Snodgrass v. Ford Motor Company

Full title: TERI SNODGRASS, ROBERT L. BAKER, KENDALL ELLIS, JILL P. FLETCHER, JUDITH…

Court: United States District Court, D. New Jersey

Date published: Mar 28, 2002

Civil Action No. 96-1814 (JBS) (D.N.J. Mar. 28, 2002)

Opinion

Civil Action No. 96-1814 (JBS)

March 28, 2002

Philip Licht, Esquire, Mont Laurel, NJ, and Beverly C. Moore, Jr., Esquire, MOORE BROWN, Washington, D.C., Attorneys for Plaintiffs.

C. Scott Toomey, Esquire, CAMPBELL, CAMPBELL, EDWARDS CONROY, P.C., Turnersville, NJ, Attorneys for Defendants Ford and UTA, Inc.

Brian C. Anderson, Esquire, Neil Gilman, Esquire, O'MELVENY MYERS, LLP, Washington, D.C., Attorneys for Defendant Ford Motor Co.

Scott L. Winkelman, Esquire, CROWELL MORING LLP, Washington, D.C., Attorney for Defendant United Technologies Automotive, Inc.

OPINION

JEROME B. SIMANDLE, United States District Judge

This action was brought on behalf of plaintiffs alleging property damage resulting from fires caused by defective ignition switches, which were manufactured by defendant United Technologies Corp., formerly known as United Technologies Automotive, Inc., ("UTA") and installed in automobiles sold by defendant Ford Motor Company. The Court's opinion filed September 4, 2001 granted in part and denied in part both plaintiffs' and defendants' summary judgment motions, leaving seven plaintiffs to proceed to trial. The Court granted partial summary judgment as to causation in favor of plaintiffs Kendall Ellis, Jill Fletcher, Frank Sherron, and Ten Snodgrass. The issue of causation remains for trial, however, for plaintiffs Tamaz Tal, Robert Baker, and Judith Shemnitz. Also, the issue of damages must be tried for all remaining plaintiffs.

Presently before the Court are defendants' motion for severance and to dismiss plaintiffs' claims for lack of subject matter jurisdiction; defendants' motion in limine to preclude evidence of Recall 95S28; defendants' motion in limine to exclude the testimony of plaintiffs' experts Dr. Jack Moshman and Mr. William Hagerty; and plaintiff's motion to admit the database of vehicle fire complaints allegedly caused by defective ignition switches.¹

1 The Court will, in a separate opinion, be addressing plaintiffs' motion for reconsideration of the September 4, 2001 order; plaintiffs' motion for leave to file supplemental summary judgment reconsideration reply regarding demand letter issue; and defendants Ford's and UTA's motion for reconsideration of the September 4, 2001 order denying summary judgment in the case of Kendall Ellis.

On September 10, 11, 12, 13, and 25, 2001, the Court held a Federal Rule of Evidence 104 hearing to determine the admissibility of the expert testimony and the database.² For reasons discussed below, defendants' motion for severance and dismissal of plaintiffs' claims for lack of subject matter jurisdiction will be denied; defendants' motion in limine to preclude evidence of Recall 95S28 will be denied as to plaintiffs Baker, Sherron and Ellis; defendants' motion in limine to exclude the expert testimony of Dr. Jack Moshman and Mr. William Hagerty will be granted as to Dr. Jack Moshman, and will be denied as to Mr. William Hagerty; plaintiffs' motion to admit the database and plaintiffs' motion for admission of narrowed database will be denied.

2 Rule 104, Fed.R.Evid., states that "[p]reliminary questions concerning the qualifications of a person to be a witness . . . or the admissibility of evidence shall be determined by the court." Fed.R.Evid. 104. "A proponent of expert testimony must establish his expert is qualified and his testimony is admissible by a preponderance of the evidence." See Poust v. Huntleigh Healthcare, 998 F. Supp. 478, 490 (D.N.J. 1998) (citing Daubert v. Merrill Dow Pharm., Inc., 509 U.S. 579, 593 n. 10 (1993)).

I Defendant's Motion for Severance and to Dismiss Plaintiffs' Claims for Lack of Subject Matter Jurisdiction

A. Standard for Permissive Joinder

Defendants argue that plaintiffs fail to satisfy the requirements of Rule 20(a), Fed.R.Civ.P., which allows permissive joinder of plaintiffs in an action, and therefore this Court should sever and dismiss their claims for lack of subject matter jurisdiction. Rule 20(a) permits joinder of parties if:

[T]hey assert any right to relief jointly, severally, or in the alternative in respect of or arising out of the same transaction, occurrence, or series of transactions or occurrences and if any question of law or fact common to all these persons will arise in the action.

Fed.R.Civ.P. 20(a). At a minimum, Rule 20(a) requires that the central facts of each plaintiff's claim arise on a somewhat individualized basis out of the same set of circumstances. See In re Orthopedic Bone Screw Products Liab. Litig., MDL 1014, WL 428683, at *2 (E.D. Pa. July 17, 1995) (citing Gruening v. Sucic, 89 F.R.D. 573, 574 (E.D. Pa. 1981); King v. Pepsi Cola Metro. Bottling Co., 86 F.R.D. 4, 6-7 (E.D. Pa. 1979)).

Rule 20(a)'s "transaction or occurrence" requirement is applied liberally by courts in order to achieve the rule's goals of promoting judicial economy and efficiency. See Bone Screw Products, WL 428683, at *2. "The purpose of Rule 20(a) is to 'promote trial convenience and expedite the final determination of disputes, thereby preventing multiple law suits.'" Miller v. Hygrade Food Products Corp., 202 F.R.D. 142, 144 (E.D. Pa. 2001) (quoting C.A. Wright et al., Federal Practice and Procedure § 1652, at 395 (3d ed. 2001)). Permissive joinder thus falls within the sound discretion of the court and is to be liberally granted because "[u]nder the rules, the impulse is toward entertaining the broadest possible scope of action consistent with fairness to the parties; joinder of claims, parties and remedies is strongly encouraged."See Miller, 202 F.R.D. at 144 (quoting United Mine Workers v. Gibbs, 383 U.S. 715, 724 (1966)).

B. Analysis

Defendants argue that the claims of plaintiffs should be severed because they arise out of circumstances too different to meet the "transaction or occurrence" requirement of Rule 20, citing, inter alia, the twelve different Ford models, seven different model years, five different states in which the vehicles were purchased, and five different cars with different mileages.³ Plaintiffs assert that the "series of transactions or occurrences" requirement of Rule 20(a) is satisfied by the likelihood that all of plaintiffs' vehicle fires were caused by the Fox ignition switch.

3 Defendants cite to the twelve plaintiffs' claims that were asserted prior to this Court's opinion filed September 4, 2001, which dismissed certain of plaintiffs' claims upon defendants' summary judgment motions. Accordingly, at this time, the claims of seven plaintiffs, Baker, Ellis, Fletcher, Shemnitz, Sherron, Snodgrass, and Tal, proceed to trial.

In the instant matter, defendant Ford manufactured all vehicles with the same Fox ignition switch, produced by defendant UTA, issued the same warranties, and in most cases, made the same representations to plaintiffs regarding the defective condition of their vehicles, specifically denying the existence of a problem with the Fox ignition switch in some cases. The cumulative evidence in this five-year case has shown that the actions of the vehicle manufacturer, Ford, in manufacturing of its product and in post-sale communications with plaintiffs, are very likely a central issue to all of plaintiffs' claims, most of which involve breach of warranty and fraud. Former Judge Bechtle's opinion in In re Orthopedic Bone Screw Products Liab. Litig., MDL 1014, 1995 WL 428683 (E.D. Pa. July 17, 1995) is persuasive in this regard. In Bone Screw Products, plaintiffs from different states went to different doctors and medical facilities, and underwent surgery at different times, yet were all implanted with the same medical device produced by a certain manufacturer. The court, applying Rule 20(a), stated:

Plaintiffs . . . represented to the court that the developing discovery in this case appears to justify linking manufacturers, distributors, and providers, and possibly doctors and others, in a confederation at best and a conspiracy at worst, in regard to the marketing of manufacturers' products that could be a central issue in this case. If that is so, or believed to be so, and good grounds exist to support such allegations, the "series of transactions" requirement of Rule 20 might be satisfied for filing purposes at least. . . .

Bone Screw Products, 1995 WL 428683, at *2. Likewise, Ford's knowledge and conduct in its manufacture and post-sales communications regarding its vehicles equipped with the Fox ignition switch, if proved at trial, are central to all of plaintiffs' claims, which involve implied warranty, enhanced damages under Mass. Gen. Laws 93(A) § 9(3), common law fraud by omission, fraudulent concealment, negligence, and UDAP by omission. This central issue is sufficient to satisfy the "series of occurrences or transactions" that Rule 20(a) requires in its liberal application.

