Opinion
2:19-CV-12333-TGB-DRG
03-31-2023
MARK D. CHAPMAN, et al., Plaintiffs, v. GENERAL MOTORS LLC, Defendant.
ORDER DENYING MOTIONS TO EXCLUDE EXPERT WITNESS TESTIMONY
(ECF NOS. 119/120, 121, 122, 124/125)
TERRENCE G. BERG, UNITED STATES DISTRICT JUDGE
The parties in this putative car defect class action have filed motions challenging the admissibility of the testimony of five expert witnesses on various grounds. These witnesses offer opinions and reports in support of and against class certification in this case. GM challenges the testimony of four witnesses, and Plaintiffs challenge one. For the reasons explained below, the motions will be DENIED.
I. BACKGROUND
A. Factual Allegations
The Court refers to its previous explanation of the defect alleged in this case. Chapman v. Gen. Motors LLC, 531 F.Supp.3d 1257, 1268-70 (E.D. Mich. 2021). Briefly, this case is about a defect in GMC and Chevrolet trucks from model years 2011-2016 with 6.6L Duramax engines, caused by GM's decision to equip the vehicles with a Bosch CP4 pump. ¶ 1, ECF No. 40, PageID.3393. According to Plaintiffs, the CP4 pump is more fragile and susceptible to wear and tear than its predecessor, the Bosch CP3 model, because of several differences in its design. ¶¶ 123-135, ECF No. 40, PageID.3478-85. Its potential for malfunction is exacerbated by a factor unique to the United States-our diesel fuel is “drier” than the diesel fuel available in other countries because of emissions-related EPA regulations. The CP4 pump relies on the diesel fuel itself for lubrication, so it is uniquely unsuited for use with our “dry” diesel. Id. at ¶¶ 148-52.
The combination of the allegedly subpar pump design and lack of lubricity from American diesel fuel leads to friction which can, among other things, cause small metal shavings to build up within the pump or engine and fuel block. If there is too much buildup, “catastrophic failure” may occur; the car suddenly stalls or loses power, generally requiring a tow and often times replacement of most of the fuel system. Id. at ¶¶ 137-140. Even if the car does not experience catastrophic failure, the wear and tear on the pump can harm the fuel injector and other parts of the engine, causing owners to suffer damages. Id. at ¶ 141.
Plaintiffs allege GM was aware of these issues even before it began to sell cars with the CP4 engine. Id. at ¶¶ 8, 185. Despite that knowledge, GM nonetheless marketed the vehicles as having increased durability and fuel efficiency, in part due to the use of diesel engine technology. Plaintiffs also allege that, instead of taking steps to remedy the problem, GM actively concealed it for as long as possible. Id. at ¶¶ 9, 186-88. GM stopped using the CP4 pump after the 2016 model year of the class vehicles, switching to another model of pump that had been in use since 2004. Id. at ¶ 218.
B. Procedural posture
The Court takes note of several procedural developments since its last Order. First, a number of named Plaintiffs have been dismissed or substituted from the case. ECF Nos. 85, 94, 95, 104. Second, a case regarding the same conduct that was separately filed in a Texas district court was transferred to this district and consolidated with this case. ECF No. 105. Parties began discovery in May 2021.
Plaintiffs have also filed a motion for class certification. ECF Nos. 111/112. That motion is addressed by separate order. Now before the Court are three Daubert motions by GM to exclude Plaintiffs' various experts in support of class certification, ECF Nos. 119/120 (Motion to Exclude Stockton); ECF No. 121 (Motion to Exclude Edgar); ECF No. 122 (Motion to Exclude Gaskin and Weir), as well as a motion by Plaintiffs to exclude Defendant's rebuttal expert. ECF Nos. 124/125 (Motion to Exclude Harrington). The Court held oral argument on all these motions on August 5, 2022.
The parties have filed several supplemental briefs regarding recent caselaw developments. See ECF Nos. 146, 147, 155, 156, 158, 162, 163. Additionally, the Court asked Plaintiffs for supplemental briefing regarding their proposed class definition at oral argument, to which Defendant has replied. ECF Nos. 154, 157. This order will resolve the parties' Daubert challenges to one another's experts.
II. LEGAL STANDARDS
Federal Rule of Evidence 702 addresses the admissibility of expert testimony:
A witness who is qualified as an expert by knowledge, skill, experience, training, or education may testify in the form of an opinion or otherwise if:
(a) the expert's scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue;
(b) the testimony is based on sufficient facts or data;
(c) the testimony is the product of reliable principles and methods; and
(d) the expert has reliably applied the principles and methods to the facts of the case.
This Rule was modified in December 2000 to add subparagraphs (b) through (d) in order to reflect the Supreme Court's emphasis that an expert's opinion should be grounded in the actual facts of the case, valid under the principles of the discipline that furnished the base of special knowledge, and structured so as to “fit” the facts of the case into the theories and methods that the expert espouses. Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 591-93 (1993); see also In re FCA U.S. LLC Monostable Elec. Gearshift Litig., 382 F.Supp.3d 687, 691 (E.D. Mich. 2019) (Lawson, J.) (discussing amendment history).
The Supreme Court in Daubert identified several factors that might bear on whether proposed expert testimony is scientifically valid and “reliable:” (1) whether the theory or technique has been tested; (2) whether the theory or technique can be or has been peer reviewed or published; (3) the known or potential error rate; (4) the existence and maintenance of standards controlling the technique's operation; and (5) the general acceptance by the relevant scientific community and the testimony's degree of acceptance therein. 509 U.S. at 593-94. These factors are not meant to establish a “definitive checklist or test;” the test of reliability is “flexible” and must be tailored to the facts of a particular case. In re Scrap Metal Antitrust Litig., 527 F.3d 517, 529 (6th Cir. 2008) (quoting Kumho Tire Co., Ltd. v. Carmichael, 526 U.S. 137, 150 (1999)).
