.
The problem with Dr. Killingsworth's analysis is that, by not using specific job classification data in his analysis, he failed to account for a major factor — that the employees being compared were similarly situated, i.e. performed substantially similar jobs. As the Bazemore Court noted, while an admissible regression analysis need not account for all variables "thought to have an effect on salary levels," it must at least account for "the major factors." Id. The specific job classifications of the workers here is such a major factor, as Plaintiffs cannot prove their pay disparity claims without pointing to similarly situated coworkers. Failing to compare Plaintiffs with coworkers with substantially similar jobs renders his analysis "so incomplete as to be inadmissible as irrelevant." Id. at n. 10.
Plaintiffs contended further that, even if Dr. Killingsworth did perform his regression analysis so broadly as to include all Shipyard workers and not just those who are similarly situated, such analysis is admissible to demonstrate discriminatory intent and pretext. This argument posits that, by being able to show that the Shipyard generally pays black workers less than it pays white worker, this is evidence that the Shipyard intended to discriminate against Plaintiff A when it paid him what it did, and that the reason the Shipyard articulated for why Plaintiff A received the pay that he did is merely a pretext for the real reason, which is discrimination. However, this argument also falls short. In some cases, broad regression analysis, such as the type performed by Dr. Killingsworth, can be admissible to show racial discrimination is a company's standard practice. This, however, occurs not in single plaintiff
Unlike the present case, which involves the alleged disparate treatment of thirty-seven individual plaintiffs, pattern and practice cases are brought by the Government as class actions against employers engaging in systematic discrimination against a protected group. Lowery v. Circuit City Stores, Inc., 158 F.3d 742, 759 (4th Cir. 1998), vacated on other grounds, 527 U.S. 1031 (1999). Consequently, in these cases the Government does not have to prove that any particular employee was a victim of a discriminatory practice, but instead must establish that racial discrimination is "the company's standard operating procedure — the regular rather than the unusual practice." Int'l Bhd. of Teamsters v. United States, 431 U.S. 335, 336 (1977). "The proof of the [discriminatory] pattern or practice supports an inference that any particular employment decision, during the period in which the discriminatory policy was in force, was made in pursuit of that policy."
Nonetheless, this is not a pattern or practice case. Each Plaintiff bears the burden of proving that, because of his race, he was paid less than persons outside his protected class who were similarly situated. A regression analysis which fails to compare similarly situated workers is flawed in the first instance, and cannot be saved by claiming it is relevant for some other issue. See Anderson v. Westinghouse Savannah River Co., 406 F.3d 248 (4th Cir. 2005) (affirming district court's decision to exclude regression analysis which failed to compare similarly situated workers). Dr. Killingsworth's opinion does not establish anything about how black welders, fitters, and laborers are paid compared to white welders, fitters, and laborers. Perhaps through no fault of his own, he has omitted from his analysis a major factor that must be considered — the job titles and duties of the relevant workers. As a result, his "regression[] [analysis is] so incomplete as to be inadmissible as irrelevant opinion." Bazemore, 478 U.S. at 400 n.10. Therefore, Defendant's Motion to Preclude Dr. Killingsworth's testimony, ECF No. 229, is
In response to Dr. Killingsworth's regression analysis opinion, the Shipyard proffered the regression analysis of its own expert, Dr. Anderson. There are two components to Dr. Anderson's opinion. First, Dr. Anderson criticized Dr. Killingsworth's analysis as improper because he did not compare workers with the same job. ECF No. 286, attach. 1 at 9. Second, when such proper analysis (as he believed was required) was performed — comparing workers in similar jobs, which he was able to do because of the Jones data — the results showed no pay disparity. Id. Looking first to the second component of his opinion, Dr. Anderson sought to compare workers in relevant job categories. The data systems maintained by Defendant did not capture specific information regarding specific job titles and salaries during the specific time periods for Plaintiffs and their white counterparts. Jones provided such specific information, at least in part from memory, to Dr. Anderson. Defendant did not provide this specific information to Plaintiffs prior to disclosing Dr. Anderson's report, despite being obligated to do so in response to Plaintiffs' Interrogatory #13 which directly asked for it. Id., attach. 6 at 3-5. Dr. Anderson relied on the Jones data in his analysis and concluded that there was no pay disparity among black and white workers with the same relevant job titles. Id., attach. 1 at 16.
Plaintiffs contend that the Jones data is unreliable, that no reasonable expert would rely on it, and therefore Dr. Anderson's opinion, which depends on the Jones data to compare similarly situated workers, is unreliable and should be stricken pursuant to FRE 702 and Daubert. The Court agrees. FRE 702 permits admission of "scientific, technical or other specialized knowledge" by a qualified expert if it will "help the trier of fact to understand the evidence or to determine a fact in issue," "the testimony is based on sufficient facts or data," "is the product of reliable principles and methods," and "the expert has reliably applied the principles and methods to the facts of the case." Fed. R. Evid. 702 ("Rule 702"). The seminal case of Daubert v. Merrell Dow Pharm., Inc. provides that expert testimony may be admitted pursuant to Rule 702 if the testimony is both relevant and reliable, considering a number of factors including whether the theory or technique "can be (and has been tested)," whether it "has been subjected to peer review and publication," whether it has been "generally accept[ed]" in the "relevant scientific community," and "the known or potential rate of error." 509 U.S. 579, 589, 593-94 (1993). Although the admissibility of an expert opinion is "flexible," the district court must function as a gatekeeper, permitting only expert testimony that comports with Rule 702's guidelines as explained in Daubert. 509 U.S. at 594. Moreover, the Court's gatekeeping responsibility extends to all expert testimony, and requires that the testimony be relevant and reliable regardless of whether it is scientific. Kumho Tire Co. v. Carmichael, 526 U.S. 137, 147 (1999). "[T]he court has broad latitude to consider whatever factors bearing on validity that the court finds to be useful; the particular factors will depend upon the unique circumstances of the expert testimony involved." Westberry v. Gislaved Gummi AB, 178 F.3d 257, 261 (4th Cir. 1999) (citing Kumho Tire Co., 526 U.S. at 149-153).