Additionally, while it is true that the seven claims proceeding to trial require "case-by-case litigation of whether each putative plaintiff's damages actually were caused by defendants' switches," as stated by this Court in its previous opinion, In re Ford Motor Co. Ignition Switch Products Liab. Litig., 194 F.R.D. 484, 492 (D.N.J. 2000),⁴ the number of claims to be addressed at trial are not so numerous and complex as to prevent swift and efficient adjudication of such claims. After the Court's opinion and order of September 4, 2001, the claims of only seven plaintiffs will proceed to trial.⁵ Even though plaintiffs' claims are distinct, the common issues will require presentation of much of the same evidence, while testimony concerning the particular vehicle fire may involve less than one day per plaintiff.⁶ In addition, after five years of litigation in this case, proceeding to trial under these terms would avert further delay for these remaining plaintiffs, which Rule 20(a) is designed to prevent. In light of the Court's broad discretion under Rule 20(a), the Court finds that permissive joinder of plaintiffs' claims is appropriate and the remaining claims shall proceed to trial.

4 This determination had been made in the context of plaintiffs' renewed motion for class certification, which this Court denied, citing the inability of plaintiffs' statistical evidence to sufficiently prove causation in its discussion on whether factual issues predominate. See Ignition Switch Products, 194 F.R.D. at 490-91.

5 The various plaintiffs' claims will include implied warranty, enhanced damages under Mass. Gen. Laws 93A § 9(3), common law fraud by omission, fraudulent concealment, negligence, UDAP by omission. See Op. of 9/4/01 at 52.

6 The details and logistics for addressing the remaining claims at trial will be addressed further at the final pretrial conference, adopting sensible procedures for a complex case trial under Rule 16(c), Fed.R.Civ.P., potentially including time limitations and grouping of evidence.

Accordingly, defendants' motion for severance and dismissal of plaintiffs' claims for lack of subject matter jurisdiction is denied, and plaintiffs' claims shall proceed to trial.

II. Defendant's Motion in Limine to Preclude Evidence of Recall 95S28

A. Standards for Relevance and Preludice

Defendants move in limine to preclude evidence of Recall 95S28, pertaining to ignition switches, pursuant to Rules 402 and 403 of the Federal Rules of Evidence because they claim that evidence of the recall is irrelevant and prejudicial. According to Rule 402, Fed.R.Evid., "[e]vidence which is not relevant is not admissible." Relevant evidence, under Rule 401, "means evidence having any tendency to make the existence of any fact that is of consequence to the determination of the action more probable or less probable than it would be without the evidence." Fed.R.Evid. 401.

"Although relevant, evidence may be excluded if its probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury, or by considerations of undue delay, waste of time, or needless presentation of cumulative evidence." Fed.R.Evid. 403. Thus, evidence may be excluded when its admission would lead to litigation of collateral issues, thereby creating a side issue which might distract the jury from the main issues. Blancha v. Raymark Indus., 972 F.2d 507, 516 (3d Cir. 1992) (citing United States v. Dennis, 625 F.2d 782, 797 (8th Cir. 1980)). Evidence also may be excluded if it relates to previous litigation involving the parties, has an aura of scientific infallibility but is likely to be used for purposes other than that for which it was introduced, or relates to statistical probabilities. See Blancha, 972 F.2d at 516 (citing J.B. Weinstein M.A. Berger, Weinstein's Evidence, 6403 (1988)).

The Third Circuit has noted that "[e]vidence should be excluded under Rule 403 only sparingly since the evidence excluded is concededly probative." Blancha, 972 F.2d at 516 (citing United States v. Terzado-Madruga, 897 F.2d 1099, 1117 (11th Cir. 1990)). The balance under the rule is struck in favor of admissibility. See Terzado-Madruga, 897 F.2d at 1117; Dennis, 625 F.2d at 797. In determining the probative value of evidence under Rule 403, the court "must consider not only the extent to which it tends to demonstrate the proposition which it has been admitted to prove, but also the extent to which that proposition was directly at issue in the case." United States v. Herman, 589 F.2d 1191, 1198 (3d Cir. 1978), cert. denied, 441 U.S. 913 (1979).

B. Analysis

Defendants assert that evidence of Recall 95S28 should be excluded because this case involves vehicles that were not subject to the recall campaign, and evidence that certain model years of certain vehicle models were subject to recall for potential ignition switch problems has no logical connection to plaintiffs' efforts to prove that their vehicles had defective ignition switches. Plaintiffs argue that only three of the plaintiffs' vehicles were not subject to the recall (Sherron, Baker, and Ellis). Further, plaintiffs state that they have never contended that the recall could establish Ford's liability. However, plaintiffs do believe that this evidence should be introduced to show that Ford should have recalled the pre-1988 model year vehicles, as well as the post-1988 vehicles, because these vehicles were equipped with the same defective Fox ignition switch.

Safety Recall 95S28 addresses "Ignition Switch Replacement" for specific Ford model years and makes.⁷ (Service Recall Bulletin, Def.'s Br. Ex. B.) Safety Recall 95S28 regarding Ignition Switch Replacement thus does not apply to plaintiff Frank Sherron's vehicle (1992 Ford Ranger); plaintiff Baker's vehicle (1986 Ford Aerostar); or plaintiff Kendall Ellis's vehicle (1987 Ford F-iSO). The recall does apply to the vehicles of plaintiffs Jill Fletcher, Ten Snodgrass, Tamaz Tal, and Judith Shemnitz, and its admissibility as to these plaintiffs is not contested.

7 The Ford Recall 95S28 recalls the following vehicles:
ALL 1988 — 1990 Model Year Escort
ALL 1988 — Model Year EXP

ALL 1988 — 1993 Model Year Mustang built through September 30, 1992

ALL 1988 — 1993 Model Year Tempo and Topaz built through September 30, 1992

ALL 1988 — 1993 Model Year Thunderbird/Cougar built through September 30, 1992

ALL 1988 — 1989 Model Year Crown Victoria and Grand Marquis

ALL 1988 — 1989 Model Year Lincoln Town Car
ALL 1988 — 1991 Model Year Aerostar
ALL 1988 — 1991 Model Year Bronco and F-Series Light

In order to determine if the recall is relevant to the claims of plaintiffs Baker, Ellis, and Sherron, we look to the elements of their claims. Baker and Sherron assert breach of implied warranty of merchantability claims, while Ellis asserts common law fraud by omission and fraudulent concealment under Mississippi state law. As to plaintiff Ellis's claim, he asserts fraud by omission, which requires plaintiff to prove (1) a representation, (2) its falsity, (3) its materiality, (4) the speaker's knowledge of its truth, (5) his intent that it should be acted upon by the hearer and in the manner reasonably contemplated, (6) the hearer's ignorance of its falsity, (7) his reliance on the truth, (8) his right to rely thereon, and (9) his consequent and proximate injury. See Franklin v. Lovitt Equip. Co., Inc., 420 So.2d 1370, 1373 (Miss. 1982) (citations omitted). The issue of Ford's knowledge of ignition switch problems in vehicles listed in the recall notice is relevant to determining whether Ford made material misrepresentations, that is, told plaintiff there were no problems when, in fact, it knew of ignition switch problems. Also, the recall is relevant to show that Ford had the capacity to notice that a trend of ignition switch problems was occurring in its vehicles due to substantial complaints, and that it took action to correct it. Thus, evidence of the recall is relevant to an issue that is central to Mr. Ellis's case, that is, whether Ford made material misrepresentations to plaintiff because it could have known that it had a problem with certain ignition switches not included in Safety Recall 95S28. In addition, evidence of the recall is relevant to show what Ford thought it should do in cases involving defects with its vehicles. Thus, the evidence is relevant and will not be excluded under Rule 402, Fed.R.Evid., as to plaintiff Ellis.

Plaintiff Baker brings an implied warranty claim under Massachusetts law. To invoke the implied warranty of merchantability under Mass. Gen. Laws ch. 106, § 2-314, a plaintiff must demonstrate that the damages complained of were proximately caused by a defect or breach which existed at the time of the sale. Fernandes v. Union Bookbinding Co., 507 N.E.2d 728, 734 (Mass. 1987) (citing Walsh v. Atamian Motors, Inc., 406 N.E.2d (Mass.App.Ct. 1980)). Here, the recall could tend to prove that the breach or defect existed at the time of sale. The fact of the recall and the reasons for it in models having the same type of ignition switch as Baker's car, would, with other evidence of the similarity between Baker's switch and the recalled ones, tend to prove that such switches can contain defects that were present at the time of sale. Evidence that the concern about manufacturing or design defects in the same type of device in other Ford vehicles led to a recall of similar vehicles is a legitimate source of inference that such switches in other vehicles can fail due to defect. The recall evidence is not relevant, however, to the issue of whether Baker's fire was caused by an ignition switch, because the existence of a recall does not help to establish that a given automobile fire was caused by a defective part. Causation of Baker's fire must be proved by the facts of his individual case and such evidence of general applicability as may be otherwise available. Defendants' motion to preclude evidence of Recall 95S28 will be denied.