Ultimately, the key inquiry in this analysis is “whether a putative expert's testimony would be inadmissible junk science or instead would be testimony falling within the ‘range where experts might reasonably differ.'” Thomas v. Novartis Pharm. Corp., 443 Fed.Appx. 58, 60 (6th Cir. 2011) (quoting Kumho Tire, 526 U.S. at 153). “[R]ejection of expert testimony is the exception, rather than the rule.'” In re Scrap Metal Antitrust Litig., 527 F.3d at 530 (internal quotations omitted). The Court's focus is not whether an opinion is “correct,” but rather on whether it “rests upon a reliable foundation.” Id. at 529-30. It is for a jury to determine what weight it should be accorded.
But in the question whether to certify class action litigation, there is no jury involved. Neither the Supreme Court nor the Sixth Circuit have addressed how motions seeking to exclude expert testimony under the Daubert framework should inform a class certification analysis. See Comcast Corp. v. Behrend, 569 U.S. 27, 39-40 (2013) (Ginsburg, J., dissenting) (describing how the Court granted certiorari to resolve the Daubert question but did not reach its merits); Hicks v. State Farm Fire & Cas. Co., 965 F.3d 452, 465 (6th Cir. 2020) (“We have yet to settle this matter … [t]his case does not present an opportunity to do so.”).
During class certification, the Court must conduct a “rigorous analysis” to determine-among other things-whether “questions of law or fact common to class members predominate over any questions affecting only individual members.” Comcast Corp., 569 U.S. at 27; Fed.R.Civ.P. 23(b)(3). Some courts have concluded that, in making this determination, “courts can consider evidence that may not be admissible at trial.” Ganci v. MBF Inspection Servs., Inc., 323 F.R.D. 249, 257 n.2 (S.D. Ohio 2017) (internal citations omitted). Others, including the Sixth Circuit, hold that a court does not abuse its discretion by evaluating whether an expert's opinion is sufficiently reliable and informative to be considered. See In re Carpenter Co., No. 14-0302, 2014 WL 12809636, at *3 (6th Cir. Sept. 29, 2014) (suggesting that “such an analysis may be required in some circumstances”); see also In re FCA U.S. LLC Monostable Elec. Gearshift Litig., 382 F.Supp.3d at 691-92 (collecting cases).
The parties do not appear to contest whether the Daubert analysis applies at class certification. The Court will therefore resolve the motions, keeping in mind that the challenged testimony is not being offered to prove the merits of the Plaintiffs' claims but only to establish whether the claims can be adjudicated through collective litigation.
III. ANALYSIS
Defendants challenge the admissibility of the opinions of four of Plaintiffs' proposed experts: Dr. Bradley L. Edgar, Steven Gaskin, Edward Stockton, and Colin Weir. Dr. Edgar provides technical opinions regarding the CP4 pump's design, the properties of American diesel and its compatibility with the CP4 pump, GM's testing and development of the engines that used the pump, and the pump's failure rates. Stockton, Weir, and Gaskin are damages experts: Stockton proposes a model for calculating the damages of the proposed “failure/repair” group, whose members experienced catastrophic failures; Gaskin and Weir propose a model for calculating the damages of the proposed “overpayment” group, whose members did not experience catastrophic failures but who overpaid for their vehicles because of the allegedly defective pump.
Plaintiffs, meanwhile, seek to exclude the opinions and testimony of GM's rebuttal expert, Ryan Harrington, who was retained to evaluate the technical opinions of Dr. Edgar.
A. Defendant's Motion to Exclude Edgar Expert Report (ECF No. 121)
1. Summary of qualifications and testimony
Dr. Edgar is the Plaintiffs' main technical expert regarding their theory of liability, as to both the fuel pump itself and GM's knowledge and awareness of the defect. He was retained to analyze the design of the fuel system in GM diesel engines, the testing GM performed before it launched the class vehicles, pump failure rates in the class vehicles, GM's response to pump failures, and the quality of American diesel fuel. Edgar also provided opinions regarding an estimate of the number of CP4 failures, likely percentages of failures, and repairs covered and not covered by warranty. His calculations were offered to determine the class size and separate the class into “failure” and “overpayment” groups. Edgar Report, ECF No. 111-2 (sealed), PageID.21054.
Dr. Edgar holds a PhD in Mechanical Engineering from the University of California, Berkeley. Id. at PageID.21053. He is presently the Chairman and Senior Director of 44 Energy Technologies, Inc., a company he co-founded nine years ago, which offers consulting services and specializes in technology development, engine emissions and performance testing, and regulatory compliance. Id. Before his tenure at 44 Energy, he served as President at Cleaire Advanced Emission Controls, another company he co-founded, which developed, verified, and manufactured diesel emission control systems for heavy-duty diesel engine applications. Id. He has published several peer-reviewed papers with the Society of Automotive Engineers and delivered presentations on diesel engines and emissions. Id. at PageID.21053, PageID.21155-56. He is also the inventor or co-inventor of eleven patents for diesel engines and emission control technologies. Id. at PageID.21053.
To prepare his report, Dr. Edgar reviewed and analyzed a variety of sources-including test records from Bosch and GM; pleadings, deposition testimony, and GM's internal records provided during discovery; publicly available information about the CP4 pump; and academic papers and texts. Id. at PageID.21054. He also physically inspected the subject vehicles and disassembled two failed pumps. Id. To estimate CP4 pump failure rates, the number of repairs that were covered and not covered by warranty, and the size of the vehicle class, he relied on data in the form of spreadsheets provided by GM. Id. at PageID.21118, PageID.21123. Where data was limited or unavailable, he described its limitations and the assumptions he made to fill the gaps. See, e.g., id. at PageID.21123-24.
In his preliminary report, Dr. Edgar concludes that the CP4 pump has a “defectively fragile” design, is “inadequate for use in the U.S. diesel market,” and “relies on a series of assumptions about diesel fuel quality that are unrealistic.” Id. at PageID.21047. After recounting the history of the pump's development, he opines that the pump's design causes wear and tear, which can generate metal shavings that make class vehicles prone to catastrophic failure and loss of power. Id. at PageID.21080. Based on an analysis of testing records and product launch data, Dr. Edgar concludes that GM knew that the quality of American diesel fuel made the CP4 pump particularly unsuited for use and sale in the American market. Id. at PageID.21131. He estimates Plaintiffs' class to comprise 563,504 vehicles, 9.6% of which have experienced pump failure. Id. at PageID.21122. The remaining vehicles, which did not experience failure, comprise the “overpayment” group. Id.