Here, the Jones data is unreliable because there is no way to verify or replicate it. When Dr. Anderson was asked in his deposition what information Jones relied on in formulating his data, he stated "I don't actually know what Scott [Jones] would say he relied on, but I believe — I believe some of the information is from his memory, yes." ECF No. 286, attach. 2 at 12. Thus, Dr. Anderson did not verify the true source of the data Jones provided. Likewise, since the data is derived from the memory of one person, it cannot be verified or replicated. Dr. Anderson needs the Jones data to support his conclusion that there is no pay disparity between welders, fitters, and laborers because the existing data under the Legacy and Navision data systems alone do not provide accurate specific job classification data for the relevant workers. Id., attach. 1 at 3-4. Consequently, the second component of Dr. Anderson's opinion relying on the Jones data must be excluded.
Furthermore, the Jones data was not timely disclosed even though Plaintiffs requested it in an interrogatory. ECF No. 286, attach. 6 at 3-5. Plaintiffs' Motion in the Alternative seeks alternative relief in the form of an Order striking Dr. Anderson as an expert based on the Shipyard's failure to produce the Jones data during discovery. ECF Nos. 272, 271. The Shipyard claimed that producing this type of information in response to Interrogatory 13 was too burdensome, so it did not do so. ECF No 286, attach 6 at 3-5. Apparently, however, it was not so burdensome that Operations Manager Jones, either straight from memory or by reviewing personnel files, could not produce it for Dr. Anderson. The Shipyard cannot have it both ways, and thus even if the Jones data were considered reliable, the Shipyard's failure to produce this information when obligated to do so should preclude it from being able to use the Jones data for its benefit.
Rule 26(e)(1) of the Federal Rules of Civil Procedure requires a party responding to discovery to supplement or correct its response in a timely manner if the party learns in some material respect the response is incomplete or incorrect. Inasmuch as the Shipyard's Operations Manager apparently possessed all the information necessary to respond to Interrogatory 13, it was incumbent on the Shipyard to provide this information to Plaintiffs, either by supplementing or correcting this interrogatory answer. This they failed to do until long after Plaintiffs were required to disclose their own expert witness report. Given the fact that this missing information forms a primary basis for Dr. Anderson's critique of Dr. Killingsworth, there can be no doubt that this information was material.
Rule 37(c) of the Federal Rules of Civil Procedure provides in pertinent part: "If a party fails to provide information or identify a witness as required by Rule 26(a) or (e), the party is not allowed to use that information or witness to supply evidence on a motion, at a hearing, or at trial, unless the failure was substantially justified or is harmless." The Defendant's failure to disclose was neither substantially justified nor harmless. Precluding Dr. Anderson from relying on the Jones data is an appropriate sanction for Defendant's failure to disclose. See Saudi v. Northrop Grumman Corp., 427 F.3d 271, 278-79 (4th Cir. 2005) ("Rule 26 disclosures are often the centerpiece of discovery in litigation that uses expert witnesses. A party that fails to provide these disclosures unfairly inhibits its opponent's ability to properly prepare, unnecessarily prolongs litigation, and undermines the district court's management of the case. For this reason, [w]e give particularly wide latitude to the district court's discretion to issue sanctions under Rule 37(c)(1)." citing Southern States Rack & Fixture v. Sherwin-Williams Co., 318 F.3d 592, 595 (4th Cir. 2003) (internal quotation marks omitted)).
Finally, with respect to the first component of Dr. Anderson's opinion — the criticism that Dr. Killingsworth failed to compare similarly situated workers, i.e. workers with the same jobs as welders, fitters and laborers — in light of the Court's rulings excluding both Dr. Killingsworth's opinion and that part of Dr. Anderson's opinion relying on the Jones data, the first component of Dr. Anderson's opinion is irrelevant and moot.
Consequently, because the Jones data Dr. Anderson utilized is unreliable, his conclusions are also unreliable, and therefore that part of Dr. Anderson's opinion purporting to demonstrate that no pay disparity exists between black and white welders, laborers and fitters does not satisfy Federal Rules of Evidence 702 and 703. Since the first component of Dr. Anderson's opinion addresses Dr. Killingsworth who has been excluded, that part of Dr. Anderson's opinion is irrelevant and must be excluded on that basis. For these reasons, Plaintiffs' Motion to Exclude, ECF No. 285, is
Consequently and in conclusion, Defendant's Motion to Preclude, ECF No. 229, is
It is so