In plaintiff Sherron's case, in which causation has already been established, plaintiff Sherron asserts breach of implied warranty of merchantability, under the Missouri statute, Mo. Rev. Stat. § 400.2-314, which requires that a plaintiff prove (1) that a merchant sold goods, (2) which were not "merchantable" at the time of sale, (3) injury and damages to the plaintiff or his property, (4) which were caused proximately or in fact by the defective nature of the goods, and (5) notice to the seller of the injury. Ragland Mills, Inc. v. Gen. Motors Corp., 763 S.W.2d 357, 360 (Mo.Ct.App. 1989) (citing Metty v. Shurfine Cent. Corp., 736 S.W.2d 527, 530 (Mo.Ct.App. 1987)). Here, the evidence of Recall 95S28 could tend to prove the existence of the element of "merchantability," because the evidence reveals the possibility that Sherron's ignition switch was not reasonably fit for its purpose. Due to the recall, it is not far-fetched for a person with a demonstrated ignition switch fire to claim that this fire was due to a more widespread type of defect in the ignition switch, which had led to a voluntary recall. This evidence will be admissible for this purpose in Sherron's case.

Although defendants claim that the admission of evidence of the recall is more prejudicial than probative and that such evidence should therefore be excluded under Rule 403, this Court thinks otherwise. Recall 95S28 will be admitted as evidence in the claims of these remaining plaintiffs. While it is conceded that evidence of the safety recall is prejudicial to Ford, it has not been shown that this particular evidence is unfairly prejudicial to Ford, which is required for exclusion under Rule 403, Fed.R.Evid. According to the Advisory Committee's Notes, the term "unfair prejudice" within its context means an undue tendency to suggest decision on an improper basis, commonly, though not necessarily, an emotional one. See Advisory Committee's Notes to Fed.R.Evid. 403. Contrary to defendants' suggestion, introduction of such evidence does not portend a foregone conclusion in plaintiffs' favor, nor does it suggest decision in plaintiff's favor due to an emotionally-charged issue, such as gender-based derogatory comments in a sex discrimination case. See Sheridan v. E.I. DuPont de Nemours Co., 100 F.3d 1061, 1076 (3d Cir. 1996), cert. denied, 521 U.S. 1129 (1997) (affirming exclusion of evidence under Rule 403); see also United States v. DePeri, 778 F.2d 963, 973 (3d Cir. 1985), cert. denied, 475 U.S. 1110 (1986) (affirming exclusion of evidence of past trauma regarding losing a child as evidence tending to show defendant did not threaten the life of any child in the instant matter). Considering that plaintiffs must still satisfy their burdens of proof at trial, it is unlikely that introduction of the recall for the purposes permitted above would lead to its use for improper purposes. These claims should not be litigated in a vacuum devoid of the context of one of the largest vehicle recalls in history. Because admission of the recall does not unfairly prejudice defendants, such evidence will not be excluded under Rule 403.

In addition, it is anticipated that the jury will be given clear limiting instructions of the purposes for which the recall may be considered for the remaining plaintiffs. Here, the Court is confident that the jury is capable of distinguishing the applicability of the evidence of the safety recall as to particular plaintiffs' claims. Further, the Court is confident that introduction of such evidence will neither mislead nor confuse the jury.

In summary, evidence of Recall 95S28 is sufficiently relevant to proof of the claims of plaintiffs Ellis, Baker and Sherron pursuant to Rule 402. In addition, evidence of Recall 95S28 is more probative than unfairly prejudicial under Rule 403. Accordingly, defendants' motion to preclude evidence of Recall 95S28 will be denied as to plaintiffs Ellis, Baker and Sherron.

III. Defendants' Motions in Limine to Exclude Exrert Testimony of Dr. Jack Moshman and Mr. William Hagerty

Defendants move to exclude the testimony of plaintiffs' experts Dr. Jack Moshman and Mr. William Hagerty. Defendants first argue that Dr. Moshman's testimony does not satisfy the relevance requirements of Rule 401. Second, defendants argue that Dr. Moshman's testimony would create unfair prejudice and undue delay and should therefore be inadmissible under Rule 403. Third, defendants argue that Dr. Moshman's testimony is not sufficiently reliable to meet the standards for expert testimony as provided by Rule 702. For the reasons set forth below, defendants' motion to exclude Dr. Moshman's testimony under Rule 702 will be granted, the Court's discussion will therefore be limited to admissibility under that rule.⁸

8 The Court notes that the testimony may be irrelevant under Rule 401 and 402, considering that Dr. Moshman has conceded that statistics do not prove causation, Tr. 9/12/01 at 113, thus making its tendency to prove causation, a disputed issue in this case, unlikely. It is also recognized that the statistical evidence may create considerable confusion in the jury under Rule 403, leading to a more prejudicial than probative effect.

A. Standard for Expert Testimony

The United States Supreme Court has held that the Federal Rules of Evidence, rather than Frye's "general acceptance" test, provides the proper standard for the admission of expert scientific testimony in a federal trial. See Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 585 (1993) (citing Frye v. United States, 293 F. 1013, 1014 (1923)). Under the Federal Rules of Evidence, the trial judge acts as "gatekeeper" to ensure that any and all expert testimony is relevant, as well as reliable. Kannankeril v. Terminix Int'l, Inc., 128 F.3d 802, 806 (3d Cir. 1997). Rule 702, which governs the admissibility of expert testimony, provides:

If scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert . . . may testify thereto in the form of an opinion or otherwise, if (1) the testimony is based upon sufficient facts or data, (2) the testimony is the product of reliable principles and methods, and (3) the witness has applied the principles and methods reliably to the facts of the case.

Rule 702, Fed.R.Evid. The Third Circuit has construed this rule to contain three main requirements: (1) the proffered witness must be an expert; (2) the expert must testify about matters requiring scientific, technical, or specialized knowledge; and (3) the expert's testimony must assist the trier of fact. See In re Paoli R.R. Yard PCB Litig., 35 F.3d 717, 741-42 (3d Cir. 1994). These three requirements are often referred to as "qualification, reliability, and fit." Poust v. Huntleigh Healthcare, 998 F. Supp. 478, 490 (D.N.J. 1998); Rutigliano v. Valley Bus. Forms, 929 F. Supp. 779, 783 (D.N.J. 1996).

Under the first requirement, the Court must (1) determine whether the proffered expert has minimal qualifications through either education or experience in a field that is relevant to a subject that will assist the trier of fact; and (2) compare the expert's area of expertise with the proffered opinion to ensure its helpfulness to the jury. Poust, 998 F. Supp. at 491 (citing Federal Judicial Center, Reference Manual on Scientific Evidence 55 (1994)).

The Third Circuit has interpreted the second requirement to mean that "an expert's testimony is admissible so long as the process or technique the expert used in formulating the opinion is reliable." Paoli, 35 F.3d at 742 (citing Daubert, 509 U.S. at 589). Daubert requires the expert to testify to scientific knowledge, meaning that the expert's opinion must be based on the "methods and procedures of science," rather than on "subjective belief or unsupported speculation." Paoli, 35 F.3d at 742.⁹ The expert must have "good grounds" for his or her belief. Id. Further, the Third Circuit has construed Daubert to hold that "the reliability of scientific evidence under Rule 702 requires a determination as to its scientific validity." Paoli, 35 F.3d at 742.

9 In determining whether an expert's methodology is reliable, the Third Circuit considers several factors, including: (1) whether a method consists of a testable hypothesis; (2) whether the method has been subject to peer review; (3) the known or potential rate of error; (4) the existence and maintenance of standards controlling the technique's operation; (5) whether the method is generally accepted; (6) the relationship of the technique to methods which have been established to be reliable; (7) the qualifications of the expert witness testifying based on methodology; and (8) the non-judicial uses to which the method has been put. Kannankeril, 128 F.3d at 807 n. 6 (citing Paoli, 35 F.3d at 742 n. 8) (combining Daubert and Downing factors).

Under the third requirement, Rule 702 requires that the expert's testimony must assist the trier of fact. Paoli, 35 F.3d at 743. "[A]dmissibility depends in part on 'the proffered connection between the scientific research or test result to be presented and particular disputed factual issues in the case.'" Paoli, 35 F.3d at 743 (quotingUnited States v. Downing, 753 F.2d 1224, 1237 (3d Cir. 1985)). Thus, an expert's proposed testimony must be "for purposes of the case." Paoli, 35 F.3d at 743 (emphasis in original). Furthermore, the requirement of "reliability," or "good grounds," extends to each step in an expert's analysis all the way through the step that connects the work of the expert to the particular case. Paoli, 35 F.3d at 743.

B. Dr. Moshman's Testimony

Plaintiffs offer Dr. Moshman to testify on the issue of causation.¹⁰ Dr. Moshman's report and hearing testimony indicate that his testimony in this case will consist of statistical opinions regarding the fire rates of two subgroups of Ford automobiles with Fox ignition switches as compared to fire rates of non-Fox-equipped vehicles. In forming his opinion, Dr. Moshman relied solely on a database compiled by plaintiffs' counsel of approximately 7, 670 fires including alleged ignition switch fires. (Tr. 9/12/01 at 113; Moshman Rpt. at 3.)

10 Dr. Moshman concedes, however, that statistics cannot be used to prove causation in an individual case. (Tr. 9/12/01 at 113.)