2. Challenges to testimony
a) Qualifications
GM first argues that Dr. Edgar is not qualified to opine on CP4 pump design or its compatibility with American diesel because his experience with diesel technology “focuses on emission control systems and technologies”-rather than fuel pump design and fuel quality assessment. ECF No. 121, PageID.372820. It asserts that he has never designed a diesel fuel pump or (aside from one other case) evaluated the cause of fuel pump failure. Id. at PageID.37280-82. Relying on several cases in which experts were determined to lack appropriate qualifications to opine in particular areas, including Early v. Toyota Motor Corp., 277 Fed.Appx. 581, 585-56 (6th Cir. 2008), and Sigler v. Am. Honda Motor Co., 532 F.3d 469, 479 (6th Cir. 2008), GM argues that the mismatch between Dr. Edgar's training and experience and his proposed testimony are too great to withstand scrutiny under Daubert.
But the cases cited by GM all involve more extreme discrepancies between an expert's area of knowledge and the subject they were engaged to opine on. Dr. Edgar has several years of experience studying and working with diesel technology. Even if his main area of expertise is emissions, his education, experience, and the detail with which he is able to describe the technology and standards at issue indicate to the Court that he has appropriately specific knowledge to be able to opine on them. See Palatka v. Savage Arms, Inc., 535 Fed.Appx. 448, 455 (6th Cir. 2013) (noting that, “[a]s a professor of mechanical engineering there is little doubt that [the proposed expert] is qualified by his knowledge, skill, experience, training, and education to proffer an expert opinion”). Moreover, GM is incorrect in saying that Dr. Edgar does not have any experience at all with diesel fuel pump design. During his deposition, he testified that he worked on the design of one at a previous job, and also that his doctoral work required him to become familiar with and modify one to accept alternative fuels. Edgar Dep., ECF No. 132-6, PageID.38646; ECF No. 144-2, PageID.40767.
GM also challenges Dr. Edgar's ability to opine on the safety risks posed by the class vehicles. ECF No. 121, PageID.37281-82. Focusing on this attack makes a mountain out of a molehill: his “opinion” on safety risks is only about 2.5 pages of his report and largely consists of synthesizing National Highway Traffic Safety Administration (NHTSA) reports on the subject. This is not “opining” so much as it is providing relevant context to the rest of his report. The Court will not exclude Dr. Edgar based on the objection that he is not sufficiently qualified.
b) Methodology
GM also contends that portions of Dr. Edgar's report are inadmissible because they are not adequately tailored to the evidence in the case or the product of reliable methods.
i. Lack of testing
GM argues that Dr. Edgar's opinion regarding defective pump design and the related safety ramifications should be excluded because Dr. Edgar did not conduct any independent testing, examination, or safety analysis of CP4 pumps, and his report is based largely on his review of GM's testing data. ECF No. 121, PageID.37282-85. According to GM, Dr. Edgar “at a minimum” would need to test a statistically representative sample of CP4 pumps to be qualified to offer an opinion about whether the pump's design is defective. In supplemental briefing, it notes that another district court has previously excluded his opinion regarding CP4 pumps on this ground. ECF No. 158. See Stevens v. Ford Motor Co., No. 2:18-CV-00456, at 6-13 (S.D. Tex. Sept. 25, 2022).
The Court is not persuaded by the reasoning in Stevens. As an initial matter, GM misstates the extent of Dr. Edgar's own examinations and analysis. While Dr. Edgar did not personally remove a CP4 pump from a class vehicle, he inspected the vehicles and also conducted teardowns of two CP4 pumps, which involved dismantling them and inspecting their key drive train components. ECF No. 144-2, PageID.40779. He then reviewed reports and photos of other failed pumps and compared them to the ones he personally tore down. Id.
Moreover, while GM is correct that lack of testing is often a “red flag” that weighs against the admissibility of an expert's opinion, see Newell v. Rubbermaid, Inc. v. Raymond Corp., 676 F.3d 521, 527 (6th Cir. 2012), “[t]here is no per se requirement that an expert conduct his or her own testing.” In re: Gen. Motors LLC Ignition Switch Litig., No. 14-MD-2543 (JMF), 2015 WL 9480448, at *3 (S.D.N.Y. Dec. 29, 2015). “Deductive reasoning and critical review of existing test data can be a reasonable and reliable method of analysis, especially where (as here) extensive data on the precise question [the expert] was asked to consider had already been collected.” Id. Here, as evidenced by GM's document production, Bosch and GM conducted several tests on CP4 pumps. The results of these tests were made available to Dr. Edgar, and his report shows that he critically analyzed them, critiquing the scope of the testing and drawing conclusions from its results in view of his training and expertise. Moreover, given the nature of the defect alleged, direct testing on pumps necessarily would involve a small sample size- and since Plaintiffs allege that failures occur unpredictably, gleaning reliable conclusions from a small sample may have been difficult.
The Court concludes that this aspect of Dr. Edgar's methodology- i.e., critically reviewing available testing data, along with reports and photographs of other pump failures, against his own inspections-is sufficiently reliable to provide a basis for his testimony regarding pump design. And as noted above, his opinions regarding safety risks presented by pump failures make up only a small portion of his report and are informed by an extensive NHTSA investigation into the issue. His opinions will not be excluded on this ground.
ii. Reliability of failure rate calculations
GM next contends that Dr. Edgar's estimate of a 9.6% pump failure rate should be excluded. ECF No. 121, PageID.37285-88. Dr. Edgar's report explains that the denominator in this calculation is drawn from the total number of cars equipped with CP4 engines for the model years 2010-2016, which he determined to be 563,504 from GM's datasets. ECF No. 111-2, PageID.21118-22. From there, Dr. Edgar used replacement pump and pump kit sales to estimate the number of pump failures, reasoning that “sales of a pump or a pump kit for any reason besides replacing a failed pump are de minimis.” Id. at ¶ 175. This number is 54,324, leading to a failure rate of 54,324/563,504 = 9.6%.