First, Dr. Moshman states in his report that Fords equipped with the Fox ignition switches, listed in the database, could be subclassed into two fire rate groups, with each group being internally homogeneous. (Moshman Rpt. at 3.) According to Dr. Moshman, based on the revised database of 7, 670 fires, see Moshman Rpt. Ex. G., Subclass 1¹¹ had a fire rate of 5.895 fires per 10,000 vehicles, and subclass 2¹² had a fire rate of 2.432 fires per 10,000 vehicles. (Id. at 5.) Second, Dr. Moshman also provides the opinion that the ignition switch fire rates for the two subgroups of Fox vehicles were substantially and significantly higher than the fire rates for non-Fox-equipped Fords. (Id.) Using the fire rates of 139 MORS II¹³ non-Fox vehicles in the database of 6,539 fires previously analyzed in his second affidavit, Dr. Moshman states that the fire rate for non-Fox vehicles is 0.432 fires per 10, 000 vehicles, while the fire rate for Fox ignition switch cars (based on the current database's 3,403 fires derived from MORS II reports) is 1.669 fires per 10,000 vehicles. (Id.) Dr. Moshman states that this fire rate of Fox-equipped cars is 3.865 times as large as the rate for non-Fox-equipped vehicles, and a statistical comparison of the two rates using a T-test revealed a highly significant difference with a p-value of significantly less than one in a thousand. (Moshman Rpt. at 5.)

11 Subclass 1 consists of the following Ford models and years: Town Car, 1984-89; Aerostar, 1986-91; Mustang, 1986-93; Bronco, 1984-91; F-Series, 1985, 87-91; and Grand Marquis, 1985-89.

12 Subclass 2 consists of the following Ford models and years: Crown Victoria, 1985-89; Thunderbird, 1984-93; Escort, 1984, 86-90; Tempo, 1985-93; Topaz, 1984-93; Cougar, 1984-93; Bronco II, 1988; Explorer, 1991-93; Continental, 1984-87; Mark VII, 1985-89, 91; Ranger, 1985-93; Econoline, 1985-88, 92-93; Lynx, 1986.

13 Dr. Moshman states that of the 7, 670 vehicle fires allegedly attributable to defective ignition switches in the current database, he had based his calculations on 3, 403 entries that had been derived from MORS II reports, thus "making them comparable to the 139 MORS nonclass fires." Moshman Rpt. at 5.

Dr. Moshman states that the difference between each of the two subclasses of the reduced class and non-Fox-equipped vehicles is also statistically significant, using only MORS II data of the subgroups in this calculation for purposes of comparison. (Id.) He states that the fire rate for Subclass 1 is 2.54 fires per 10,000 vehicles, and the rate for Subclass 2 is 1.15 fires per 10,000 vehicles. (Id.) Dr. Moshman's report also indicates that the rate for Subclass 1 is 5.9 times as large as the nonclass vehicles, and the rate for Subclass 2 is 2.7 times as great as the nonclass vehicles. (Id.)

1. Analysis

Dr. Moshman qualifies as an expert in statistics, and defendants so stipulated. (Tr. 9/12/01 at 136.)¹⁴

14 Dr. Moshman's qualifications and experience in the field of statistics are extensive. He holds a doctorate in mathematical statistics from the University of Tennessee and a master's degree in mathematical statistics from Columbia University. (Moshman Rpt. App. A.) He has also served on the faculty of the University of Tennessee and Queens College, has published several articles in the field of statistics, and has been retained as an expert on statistics in trial proceedings. Dr. Moshman has designed, conducted, and analyzed statistical surveys for both government and industry, and has been a member and officer of several professional statistical associations. Defense counsel in fact stipulated to Dr. Moshman's qualifications and credentials at the Rule 104 Hearing. Accordingly, the Court accepted and qualified Dr. Moshman as an expert in statistics at the Rule 104 hearing. (Tr. 9/12/01 at 136.)

Under the second requirement of Rule 702, Dr. Moshman must show that the process or technique he used in formulating his opinion is reliable.Paoli, 35 F.3d at 742. The issues to be discussed here are (a) Dr. Moshman's sole reliance on plaintiffs' database, see Moshman, Tr. 9/12/01 at 113; Moshman Rpt. at 5; and (b) Dr. Mosliman's methodology in calculating the fire rates for Subclass 1 and Subclass 2.

a. Dr. Moshman's Reliance on the Database

In essence, the database is being created and introduced for the purpose of showing a trend in fire rates of vehicles equipped with the Fox ignition switch.¹⁵ Pl.'s Mot. Admit Database Br. at 15. The database at the time of the Rule 104 hearing was composed of 7,562 complaints of fires possibly attributable to a defective ignition switch, and later was reduced to 6,227 complaints as asserted in Plaintiffs' Motion for Admission of Narrowed Database.¹⁶,¹⁷ Pl.'s Br. Admit Database at 8. However, as shown below, it has never been fully verified that the causes of the fires entered into the database were defective ignition switches. Mr. Hagerty discussed the genesis of the database, explaining that the alleged fires were initially included in the database based on information provided in

15 For this Court's discussion of plaintiffs' motions to admit the database and the narrowed database, see discussion infra Part W.

16 For purposes of Dr. Moshman's report, the database contained 7, 670 entries at the time his opinions were formulated. (Moshman Rpt. at 5.)

17 Plaintiffs' Motion for Admission of Narrowed Database asserts that the database at this time consists of 6, 227 complaints, due to 303 net deletions (321 deletions less 18 additions). This motion does not change this Court's present analysis, which relies upon the submissions of parties as of the date of the Rule 104 hearing.

the State Farm database¹⁸,¹⁹ wand probably every third or fourth one of those they have had a fire cause and origin expert go out and investigate that case before they decided to put it in their database. . . . [I]n many cases . . . the insurance company's office has gone out to determine that they were cause and origin ignition switch fires. . . . They're in there because someone had decided they were fires including the insurance investigators. . . . They didn't get in there unless State Farm had already decided independently of Mr. Moore's office that they were ignition switch fires. And in many of them, there are private investigators and public fire reports attached.

18 Testimony of Mr. Hagerty revealed that State Farm filed suit on behalf of the people whose vehicle fires were included in their database to recover subrogation for moneys paid those people. (Hagerty, Tr. 9/10/01 at 142.)

19 Approximately 650 database entries comes from State Farm/CSAA insurance data, and about 2, 100 database entries come from ASES fire reports. (Hagerty Rpt. at 6.) In addition, other fire reports were obtained through Ford discovery, such as Ford Office of the General Counsel complaints, expert reports, and from independent sources, including a website entitled "Flaming Fords." See Moshman Rpt. at 3.

(Hagerty, Tr. 9/10/01 at 140, 141.) Although initially included in the database because "someone" had thought that these fires were caused by a defective ignition switch, Mr. Hagerty and others on behalf of plaintiffs have since initiated a review of the entries in order to determine whether they should or should not be included in the database. (Hagerty, Tr. at 141.)

Specifically, Mr. Hagerty testified that he has personally reviewed the documents underlying between 2, 200 and 2, 500 entries in order to determine whether specific fires should be included or excluded from the database. (Tr. 9/10/01 at 138.) Several thousand questionnaires had previously been sent to database members for whom plaintiffs had full mailing addresses, asking questions such as: (1) where did the fire start; (2) was there an inspection or report by a fire department, insurance company, independent mechanic, or Ford investigator as to the fire cause; and (3) other relevant questions. (Hagerty Rpt. at 5.) These questionnaires were distributed by Rust Consulting, Inc., in September 2000. (Id.) After Rust forwarded information to Hagerty, he then reviewed each questionnaire response and any attached documentation, including photographs, fire inspection reports, and correspondence, to advise in his expert opinion whether (1) it appeared to be an ignition switch fire, (2) it looked like it might be, but more information should be sought directly by counsel as to whether it was; or (3) it looked like it was not an ignition switch fire. (Id.) Thus, entries are to remain in the database until Mr. Hagerty reviews them and advises whether the complaint should be included, excluded, or followed up with a questionnaire. Hagerty, Tr. 9/10/01 at 142.

Mr. Hagerty testified, "I'm trying to look independently at each fire to determine whether or not I believe it was more than likely caused by the ignition switch without regard to what vehicle it belonged to or what year." (Hagerty, Tr. 9/10/01 at 143.) In addition, employees of plaintiffs' counsel's firm, Mr. Alexander and Ms. Kelly Best, have conducted reviews of the underlying documents for approximately 100 database entries. (Tr. 9/10/01 at 138, 140.) Mr. Alexander and Ms. Best were trained in the following manner: they worked under the tutelage of Mr. Hagerty for two days to look for certain words and photos that indicated a fire caused by the ignition switch. When the indicated factors were present during their review, Mr. Alexander and Ms. Best "put them in one stack, [and] when they are not present, they're sending them to [Mr. Hagerty] to be looked at." (Tr. 9/10/01 at 138.) Mr. Hagerty stated that although the two employees had no engineering background, there were "numerous phone calls and conversations on individual cases" between them. (Tr. 9/10/01 at 139.)

Thus, the database is composed of entries to be reviewed or already reviewed by Mr. Hagerty or the two employees of Mr. Moore's office, thereby consisting of the following two categories: (1) fires more likely than not caused by a defective ignition switch, and (2) entries that have not been reviewed yet, thus including both (a) fires probably not caused by a defective ignition switch, and (b) fires that may be attributable to a defective ignition switch but are unverified. As to the certainty of ignition switch causation among the fires allegedly confirmed in the 2, 500 or so entries reviewed by Mr. Hagerty, the certainty "varies from minimal information to excellent private fire report, photographs, witness statements, all that sort of thing, because the quality of the reports vary tremendously." (Hagerty, Tr. 9/10/01 at 143.)

The concern of the Court is twofold: that Dr. Moshman's opinion of fire rates will be introduced for the purpose of proving causation (admittedly a probable issue under Rules 402 and 403), and his statistical opinions regarding fire rates are based on complaints in the database, thus operating on the wrongly held assumption that the plaintiffs' database is an accurate compilation of verified fires caused by ignition switch defects. The fires were initially included in the database based on fire reports of tremendously variable reliability and are being reviewed after the fact and solely for purposes of this litigation to determine whether they indeed should be "included." See Hagerty, Tr. 9/10/01 at 139 ("The understanding is that they're already included in the database and they're looking at them to be included."). Using a database the entries of which may or may not be fires attributable to ignition switch defects as the foundation for Dr. Moshman's opinions regarding the fire rates of such entries suggests that the "existence and maintenance of standards controlling the technique's operation," required by Rule 702 andKannankeril, 128 F.3d at 807 n. 6, is inherently lacking. If the database is not an accurate portrayal of what it purports to be — namely, a compilation of actual incidents of fires caused by ignition switches — then Dr. Moshman's use of this database, which assumes the conclusions of ignition switch fires to be accurate, is methodologically unsound.

Methodical assessment throughout the database was not conducted to determine that its contents were composed entirely of vehicle fires caused by an ignition switch defect. Essentially, there were no definite standards that controlled whether vehicle fires were to be initially included in the database, besides the fact of the inclusion of the vehicles from the State Farm database. Even so, not all of the database entries came from State Farm. Other sources include Ford, ASES reports, MORS reports, and the internet site entitled "Flaming Fords." Notwithstanding further concerns for the reliability of these sources,²⁰ purification of at least all of the database entries using Mr. Hagerty's method would have provided for uniformity. Here, however, as of the Rule 104 hearing, Mr. Hagerty had looked at only 2,200 to 2,500 entries, leaving at best approximately 5,000 of the total 7,500 unreviewed. Although plaintiffs emphasize that 303 deletions have been made since that time, in its Plaintiffs' Brief in its belated Motion for Narrowed Database, recently filed, plaintiffs fail to clearly indicate the total number of complaints reviewed by Mr. Hagerty thus far. Indeed, a brief glance at the proposed database indicates that the entirety of its contents still has not yet been reviewed. Dr. Moshman himself has testified that he has done nothing to verify the contents of the database. (Moshman, Tr. 9/12/01 at 113.) Dr. Moshman's opinions regarding calculations of fire rates using rates of vehicle fires, purported to be caused by defective ignition switches, in a database that has been only partially verified, thus appear to rest on a process or technique that is unreliable.²¹

20 Plaintiffs have taken steps to address the reliability of these sources. See Pl.'s Mot. Narrowed Database Br. at 10-12. However, confirmation of a percentage of the complaints does not fully address these reliability concerns.

21 One of Dr. Moshman's opinions involves only fire rates of vehicles from MORS II report sources. Even these reports have not been fully substantiated. Plaintiffs rely on the fact that of about 2,000 of the over 3,000 database people who sent back questionnaires, "most were MORS II people." Pl.'s Mot. Admit Database Br. at 9. To the extent that Dr. Moshman's opinion utilizes over 3,000 MORS II reports which includes both verified and unverified data in calculating what were expected to be normalizing rates, this, too, provides little reliability.

Plaintiff responds to defendants' arguments regarding the flawed database by asserting that there are "ongoing efforts to verify that each person on the present database actually had an ignition switch fire, making the database as finally refined prior to trial admissible as reliable evidence. . . . [M]ost questionable people have already been eliminated (and more will be prior to trial). . . ." Pl's Mot. Admit Database Br. at 2. This Court, however, stated that the July 20, 2001 database was final for trial "unless it's amended by order of the Court." Tr. 9/12/01 at 199-201. Plaintiffs' Motion for Admission of Narrowed Database, discussed further in Part IV, does nothing to persuade the Court that plaintiffs' proposed experts have completely reviewed the entirety of the database's contents to reasonably assure the accuracy of the entries.

In the five years since this litigation began, plaintiffs have had the time and capability to ensure that the database and its contents were verified as to what it purports to be, that is, records of fires demonstrably caused by ignition switch defects. The remoteness of the fire events and the incompleteness of meaningful, reliable data for so many events, has required plaintiffs to attempt the task of garnering reliable expert opinion on the cause and origin of a myriad of fires lying far beyond the discrete events at issue in plaintiffs' own vehicle fires.

Since Dr. Moshman's opinions rest completely on the factual assumption that all incidents in the database are ignition switch fires, as distinct from just reports of fire events that may or may not involve the ignition switch, the inclusion of an incident in the database amounts to an expression of opinion, for purposes of litigation, which must be based upon a reasonable certainty employing the relevant discipline of determining the cause and origin of suspected fires. Confirmation of the data would have ensured that the standards for controlling the database were maintained properly and that Dr. Moshman's opinions, which relied solely on the correctness and completeness of the database, were grounded in sound data. Lacking that, plaintiffs have submitted a database the contents of which provided Dr. Moshman's expert opinion with little more than "unsupported speculation." Daubert, at 590.

In addition, the database was compiled and developed by employees of plaintiffs' counsel and Mr. Hagerty for the purposes of litigation. It does not derive — in terms of the opinions expressed — from some reliable historical source. The database's creation aided by the help of two employees with two days of training and oversight ensures little reliability that the fires included in the database were caused by a defective ignition switch, given that such conclusions are presumably normally made by experts in the field. Indeed, Mr. Hagerty's expert testimony in this case would not be required if determinations of cause and effect in vehicle fires could be made so easily by lay persons. Unfortunately, original investigatory data are lacking, or the on-the-scene investigators did not reach conclusions as to cause and origin being located in the ignition switch. While entries deemed questionable by Mr. Hagerty or by plaintiffs' counsel were deleted, other entries have not been examined. Because Dr. Moshman's testimony relies solely on the database, the majority of which consists of entries of alleged events in which a reliable determination of correctness (i.e., that the event was an ignition switch fire has not been made), this weighs in favor of finding Dr. Moshman's testimony unreliable and ultimately not useful to the jury.

b. Dr. Moshman's Methodology in Calculating Subgroup Fire Rates to Support Internal Homogeneity

Furthermore, this Court is hesitant to find that Dr. Moshman's methodology in calculating the fire rates for Subclass 1 and Subclass 2 comports with the factors outlined in Daubert and Downey. See Kannankeril, 128 F.3d at 807 n. 6. In his opinion, Dr. Moshman first determined that the entries in the database were not homogeneous in their statistical fire rates. (Tr. 9/12/01 at 114.) Plaintiffs' counsel then excluded from the set those Ford vehicle classes equipped with Fox ignition switches that had low to zero fire rates, after which Dr. Moshman split the records into two groups, which, he testifies, each internally exhibited homogeneity. (Moshman, Tr. 9/12/01 at 114.) Three groups were thus created: the low/zero fire rates, middle fire rates, and high fire rates. (Tr. 9/12/01 at 114.) Dr. Moshman explained his method of ascertaining the homogeneity of the high and middle fire rates:

I examined the data, looked to see what the rates were, tried to see whether or not the entire — what you call the high and middle group, whether they constituted a homogeneous group within themselves. And I found they did not according to the tests that I used. . . . I then made an arbitrary division into two subsets and each of these subsets then passed my requirements for exhibiting homogeneity internally. These are not the only way they could be divided, but they served the purpose.

Moshman, Tr. 9/12/01 at 115 (emphasis added).

Here, Dr. Moshman relies upon an "arbitrary" division in order to "serve the purpose" of achieving homogeneity within the subclasses. This Court questions whether this methodology, used to find internal homogeneity for two subgroups, consists of a hypothesis that is capable of being tested over and over again. See Kannankeril, 128 F.3d at 807 n. 6 (listing "whether a method consists of a testable hypothesis" as a factor in determining the reliability of an expert's methodology). Given that Dr. Moshman acted arbitrarily in excluding some vehicle classes entirely and in defining the remaining subsets, the Court finds it doubtful that another expert could ascertain the same conclusion, due to the innumerable subclasses that may exist within the set. Dr. Moshman's continuous arbitrary divisions into two subsets instead suggests groupings that were intentionally selected in order to result in rates that bolstered the purpose for which it is being introduced, causation in this case. Thus, it appears rather that Dr. Moshman determined the conclusion before the hypothesis was put forth, in light of his testimony that he tried different subclasses to find the sought-after homogeneity attributes. The Court agrees with defendants' supplemental affidavit of statistician William Wecker, who demonstrated in his first affidavit that Dr. Moshman subjectively chose models and model years to include or exclude from his reduced class, with the help of plaintiffs' counsel, and that his arbitrariness is demonstrated by the changes in composition of the "reduced class" between his first, second, and third affidavits. See Wecker Supp. Aff. ¶ 15. The Court is also persuaded by defendants' argument that Dr. Moshman relied on the incorrect assumption that the various years of a particular model should be considered successive samples, and that these vehicles in successive model years share the same underlying incident rate. This is not just a clash of viewpoints between experts about two equally reliable ways of looking at the data, but it is rather a convincing showing by defendants that Dr. Moshman's subgrouping methodology is not methodologically justifiable. The subjective inclusion and exclusion of data suggest that Dr. Moshman manipulated the data to achieve a desired result. This weighs against the reliability of Dr. Moshman's methodology.