GM argues that Dr. Edgar's use of replacement pump sales as a proxy for pump failures renders his calculations unreliable because it wrongly assumes that every sale reflects a failure under Plaintiffs' defect theory. GM points to Kondash v. Kia Motors Am., Inc., __ F.R.D. __, 2020 WL 5816228, at *8 (S.D. Ohio Sept. 30, 2020), where another district court rejected an identical method for calculating vehicle component failure rates. According to GM, using data from actual repairs and replacements of pumps for the calculation would have yielded more reliable results. And since its warranty data reflects that about a third of pumps that were replaced were misdiagnosed as failures, it contends that his calculations would be unreliable even if he had done so. ECF No. 121, PageID.37285-86.
In Kondash, the specific component at issue was a panoramic sunroof, comprised of three panels. Kondash, 2022 WL 5816228, at *8. The expert calculated the sunroof failure rate based on sales of individual replacement panels, without regard for whether one, two, or three panels were required to repair the sunroof of a class vehicle. Id. In addition to potentially double or triple counting the number of sunroof failures, the expert's calculations made no effort to account for sunroof repairs unrelated to defective design, such as leaking, wind, or car crashes. Id. In excluding that expert's report and opinions, the court noted that there was also evidence that the expert had ignored calculation methods that were more widely accepted in the industry. Id.
But Kondash is only authoritative to the extent it is persuasive, and given the different circumstances here, it is not. GM's criticisms of Dr. Edgar's calculations do not expose a defect in reliability so grave as the potential double or triple counting component failures in Kondash. As Plaintiffs noted during argument, here there is only one replacement component that provides the basis for Dr. Edgar's calculation-i.e., one replacement pump per pump failure, rather than one to three panels per defective sunroof. Hr'g Tr., ECF No. 151, PageID.41039. To be sure, Dr. Edgar's calculations do not readily account for pump failures unrelated to the alleged design defect (such as customer negligence). He simply assumes that such failures are “de minimis.” ECF No. 121, PageID.37285-86. The correctness of these assumptions goes towards the weight of Dr. Edgar's calculations, however, not their admissibility. “Disputes about the accuracy of a theory's results, generally speaking, provide grist for adversarial examination, not grounds for exclusion.” United States v. Gissantaner, 990 F.3d 457, 464 (6th Cir. 2021) (internal quotations omitted).
iii. Use of secondary sources
GM further argues that large portions of Dr. Edgar's report merely “summarize secondary sources and record evidence.” ECF No. 121, PageID.37288-90. Information presented “without any expert analysis or other application of the expert's expertise” generally should come to the jury directly from the sources rather than through the mouth of the expert. In re Davol, Inc./C.R. Bard, Inc., Polypropylene Hernia Mesh Prod. Liab. Litig., 546 F.Supp.3d 666, 677 (S.D. Ohio 2021). But “expert testimony that relies on expert knowledge and experience to contextualize, analyze, and interpret historical facts is admissible.” Id. GM does not specify what portions of the report it believes are inadmissible under this standard. And after its own review of Dr. Edgar's report, the Court concludes that it is highly technical, related to his expertise, and would provide helpful context and analysis for a jury, as well as the Court.
Almost as an aside, GM contends that Dr. Edgar does not have any basis to opine that GM frequently denied warranty claims resulting from out-of-spec fuel. ECF No. 121, PageID.32787-88; see ¶ 101, ECF No. 111-2, PageID.21090. On this point, the Court agrees. Dr. Edgar states in his deposition that he does not have any basis for such an opinion. Edgar Dep. 175:1-8, ECF No. 121-2, PageID.37333. Because Dr. Edgar does not arrive at this conclusion based on his own analysis, he will not be permitted to offer an opinion on this alleged practice.
Subject to the qualification above, the Court concludes that GM has not raised sufficient challenges warranting the exclusion of Dr. Edgar's testimony under either Federal Rule of Evidence 702 or 403.
B. Defendant's Motion to Exclude Stockton Report (ECF No. 119/120)
1. Summary of qualifications and testimony
Plaintiffs engaged Stockton to evaluate whether it is possible to calculate the typical and aggregate expenses for the “failure/repair group”-i.e., consumers who paid for repairs after catastrophic failures. ECF No. 111-34, PageID.32370; ECF No. 129-2, PageID.38172. Assuming an affirmative answer, he was asked to describe the data and methods he would use to make those calculations.
The day before Stockton submitted his original report, GM advised Plaintiffs that some of the data files it had provided contained errors. Once the new files were provided, Stockton issued an update to his report. ECF No. 119-2. The original report contains the most thorough explanation of his methods and calculations, while the update and declaration contain his most up-to-date findings. There is not a significant difference between the Stockton's original and updated figures, which Plaintiffs point to as evidence of the soundness of his methodology in excluding erroneous or outlier data.
Stockton has a bachelor's degree in economics from Western Michigan University, and a master's degree from the Department of Agricultural and Resource Economics at the University of Arizona, where his concentration was applied econometrics. ECF No. 111-34, PageID.32404. He has worked at Fontana Group, Inc. since 1998, where he has held positions as an analyst, senior analyst, senior financial analyst, case manager, Director of Economic Services, and now Vice President. Id. His work includes over 100 studies concerning economic problems in the automotive industry. Id. at PageID.32371.
In his report, Stockton explains that he arrived at his conclusions using the following steps: (1) determining what goods and services are required for a repair; (2) determining what data sources concerning repair costs are available; (3) determining how that data was prepared and making relevant adjustments; and (4) performing calculations using the data to estimate repair costs. Id. at PageID.32375-76. He drew his data from GM's warranty records, GM's internal repair cost estimates, dealer records of customer pay repairs, GM's records of dealers' customer pay and warranty labor rates, and GM's internal CP4 repair bulletins. Id. at PageID.32378. He also referenced Dr. Edgar's calculations. Id. at PageID.32380. His report describes which data he believes to be the most reliable, the assumptions he made in working with that data, and how he prepared it. Id. at PageID.32380-89.
Stockton's report concludes that calculating the typical and aggregate damages of the “failure/repair” group is feasible. He has created tables of average repair costs and estimates the aggregate damages across the repair class. ECF No. 129-2, PageID.38181. His calculations are meant to take into account available warranty protections. The report also proposes a method for allocating damages among multiple owners of vehicles which never experienced failures, to compliment Gaskin and Weir's findings. ECF No. 111-34, PageID.32400-02. This method would base each owner's share on relative participation in the vehicle's ownership.