Under the factors to consider in determining whether an expert's methodology is reliable, Dr. Moshman's method in forming his opinion does not consist of a testable hypothesis because he had determined his conclusion regarding the homogeneity of the subgroups' fire rates before formulating his hypothesis; he has relied upon incorrect assumptions in calculating his statistical opinions; the vehicle fire complaints in the database, on which Dr. Moshman solely relies, were not and have not been subject to reliable scientific analysis to support identifying all events as actual ignition switch fires; and the database, created solely for purposes of this litigation, has not been put to any non-judicial uses.²² These factors weigh against reliability of Dr. Moshman's expert opinion, and Dr. Moshman's testimony will therefore not be admissible under Rule 702. Accordingly, defendants' motion in limine will be granted on this ground, and neither this determination nor the reduced database may be admitted into evidence in this trial as a basis for any expert opinion.

22 For a complete list of factors to consider in determining whether an expert's methodology is reliable, see supra note 9.

C. Mr. William Hagerty's Expert Testimony

Plaintiffs offered Mr. Hagerty as a cause and origin expert, as well as an expert in the design of motor vehicle components. (Tr. 9/10/01 at 23.) Mr. Hagerty holds a bachelor of arts degree in economics from Lafayette College and a master's degree in business administration from the Wharton School of Business at the University of Pennsylvania. (Hagerty Rpt. App. 1.) The Court finds that Mr. Hagerty is qualified as an expert under Rule 702 in the cause and origin of fires. Mr. Hagerty owned and operated an independent wholesale/retail automotive store from 1987 to 1993, and then worked as a consultant from 1993 to 1995. (Hagerty Rpt. App. 1.) Since 1995, Mr. Hagerty has worked as a forensic consultant and automotive fires, tires and brakes expert. (Hagerty Rpt. App. 1.) Mr. Hagerty has attended approximately twelve fire courses, including mechanic and vehicle fire, (Tr. 9/10/01 at 12), and has given training to groups including the Insurance Association and the Association of Arson Investigators. (Tr. 9/10/01 at 12.) In addition to the 911 investigations he has conducted (for both plaintiffs and defendants) in the role of lead investigator, (Hagerty Rpt. App. 1), Mr. Hagerty has also assisted in 100-150 fire investigations. (Tr. 9/10/01 at 13.) In addition, Lee Cole, a leading authority and author of "An Investigation of Motor Vehicle Fires," often enlisted Mr. Hagerty's assistance in teaching vehicle fire courses and had assigned Mr. Hagerty's office about three or four cases each month. (Pl.'s R. 104 Hrg. Ex. 5.; Tr. 9/10/01 at 14.) Mr. Hagerty also testified that Lee Cole is referenced as an authoritative source in the National Fire Protection Association 921 ("NFPA"), which is the overall guide for fire and explosion investigations. (Tr. 9/10/01 at 16.) Accordingly, because of his extensive experience in the field of automotive fires, the Court qualified Mr. Hagerty on the issue of cause and origin of vehicle fires, satisfying the first requirement of the Rule 702. (Tr. 9/10/01 at 39.)

On the issue of whether Mr. Hagerty may be offered as an automotive design expert, Mr. Hagerty does not possess any education or training regarding the design of automotive components. (Id. at 38.) His knowledge of component design comes exclusively from the facts developed in this case and other cases he has worked on, rather than from previous experience or education. (Id.) In addition, Mr. Hagerty has not conducted an independent study on design issues or engineering issues in this case. (Id.) As a result, this Court did not qualify Mr. Hagerty as an expert on the issue of automotive design at the Rule 104 hearing, and will not do so here. (Id. at 38.) The Court does permit Mr. Hagerty, however, to explain the mechanism to the extent that it comes from opinions regarding the design as expressed by Dr. Malladi, McCarthy, and Ford and UTA documents. (Id.)

Under the second requirement of Rule 702, the Court must then determine whether "the process or technique the expert used in formulating the opinion is reliable." Paoli, 35 F.3d at 742 (citing Daubert, 35 F.3d at 742). Plaintiffs offer Mr. Hagerty to provide a causation opinion as to the remaining plaintiffs Baker, Tal, and Shemnitz, for whom causation still remains at issue. In reaching his opinions as to plaintiffs Baker, Tal, and Shemnitz, Mr. Hagerty relied on the 1998 version of the NFPA 921. (Tr. 9/10/01 at 103.) Mr. Hagerty testified that he "generally" followed the concepts the 1998 edition of NFPA 921 "[t]o the best of [his] ability," in reaching his opinions. (Tr. 9/10/01 at 105-06.) In order to determine whether Mr. Hagerty utilized the proper process or technique under the second factor, we must consider the NFPA 921, which provides the basic methodology for fire investigators and specifically lists steps as to how the scientific method should be applied in fire investigations. (Tr. 9/10/01 at 93.) Section 2-3 of the NFPA 921, entitled "Relating Fire Investigation to the Scientific Method," states, "The scientific method is a principle of inquiry that forms a basis for legitimate scientific and engineering processes, including fire incident investigations." (Hagerty, Tr. 9/10/01 at 94.) The NFPA 921 describes six steps as to how the scientific method should be applied in fire investigations. (Hagerty, Tr. 9/10/01 at 94.) The third step of the scientific method as applied to fire investigations relates to data collection and emphasizes that a complete and thorough fire investigation involves the collection of all relevant data that exists, where possible. (Hagerty, Tr. 9/10/01 at 95.)

In terms of the analysis of the data, Section 2.3.6.7 in the 2001 edition of the NFPA 921, entitled Presumption of Cause, states that "[u]ntil data have been collected, no specific hypothesis can be reasonably formed or treated." (Hagerty, Tr. 9/10/01 at 97.) Section 2.3.6.7 then states that "[a]ll fires should be approached by the investigator without presumption." (Tr. 9/10/01 at 95-96.) In addition, Section 2.4.3 of the 2001 revised edition of the NFPA entitled "Conducting the Investigation" provides that "[t]he investigator should conduct an examination of the scene if it is available and check data necessary to the analysis." (Hagerty, Tr. 9/10/01 at 97.) Mr. Hagerty testified that the collection of the data would include reviewing or interviewing witness statements and statements made by people who know something about the fire. (Hagerty, Tr. 9/10/01 at 98.)

Although the 1998 version of the NFPA 921 states that "[t]he thorough fire investigation always involves the examination of the fire scene either by visiting the actual scene or by evaluating the prior documentation of that scene," Mr. Hagerty testified that this section was modified in the 2001 version because "there are extenuating circumstances which would prohibit visiting the scene. . . ." (Tr. 9/10/01 at 104.)

Mr. Hagerty stated that he conducted a thorough investigation of the six vehicles which he actually inspected, but did not conduct an equally "thorough" investigation for the six vehicles that he did not inspect, according to the concepts as outlined in NFPA 921, including Baker, Shemnitz, and Tal, the plaintiffs at issue. (Tr. 9/10/01 at 107.) In his expert report of March 16, 2001, Mr. Hagerty states that he personally inspected vehicles or relevant components of plaintiffs Snodgrass, Sherron, Fletcher, and Ellis, and concludes that the fires in those vehicles were caused by the Fox ignition switch. (Hagerty Rpt. at 1-2.) As for the plaintiffs whose vehicles were no longer available for inspection, that is, Baker, Shemnitz, and Tal, Mr. Hagerty concludes, based on the Fifth Amended complaint accounts, relevant deposition transcript excerpts, and evidentiary documents for each of these plaintiffs, that each of those fired "occurred under conditions and with burn pattern results that are highly consistent with ignition switch causation." (Hagerty Rpt. at 2.) He offers the opinion that "[m]ost were likely caused by the ignition switch." (Id.) Mr. Hagerty stated that he bases his Fox ignition switch causation probability conclusion on the following:

I reach this conclusion based on the fire circumstances and burn patterns described by witnesses, even where the burned vehicles are later no longer available for inspection by a forensic fire expert. I base this Fox ignition switch causation probability conclusion on (1) Database statistics provided me by counsel suggesting that these Fox Fords experienced such fires some 3.5 times more frequently than did non-Fox Fords and (2) my knowledge of the gap width characteristics of the pre-redesigned Fox switches making them more susceptible to an electrical arcing (and fire causing) short circuit than other differently designed ignition switches."

(Hagerty Rpt. at 3.) Although Mr. Hagerty testified at the Rule 104 hearing that he did not rely on the proposed database in determining his conclusions regarding causation as to the vehicle fires of plaintiffs Tal, Shemnitz, and Baker, he subsequently stated that the database statistics only augment the probability that the vehicle fires were caused by defective ignition switches. (Hagerty Suppl. Statement, Pl.'s Post-Oral Arg. Br. Ex. B.) In addition, Mr. Hagerty testified that he relied on his experience, internal Ford documents, and Dr. Maladdi's report, which discussed many engineering studies on the cause of ignition switch fires. (Tr. 9/10/01 at 32.)

Mr. Hagerty's expert report also provides a list of "telltale signs" that "reinforc[e] the possibility that the ignition switch was likely the cause." (Hagerty Rpt. at 4.) The "telltale signs" include: (1) the ignition switch had been off and the engine not running for some time before the fire ignited; (2) when the fire started, the horn sounded, the lights flashed, the motor tried to start itself even though the ignition was not on, or smoke emanated from the steering column; (3) prior to the fire occurring, the key when removed felt warm to the touch; the ignition key was hard to turn; the radio, heater, or defroster did not function; or there was a crackling sound when the ignition switch was turned on. (Hagerty Rpt. at 4.) Mr. Hagerty stated that plaintiffs Shemnitz's and Tal's vehicle fires exhibited characteristics of the last category, and that plaintiff Baker's vehicle fire exhibited characteristics of the first category. (Id.)

Defendants, at the Rule 104 hearing, argued that plaintiffs' expert Mr. Hagerty did not comport with the guidelines of the NFPA in forming his opinions regarding the cause of plaintiffs' vehicle fires because he failed to investigate the vehicle or fire scene. Here, Mr. Hagerty reviewed all of the documentation available to him in cases in which the vehicle was unavailable or had been disposed of after the fire. In many cases, Mr. Hagerty reviewed the complaint accounts, deposition transcripts, and evidentiary documents, most of which described the fire incidents as occurring under conditions and with burn patterns resembling those fires attributable to ignition switch defects, according to his report. Indeed, Mr. Hagerty's vast experience as lead investigator in over 900 vehicle fires provides him with the capability and expertise to diagnose many different kinds of vehicle fires. See Hagerty Rpt. Ex. 1 at 3 (listing 17 different causes of vehicle fires, the investigations of which Hagerty had led). Mr. Hagerty's methodology, that of reviewing all documentation and available evidence when the car and fire scene no longer exist, substantially complies with the guidelines set forth in the NFPA 921, particularly the six steps of the scientific method as applied to fire investigations and Section 2.4.3 relating to "Conducting the Investigation," both of which emphasized examination of fire scenes, if available. Because, in this case, they were not, Mr. Hagerty's efforts to analyze the fire data were adequate.

Defendants question plaintiffs' reliance on documents and statements of plaintiffs as contrary to Chapter 11 of the NFPA, citing plaintiffs' reports as untrustworthy because of their interest in the case and therefore unreliable. Section 11.4.2 of the NFPA 921 indicates that "[i]nterviews can generally be categorized into three different types," and that "[t]hese include interviews with those you can approach with caution and interviews with those you should approach with an attitude of distrust." (Tr. 9/10/01 at 99.)

NFPA Section 11.4.4, entitled "Interviews With Those You Can Approach With an Attitude of Trust," indicates that this type of interview is conducted with those who can be considered reliable, identifying such people as government officials, citizen witnesses, and others who have no specific interest in the result of the investigation. (Tr. 9/10/01 at 99-100.) Section 11.4.5 describes interviews with those an investigator should approach with caution, described as having a degree of interest in the outcome of the investigation. (Tr. 9/10/01 at 100.) Section 11.4.6, entitled "Interviews With Those You Should Approach With an Attitude of Distrust," involves "persons whose information should be considered as unreliable unless substantially verified. Such persons include those who have an obvious or documented specific interest in the results of the investigation such as the suspect in an incendiary fire investigation." (Tr. 9/10/01 at 100-01.) Mr. Hagerty, however, testified that "[t]his is specifically addressing people who would be suspects in an arson fire burning their own property down," and that this is indicated by the "specific example that they use in the manual." (Tr. 9/10/01. at 101-02.)

The Court agrees with Mr. Hagerty that this provision deals with situations such as the specific example of a suspected arsonist. Although plaintiffs may have a specific interest in the results of the investigation, the Court believes that their evidence and testimony are not so distrustful as to be unreliable. The plaintiffs' reports provide important details and firsthand accounts of the fire scene and of the wreckage of the vehicle, which cannot be found elsewhere. Also, plaintiffs have answered interrogatories and given deposition testimony under oath, which enhances reliability. Despite the plaintiffs' self-interest, NFPA Section 11.4.5 would place them into the category of persons whom an investigator "should approach with caution," and not "with distrust."

Mr. Hagerty's use of documents provided by plaintiffs Baker, Shemnitz, and Tal does not render them unreliable when their accounts provide perhaps the most probative descriptions. On the contrary, this Court finds that Section 11.4.6 is meant to apply to suspected arsonists, which has not been alleged to be the situation here. Mr. Hagerty's vast experience in the field demonstrates to the Court that he is capable of properly weighing the evidence in forming his conclusions. Dr. Hagerty's methodology in forming his opinions as to plaintiffs Baker, Shemnitz, and Tal is sufficient to satisfy Rule 702's second requirement.

To the extent, however, that Mr. Hagerty's conclusions rely on Dr. Moshman's database statistics, the opinions of which were found above to be inadmissible, such conclusions are not reliable and will not be admissible. Thus, references in Mr. Hagerty's testimony to the probabilities of fires in various types of vehicles equipped with the Fox ignition switch, derived from Dr. Moshman's findings, will be stricken for lack of foundation.

Under the third requirement, Mr. Hagerty's opinion must assist the trier of fact. Admissibility depends on the connection between the scientific research or result and the particular issues in the case. See Paoli, 35 F.3d at 743 (citing Downing, 753 F.2d at 1237). Here, Mr. Hagerty's opinions relate to the causation of the vehicle fires of plaintiffs Baker, Shemnitz, and Tal. Mr. Hagerty will testify that the vehicle fires were more than likely caused by an ignition switch defect. Because Mr. Hagerty's opinions directly relate to and provide a scientifically-based explanations for a particular issue to be decided at trial, the third requirement under Rule 702 is satisfied.

Mr. Hagerty has satisfied the requirements of Rule 702, governing scientific expert testimony with respect to the cause and origin of the Baker, Shemnitz and Tal fires. Accordingly, Mr. Hagerty's opinions on causation as to plaintiffs Baker, Shemnitz, and Tal, to the extent that they do not rely on the probability statistics based on the database as furnished by Dr. Moshman, will be admissible at trial.

IV. Plaintiffs' Motion to Admit the Database and Motion for Admission of Narrowed Database for Purposes of Notice

The Court presently considers plaintiffs' motion to admit the database and motion for admission of the narrowed database. Plaintiffs move to admit the database of ignition switch fire complaints for purposes of showing that Ford was on notice of the defect. In addition, plaintiffs contend that the "database contains people who like plaintiffs had ignition switch fires in circumstances substantially similar to plaintiffs as verified through questionnaires, video tapes, State Farm Insurance expert reports, and Ford ASES documents." (Pl.'s Br. at 15.) Thus, plaintiffs intend to introduce the database for purposes of putting Ford on notice of the defect and proving causation. The Court herein incorporates its discussion in Part III.B.1.a, supra, relating to the unreliability of the database and will hereby proceed with its analysis.

Defendants argue that the database is inherently unreliable, that it is irrelevant under Rule 402, that the database is inadmissible hearsay under Rule 801, and that the database is unduly prejudicial under Rule 403.

A. Relevance of the Database for Purposes of Proving Notice

The court has determined, for reasons discussed in Part III.B.1,above, and Part IV.C. below, that the proposed database and narrowed database are inadmissible to prove causation, i.e., that the fires at issue in the trial were caused by ignition switch defects. That analysis will not be repeated here. The court will focus instead upon the admissibility for the purpose of proving Ford had notice of complaints of ignition switch fires at relevant points in time.

The database, if sought to prove solely the nature of complaints received by Ford and the dates of such complaints, is highly relevant to demonstrating what defendant Ford knew about allegations of ignition switch fires and when Ford received this information. If, for example, plaintiffs can demonstrate that Ford had early knowledge of the complaints of ignition switch fires and ignored them, or misrepresented that there was no such problem, then this evidence would be probative of issues of Ford's alleged knowledge and misrepresentations as claimed by various plaintiffs.

When the database is viewed as a compilation of complaints received by Ford in its various corporate capacities, gleaned from indecisive and time-consuming searches of Ford archives during the discovery phase of this case, then it becomes a summary of documents containing complaints and received by Ford at relevant points in time.