2. Challenges to testimony
a) Data selection and “cleansing”
GM first challenges Stockton's damages calculations for the “failure/repair” group, attacking his choice of data and how he prepared it. In his report, Stockton notes that he relied primarily on two data sets: GM's warranty data file, and GM's customer pay data file. ECF No. 111-34, PageID.32388. He describes how he “cleansed” this data, excluding records unless: (1) the vehicles at issue were class vehicles; (2) the part numbers corresponded to part numbers at issue in the case; and (3) the listed part costs were less than $2,000 (in other words, far below the part cost for a CP4 pump). Id. at PageID.32389. He also explains his process for identifying outlier records and testing the sensitivity of his calculations using fuller data sets. Id. at PageID.32389-91.
GM attacks Stockton's data cleansing methods as unreliable and argues that his calculations are not adequately tethered to the facts of the case. ECF No. 119, PageID.36717-20. It asserts that Stockton's data “cleansing” led him to omit thousands of repair records and arbitrarily exclude another 50% of those remaining, such that the data underlying his calculations no longer accurately reflects the record. Echoing its complaints about Dr. Edgar, GM additionally argues that Stockton's calculations are unreliable because he used data that may reflect instances where there was not actually a pump failure. Id. at PageID.36710-14. According to GM, Stockton's calculations are therefore simultaneously underinclusive and overinclusive.
In supplemental briefing, GM again relies on Stevens v. Ford Motor Co., 2:18-CV-00456 (S.D. Tex. Sept. 25, 2022), as authority weighing in favor of exclusion. In Stevens, the court excluded Stockton's testimony and opinions under Federal Rule of Evidence702, concluding that Stockton's methodology was “incomplete” because he apparently did not supply any preliminary analysis or calculations. But Stockton has supplied both analysis and calculations in this case, so the Court finds Stevens unpersuasive.
But none of GM's challenges are specific or concerning enough to warrant exclusion of Stockton's testimony. Cleansing of statistical data is commonplace; any data set will have outliers or irrelevant information that will skew calculations if not removed. See In re RFC & RESCAP Liquidating Tr. Action, 332 F.Supp.3d 1101, 1146 (D. Minn. 2018) (“As a general matter, statistical sampling is a commonly used and accepted means of assembling and analyzing data, particularly in complex litigation.”). And as GM admitted during oral arguments, some data “cleansing” was necessary because GM's data production included numerous records not relevant to Plaintiffs' claims. Hr'ing Tr., ECF No. 141, PageID.41056.
The Court concludes that Stockton's methodology is sufficiently based in economic principles and industry-backed to be reliable. In his report, Stockton explains that his decision to focus on the middle 50% of the dataset was part of a standard “interquartile range” test, designed to eliminate statistical outliers and focus on averages. ECF No. 111-34, PageID.32390. Moreover, he did not rely exclusively on data culled through this exercise; GM's arguments ignore that he also conducted calculations with the fuller dataset to test the sensitivity of his methods. Id. He offered plausible and transparent reasons for culling the data as he did. GM's criticisms of his methods and the assumptions underlying his calculations may affect how the jury should weigh his opinions, but they do not render it inadmissible. See Bledsoe v. FCA U.S. LLC, 4:16-cv-14024-TGB-RSW, 2022 WL 4596156, at *29 (E.D. Mich. Sept. 30, 2022) (Berg, J.) (“An expert is permitted to make reasonable assumptions, which Stockton has done here. … Any difference in opinion about those assumptions should be resolved by a jury and is not a proper basis to strike his opinions.”); Teenier v. Charter Commc'ns, LLC, No. 16-CV-13226, 2017 WL 3141051, at *4 (E.D. Mich. July 25, 2017) (Drain, J.) (when there is “some sort of explanation, the issue becomes one of weight, which should be left to the trier of fact.”)
GM's criticism that Stockton's damages model is not sufficiently tied to Plaintiffs' liability theory because the underlying data may reflect instances where there was not actually a pump failure again goes to weight rather than admissibility. What's more, GM wrongly frames Plaintiffs' theory as being only about U.S. diesel fuel. It appears to suggest that Stockton should have included only repairs that were identified as being related to out-of-spec fuel. ECF No. 119, PageID.36710, PageID.36719-20. But the record does not establish that data with this level of granularity was available. And Plaintiffs' theory of liability is that the design of the CP4 pump itself is defective. Stockton's model therefore assumes that any repair relates to the defect alleged. GM is fully capable of challenging and a jury is capable of understanding the implications of this assumption.
b) Allocation of “overpayment” damages
GM also challenges Stockton's proposed model for allocating overpayment damages across class vehicles with multiple owners, arguing that it is unreliable, unsupported by economic principles, and unworkable in practice. ECF No. 119, PageID.36720-23. It asserts that Stockton's proposed method was previously rejected under Evidence Rule 702 by another court in Sloan v. General Motors, No. 16-cv-07244-EMC, 2020 WL 1955643, at *48 (N.D. Cal Apr. 23, 2020).
The Sloan decision on which GM relies did not exclude Stockton's overpayment model under Rule 702; it did not conduct a Rule 702 analysis at all. Instead, in the context of a class certification motion, it asked whether former owners and lessees of class vehicles would ever be entitled to damages as a legal matter. After answering this question in the negative, it modified the proposed class definition accordingly. Sloan, 2020 WL 19555643, at *48. This inquiry does not neatly map onto whether expert testimony is admissible under Daubert.
As with his “failure/repair” class calculations, Stockton clearly and transparently set forth his proposed methodology for allocating overpayment damages. And in any event, damages allocation is not something that is generally resolved by courts at the class-certification stage. Counts v. Gen. Motors, LLC, 606 F.Supp.3d 547, 578 (E.D. Mich. 2022) (Ludington, J.) (collecting cases). GM's additional criticism that Stockton cannot explain how individual damages would be allocated assuming this class were to be certified is similarly flawed. ECF No. 119, PageID.36723-24. Stockton is meant to be an expert as to the issue of class-wide damages, nothing more. ECF No. 129, PageID.38159; see also In re Whirlpool Corp. Front-Loading Washer Prod. Liab. Litig., 722 F.3d 838, 861 (6th Cir. 2013) (noting that individual damages calculations do not preclude class certification).
Stockton's opinions and report are sufficiently based on reliable data and methods-including the factual record, basic economic principles, and industry-standard data, and they are helpful to the Court. GM's motion to exclude Stockton will therefore be denied.
C. Defendant's Motion to Exclude Gaskin and Weir Reports (ECF No. 122)
1. Summary of qualifications and testimony
Plaintiffs engaged Gaskin to develop a market research survey and analysis to determine the reduction in market value of class vehicles resulting from the allegedly defective fuel pump. ECF No. 111-32, PageID.31790. Weir assisted Gaskin in developing the survey and was also asked to develop and validate a framework for calculating overpayment damages in this lawsuit-i.e., the damages owed to consumers who have not experienced a fuel pump failure but who allegedly suffered a diminution in the market value of their vehicle on account of the pump or consumers who received a no-cost repair under warranty. ECF No. 111-33, PageID.32326.
Gaskin is an independent survey expert who holds Bachelor of Science and Master of Science degrees in Management from the Sloan School of Management at the Massachusetts Institute of Technology. ECF No. 111-32, PageID.31788. He has co-authored several articles in peer-reviewed publications, including in Marketing Science and Management Science. Id. Additionally, he has given several conference presentations on aspects of conjoint analysis. Id.
Weir, meanwhile, holds a Masters of Business Administration degree from the High Technology program at Northeastern University in Boston and a Bachelor of Arts in Business Economics from the College of Wooster, Ohio. ECF No. 111-33, PageID.32324. He is the Vice President of the firm Economics and Technology, Inc., where he conducts economic, statistical, and regulatory research and analysis. Id. at PageID.32341. He has consulted on a variety of consumer and wholesale products cases, calculating damages relating to automobiles, food products, household appliances, herbal remedies, health/beauty care products, electronics, and computers. Id.
Gaskin's report describes the survey methodology he used, which is choice-based conjoint analysis. Id. at PageID.31790. He explains that conjoint analysis is a widely accepted method of determining the relative market value of a product in light of disclosure or non- disclosure of a defect at the point of sale. Id. at PageID.31790-91. Rather than asking consumers directly how much they would pay for a product, conjoint analysis asks them to choose between various options and extrapolates the value of different features based on their answers. Id. at PageID.31792. Gaskin administered the survey he developed to a total of 400 respondents. Id. at PageID.31802. Based on its results, Gaskin concluded that the reduction in class vehicle market value from a change in “diesel with fuel pump that works reliably for the life of the truck” to “diesel with fuel pump that leads to increased engine wear but will not lead to engine failure” is 9.4%. Id. at PageID.31791.
Weir's report elaborates on the economic validity of using conjoint analysis to estimate overpayment damages and also describes how he worked with Gaskin on developing portions of the survey. ECF No. 111-33, PageID.32327-35. He details economic supply-side considerations taken into account in designing the survey-including that the quantity of vehicles supplied was known and fixed, historical price data was available, the vehicle market was highly competitive, and GM ultimately did not control the retail price paid by consumers. Id. at PageID.32331-35. The report also describes Weir's use of GM's data and Dr. Edgar's calculations to estimate the overpayment damages. ECF No. 111-33, PageID.32335-38. Id. Using Gaskin's 9.4% reduction in market value figure, Weir estimates the overpayment damages across the class to be $1,608,647,106. Id. at PageID.32338.
2. Challenges to testimony
As a preliminary matter, the Court notes that GM does not appear to argue that conjoint analysis is not an accepted methodology in the field of market research. Conjoint surveys like the one challenged here have routinely been approved a means to estimate overpayment damages of car purchasers in class action cases involving allegedly defective automobile components. See, e.g., In re MyFord Touch Consumer Litig., 291 F.Supp.3d 936, 943 (N.D. Cal. 2018). GM's challenges focus on its specific application.
a) Consideration of supply-side factors
GM first attacks Gaskin and Weir's survey as not adequately accounting for supply-side considerations. It acknowledges that Gaskin and Weir's analysis uses real-world data in their analysis but contends that the use of such data is fundamentally erroneous. In short, it asserts that the analysis wrongly assumes a fixed and stable supply of vehicles, despite evidence of a reduction in demand. ECF No. 122, PageID.37433. It further argues that, even if Gaskin and Weir's assumptions about supply are appropriate, the price data used to inform the survey is flawed because it consists of dealer invoice prices, manufacturer suggested retail prices (MSRPs), and public websites to estimate actual market prices for the conjoint survey. Id. at PageID.37437-38. According to GM, reliance on this information is inappropriate because marketplace realities, such as discounts and negotiation, mean that hardly anyone pays the MSRP when they purchase a vehicle. Id. at PageID.37433-39.
GM relies on a number of district court decisions, which it says have rejected similar analyses-including analyses done by Gaskin and Weir-as not adequately incorporating supply-side factors. But only one of these decisions, In re Volkswagen “Clean Diesel” Mktg., Sales Pracs., & Prod. Liab. Litig., 500 F.Supp.3d 940, 949 (N.D. Cal. 2020), excludes a conjoint survey as unreliable under in the context of a Daubert motion. Another decision rejects a similar analysis as a “failure of proof” at the summary judgment stage. See In re Gen. Motors LLC Ignition Switch Litig., 407 F.Supp.3d 212, 236 (S.D.N.Y. 2019). In a third decision from this district, Weir and Gaskin's method of analysis was found to withstand Daubert scrutiny-although as the case progressed, the court eventually concluded that Gaskin and Weir's analysis was insufficient to support the predominance analysis at the class certification stage. See Schechner v. Whirlpool, No. 16-12409, ECF No. 168 (admitting expert reports); 2019 WL 4891192, at *7-8 (E.D. Mich. Aug. 13, 2019) (denying class certification) (Murphy, J.).
Many courts have approved of the ways in which Gaskin and Weir incorporated supply-side factors into their analysis under Daubert. See, e.g., Fitzhenry-Russell v. Dr. Pepper Snapple Grp., Inc., 326 F.R.D. 592, 605 (N.D. Cal. 2018); In re Dial Complete Mktg. & Sales Practices Litig., 320 F.R.D. 326, 332 (D.N.H. 2017). These courts have found that the factors are reliably incorporated when (as here): “(1) the prices used in the surveys underlying the analyses reflect the actual market prices that prevailed during the class period; and (2) the quantities used (or assumed) in the statistical calculations reflect the actual quantities of products sold during the class period.” Hadley v. Kellogg Sales Co., 324 F.Supp.3d 1084, 1105 (N.D. Cal. 2018). In supplemental briefs, Plaintiffs flag two recent decisions rejecting arguments like the ones GM makes here, including one from this district. Won v. General Motors LLC, No. 19-11044, 2022 WL 3010886, at *5 (E.D. Mich. July 28, 2022) (Lawson, J.); Johnson v. Nissan N. Am., Inc., No. 3:17-cv-00517-WHO, 2022 WL 2869528, at *7 (N.D. Cal. July 21, 2022).
As the Johnson decision notes, objections that conjoint survey analyses do not adequately account for supply-side factors have become routine. Johnson, 2022 WL 3869528, at *7. The dispute over how to appropriately incorporate these factors has come up in district courts across the country-often in cases involving these same two experts. See, e.g., Cardenas v. Toyota Motor Corp., No. 18-22798-CIV, 2021 WL 5811741, at *4 (S.D. Fla. Dec. 6, 2021) (summarizing dispute and collecting cases). Some courts conclude that Gaskin and Weir's methods do not adequately take into account these factors, whereas others conclude that they incorporate certain factors and leave the question of which factors matter for juries. Compare In re Volkswagen “Clean Diesel” Mktg., Sales Pracs., & Prod. Liab. Litig., 500 F.Supp.3d at 949 (“[P]resuming that Defendants would have sold the same number of cars, at the exact price that consumers would have been willing to pay, is not a way to reliably incorporate supply-side considerations.”), with Hadley, 324 F.Supp.3d at 1105-06 (concluding that conjoint analysis sufficiently accounted for supply-side factors by considering actual market prices and actual sales during the class period). The Sixth Circuit has not considered the issue.
This Court finds the latter line of cases persuasive. The key issue for the Court is whether, as required by Daubert, the analysis reliably determines diminution to the class vehicles' reasonable market value caused by the allegedly defective pump. Gaskin and Weir's reports set out which supply-side factors they considered and offer reasonable explanations for the assumptions they made. Their analysis is informed by historical marketplace and quantity data. GM offers nothing to suggest that the analysis deviates in any significant way from the usual principles by which a conjoint analysis survey is designed. And since there are multiple accepted ways of conducting a conjoint analysis, any dispute with their method goes to its weight, not its admissibility. See Beaty v. Ford Motor Co., No. C17-5201 TSZ, 2021 WL 3109661, at *5 (W.D. Wash. July 22, 2021) (concluding that defendant “simply disagrees with Gaskin and Weir about whether Plaintiffs' proposed conjoint model … should incorporate certain supply-side information” and reasoning that “[s]such disputes go to the weight that should be afforded to the experts' model, not its admissibility”).
Plaintiffs do not respond in their briefing to GM's challenge regarding Gaskin and Weir's use of MSRP and dealer invoice data instead of what consumers ended up paying for their cars as a measure of “actual market prices” in the consumer survey. But Weir notes in a rebuttal declaration that GM's experts could not identify any transaction price for a class vehicle that fell “outside the range of prices used in the Gaskin survey.” ECF No. 130-8, PageID.38442. So even if Weir and Gaskin are subject to criticism for their use of dealership invoices and MSRPs to develop their price ranges, they did not include ranges so off-base as to render their analysis meaningless. At all events, other courts have previously approved the use of MSRPs as reliable in this context. See, e.g., Johnson, 2022 WL 2869528, at *7. And the purpose of a conjoint analysis is to determine the price differential between what was paid and what would have been paid but for the alleged misrepresentation, no matter the baseline. Johannesohn v. Polaris Indus., Inc., 450 F.Supp.3d 931, 973 (D. Minn. 2020) (concluding that, in “seek[ing] to find the difference between Polaris's prices in the actual world … and in the Plaintiffs' counterfactual world, it “is not unreasonable to assume that MSRP data may be used to capture this difference-i.e., the overcharge percentage-because there is some correlation between MSRP and actual price. That is, if [the expert] captured an 8.8 percent change in MSRP, then it is reasonable to assume that mirrors an 8.8 percent change in actual price”).
b) Other methodological flaws
GM mounts a number of additional methodological objections to Gaskin's survey-attacking his choice of attributes included as options as unverified, his written descriptors as biased, and his sample as unrepresentative. ECF No. 122, PageID.37439-43. But district courts across the country have recognized that these sorts of challenges go to weight, rather than admissibility. See, e.g., Cardenas, No. 2021 WL 5811741, at *5 (“Notwithstanding the potential merit of these objections … district courts … have recognized that they go to weight, not admissibility”); FCOA, LLC v. Foremost Title & Escrow Servs., LLC, No. 17-2391-Civ-WILLIAMS/TORRES, 2019 WL 416817, at *4 (S.D. Fla. Feb. 1, 2019) (objections like poor sampling and poorly designed questions are “technical deficiencies that affect the survey's weight” but “not its admissibility”). These are arguments for a jury.
c) “Fit” with plaintiffs' liability theory
GM further attacks Gaskin and Weir's analysis as being inconsistent with Plaintiffs' liability theory. ECF No. 122, PageID.37443-44. Zeroing in on survey options that asked participants to compare “diesel with fuel pump that works reliably for the life of the truck” with “diesel with fuel pump that leads to increased engine wear but will not lead to engine failure,” it argues that the survey choices do not adequately reflect liability premised on the notion that every truck will eventually have catastrophic failure. But Gaskin and Weir were engaged to calculate damages for an “overpayment” group. By definition, this group includes only individuals who never actually experience a catastrophic engine failure. The survey logically models the circumstances of this particular group.
d) Reliability of class-wide damages calculation
GM lastly attacks Weir's class-wide damages calculation as unreliable. ECF No. 122, PageID.37444-47. Its first challenge is that, to the extent Weir relies on figures calculated by Gaskin and Dr. Edgar, it does not have an adequate factual basis in the record. But the Court has concluded that Gaskin's and Dr. Edgar's testimony and opinions will be admitted, so this challenge is unavailing.
GM also argues that Weir's calculations are unreliable because his calculations are based on a “mean price paid” for the putative class vehicles, calculated using dealer invoice prices, as opposed to the actual prices paid by consumers. Id. at PageID.37445-46. But in the absence of information about actual transaction prices, which is unavailable, the use of averages derived from sources such as MSRPs and dealership invoices is a reasonable proxy measure. GM may conduct cross-examination on the implications of this use of averages, but this methodology is not so unsound as to warrant exclusion under Rule 702.
GM's motion to exclude the opinions and testimony of Gaskin and Weir will be denied.
D. Plaintiffs' Motion to Exclude Harrington Report (ECF Nos. 124/125)
1. Summary of qualifications and testimony
Harrington was retained by GM to respond to Dr. Edgar's report and testimony. Specifically, he was asked to “review documents and warranty data, perform analysis, and provide opinions related to the development, deployment, and performance of the CP4 fuel pump used in vehicles containing the GM Duramax engine,” to “respond to Dr. Edgar's opinions on these topics,” and to “evaluate whether potential failures of the CP4 fuel pumps … can be assessed based on common characteristics.” ECF No. 116-7, PageID.33916.
Harrington holds a Master of Science degree in Automotive Engineering from the University of Michigan, Ann Arbor. Id. at PageID.33917. He is presently the Principal of Exponent Inc., an engineering consulting firm. Id. Before joining Exponent, Harrington worked first as a senior engineer and then the Chief of the Technology Innovation and Policy Division at the United States Department of Transportation's Volpe National Transportations Systems Center, where he oversaw teams studying transportation technology. Id. at PageID.33917-18. And before that, he was a technical support manager at Cummins Inc., where he analyzed failure and warranty data, conducted investigations to identify root causes of fuel system failures, and assessed out-of-warranty reports. Id. He is a peer reviewer at Department of Energy's annual merit review and a volunteer design judge for the Society of Automotive Engineers' Formula Hybrid Competition. Id. at PageID.33918.
To prepare his report, Harrington and his staff reviewed Dr. Edgar's report, a transcript of his deposition, and materials he requested and received from counsel. Id. at PageID.33923, PageID.34021; ECF No. 144-3, PageID.40823, PageID.40849-50. He also spoke with Mr. Peter Hubl and Mr. Nallaya Chinnusamy, two of GM's design engineers. ECF No. 144-3, PageID.40824.
Harrington's report begins by evaluating diesel fuel and diesel engine technology, as well as the CP3 and CP4 pumps. ECF No. 116-7, PageID.33928-37. It then reviews GM's development and testing process for the CP4 pump and finds it to be appropriate and within industry standards. Id. at PageID.33938-72. Ultimately, the report concludes that Dr. Edgar's failure rate calculations are based on faulty assumptions and incomplete review of data, and it disagrees with his theory regarding the “root cause” of the CP4 pump design defect. Id. at PageID.33919-23, PageID.34001-04.
2. Challenges to testimony
Plaintiffs first argue that Harrington's testimony does not fit the facts of this case and should therefore be excluded as irrelevant. They criticize his characterization of the “root cause” of the defect in the CP4 pump and his analysis of Dr. Edgar's report, saying that his mischaracterization renders the rest of his testimony unreliable. ECF No. 124, PageID.37737-40. But Plaintiffs do not cite to any case law that persuasively indicates why Harrington's testimony should be excluded altogether when they essentially criticize him for doing what a rebuttal expert is hired to do: challenge the characterizations and conclusions of another expert. No parts of the report are so untethered from the facts as to make them inadmissible under Rule 702.
Similarly, Plaintiffs state that Harrington provides testimony that is inconsistent with the evidence in the case, and that he should be excluded on that basis. ECF No. 124, PageID.37740-44. They point to a number of purported “inconsistencies” between his deposition testimony and other evidence in the case. But a close reading of the alleged materials in question reveals that none of the issues raised are truly direct contradictions, merely differences in opinion. All of these differences can be inquired into on cross-examination.
The remainder of Plaintiffs' motion involves attacks on Harrington's practice generally as a “defense expert” and accuse him of bias in favor of GM. ECF No. 124, PageID.37744-46. Plaintiffs also cite a decision from this district in which his opinions were excluded in part. See Counts, 2022 WL 2078023 at *23-26. In that decision, Judge Ludington excluded Harrington's opinion in part because the court concluded that Harrington did not have the experience or qualifications to opine regarding a specific type of emissions testing at issue.
But Plaintiffs cannot make that kind of allegation here-indeed, some of the criticisms Plaintiffs make toward Harrington's experience are the same that GM directed against Dr. Edgar. Harrington, like Dr. Edgar, has several years of experience with diesel technology, and the Court cannot find that he is unqualified to offer an opinion on the subject matter at issue. As for Plaintiffs' accusation of bias, this is a traditional ground for cross-examination. See In re Welding Fume Prods. Liab. Litig., 534 F.Supp. 761, 766 (N.D. Ohio 2008) (“[A]bsent a showing of bias so extreme that exclusion is appropriate under Daubert, the Court believes that disclosure of possible financial bias coupled with cross-examination by the parties is a more appropriate and fine-tuned mechanism for arriving at the truth.”) Plaintiffs would have this Court conflate the conclusions of other judges about Harrington and his firm rendered in different contexts with the question of whether he has sufficient expertise to offer an opinion in this case. This is not an appropriate basis upon which to exclude an expert.
Plaintiffs' motion will be denied.
IV. CONCLUSION
For the reasons explained above, the motions to exclude the opinions of Dr. Bradley L. Edgar (ECF No. 121), Edward Stockton (ECF Nos. 119/120), Steven Gaskin and Colin Weir (ECF No. 122), and Ryan Harrington (ECF Nos. 124/125) are DENIED. Neither party has established that the testimony of their opponents' putative experts is so unreliable as to warrant exclusion under Federal Rule of Evidence 702.
IT IS SO ORDERED.