Seen as a compilation of documents from Ford's files, and probative only of the issue of notice, it is clear that the compilation is not admitted into evidence for the truth of the matters contained in the complaints and incident reports, but only for the fact that such reports were received. The issue of whether Ford was aware of such complaints is dispositively different from the issue of whether these events were actually ignition switch fires. Undoubtedly many of them were merely alleged or suspected, while others were confirmed. Whether or not the information on these complaints is true and confirmed is not the point for which this evidence is offered; it is instead the fact that such complaints and reports were known to Ford at a particular time.

The concerns of this court which fatally impaired the Moshman database (because it contained unreliable opinions that each of the events was an actual ignition switch fire) are cured if the database is reduced only to those complaints that were reported to Ford and obtained from Fords files, and if no "diagnosis" or "opinion" is presented as to the actual cause. If the compilation summarizes the date and nature of each such complaint which alleged to implicate an ignition switch fire, then such a compilation can serve to inform the jury, in an economical way, of the chronology and volume of such allegations, from which, together with other evidence, plaintiffs can seek to prove the extent of Ford's knowledge of the problem at various relevant points in time.²³

23 Because the parties have not focused upon which dates are relevant for the upcoming trial of this case, the court will not address specific time intervals now. Obviously, if the relevant dates are the dates of sale of plaintiffs' vehicles, an early cutoff coterminous with the last date of sale is appropriate and the task of compiling the Redacted Database is reduced.

B. Standard for Unfair Preludice

As discussed above, Rule 403 provides that "[a]lthough relevant, evidence may be excluded if its probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury, or by considerations of undue delay, waste of time, or needless presentation of cumulative evidence." Fed.R.Evid. 403. Thus, evidence may be excluded when its admission would lead to litigation of collateral issues, thereby creating a side issue which might distract the jury from the main issues. Blancha v. Raymark Indus., 972 F.2d 507, 516 (3d Cir. 1992) (citing United States v. Dennis, 625 F.2d 782, 797 (8th Cir. 1980)). Evidence also may be excluded if it relates to previous litigation involving the parties, has an aura of scientific infallibility but is likely to be used for purposes other than that for which it was introduced, or relates to statistical probabilities. See Blancha, 972 F.2d at 516 (citing J.B. Weinstein M.A. Berger, Weinstein's Evidence, ¶ 403 (1988)).

In determining the probative value of evidence under Rule 403, the court "must consider not only the extent to which it tends to demonstrate the proposition which it has been admitted to prove, but also the extent to which that proposition was directly at issue in the case." United States v. Herman, 589 F.2d 1191, 1198 (3d Cir. 1978), cert. denied, 441 U.S. 913 (1979).

C. Analysis

For its first contention, plaintiffs cite the proposition that "[g]enerally, where negligence is an issue, prior complaints to a defendant concerning an allegedly hazardous condition are admitted as being probative of the defendant's knowledge." See Pl.'s Br. at 15 (citing Julander v. Ford Motor Co., 488 F.2d 839, 846 (10th Cir. 1973)). Here, plaintiffs' database is compiled to show individual records of complaints of vehicle fires allegedly caused by ignition switch defects. The database lists the name, year, model, and source of complaint, such as Ford, MORS reports, ASES reports, State Farm, or the "Flaming Fords" internet website. See discussion supra Part III.B.1.a. In addition, the fields entitled "Miles," "Complaint date/Fire date," and "Diagnosis" (presumably determined by Mr. Hagerty), are empty for a substantial number of entries. See Moshman Rpt. Ex. G.

The Court is persuaded that a suitable compilation is probative of the fact that complaints put Ford on notice that there may be an ignition switch defect. On the other hand, the database entry for "Diagnosis" is generally the Hagerty opinion upon review of available information for the particular complaint that it was probably an ignition switch fire. Ford, of course, did not receive Hagerty's opinion with any of these complaints, since his opinions were only recently proferred with respect to this litigation. References in the database to Hagerty's opinions must be stricken.

Likewise, reports of incidents that were not reported to Ford, and which plaintiffs may have developed independently, must be stricken. Such complaints, unreported to Ford, are irrelevant to the issue of notice.

Also, where relevant to notice, the date when Ford received notice (actual or approximate) must be known. Alternately, if it can be determined that Ford received a complaint "not later than" a certain date, for instance because that date is referred to in other documents, then it can be included with the "not later than" date. If the date of Ford's notice is unknown, then the entry is irrelevant to the issue of when Ford learned of the allegation. Entries for which the date of Ford's receipt are unknown must be stricken.

Ford argues that such a database is more prejudicial than probative because its voluminous nature creates the inference that Ford was at fault in all of these database entries, numbering over 6, 000, even though no one has shown that this is so, at a minimum, after applying reasonable investigation techniques. The Court has remedied these legitimate concerns by striking the "opinion" aspects and narrowing the realm of permissible entries by date, as discussed above. The jury will be told that this redacted database constitutes solely a chronology of the complaints received by a Ford entity alleging a problem with an ignition switch fire. If demonstrating that complaints put Ford on notice is the aim of the plaintiffs' database, the Court envisions that a suitably redacted database will provide a time line showing when Ford received such complaints and how many complaints were in fact received, which will be sufficient in this regard, and which does not turn each such report into a causation minitrial. The database is not presently structured, however, to demonstrate that proposition nor are the dates actually known for a substantial number of records. As a matter of relevance under Rule 402, and substantial undue prejudice to Ford under Rule 403, the proposed database must be significantly restructured to be admissible for this purpose.

In addition, plaintiffs contend that the "database contains people who like plaintiffs had ignition switch fires in circumstances substantially similar to plaintiffs." See Pl.'s Br. at 5. Here, plaintiffs seek to admit evidence that purports to be a database of ignition switch fires. However, as discussed above, see Part II.B.1.a, the entirety of fires has not been reviewed by plaintiffs' expert Mr. Hagerty and adequate demonstration for a reliable determination of the causes of many database fires does not exist because he has been unable to apply the recognized methods of a cause and origin expert to the existing data. Thus, the database still contains fires that have not been determined to a reasonable probability, to be attributable to a defective ignition switch. Out of 7, 500 entries, Mr. Hagerty had reviewed as of the date of the Rule 104 hearing, between 2, 200 and 2, 500, leaving 5, 000 at best as unreviewed. As discussed above, the court is further unwilling to permit the jury to infer that the initial inclusion of each record in the database is proof that the fires were actually caused by an ignition switch defect.

Plaintiffs' Motion for Admission of Narrowed Database, filed January 8, 2002, provides little proof that the Court's concerns, regarding the reliability of plaintiffs' database, addressed in Part III.B.1.a, have been remedied satisfactorily. Plaintiffs assert that 303 net complaints have been deleted (321 deletions less 18 additions), and that the proposed database is now reduced to 6,227 complaints, due to the further setting aside of 1,227 entries, the additional documents for which had not yet been obtained. Plaintiffs state that "Hagerty has personally reviewed the well over 2,000 Questionnaires received and has spent much time reviewing individual MORS IT and other source records, resulting in hundreds of deletions." See Pl.'s Mot. Narrowed Database Br. at 5. A scan of the proposed database, see Pl.'s Mot. Narrowed Database. Br. Ex. D1, however, indicates that a substantial number of records, the number still unclear from plaintiffs' submissions, have not been reviewed by Mr. Hagerty. The database, therefore, lacks uniformity and standardized purification that would render it sufficiently reliable for purposes of providing the foundation to Dr. Moshman's expert opinions and for its own admission into evidence.²⁴

24 In the context of these motions, the Court is not ruling that the list of approximately 162 ignition switch fire complaints reported to Ford prior to 1990, as referred to in plaintiffs' motion for reconsideration, is inadmissible. This list and plaintiffs' motion for reconsideration of the Court's Order and Opinion filed September 4, 2001, are currently being considered by the Court.

In summary, the Court finds that the presently constituted database and the narrowed database identified in plaintiffs' motion of January 8, 2002, are inadmissible in their present form. Such a database, if redacted to delete events which were not reported to a Ford entity at a known point in time, and to delete references to opinions reached by Mr. Hagerty, would be probative of Ford's knowledge of complaints of ignition switch fires at relevant points in time, and not unfairly prejudicial to Ford. With an appropriate limiting instruction to the effect that this redacted database may be considered for determining issues of notice, but not of causation of any disputed fire event, the jury would properly be focused. The resulting redacted database will be admissible at trial for this limited purpose.²⁵

25 Since the data in the underlying redacted database, as permitted by this Court's opinion today, derives solely from documents received by a Ford entity at a relevant date in time, it is assumed that plaintiffs' characterization of such documents will be accurate. Plaintiffs are cautioned, in preparing the redacted database, to preserve and cross-reference the underlying documents from which the descriptions are taken, to avoid any dispute about accuracy.

The plaintiffs motion to admit the present database is denied, without prejudice to the admissibility of a database redacted in accordance with today's opinion.

Plaintiffs shall proffer the redacted database within thirty (30) days of today's date.

CONCLUSION

For the reasons set forth above, defendants' motion for severance and dismissal of plaintiffs' claims for lack of jurisdiction will be denied; defendants' motion in limine to preclude evidence of Recall 95S28 will be denied; defendants' motion in limine to exclude the expert testimony of Dr. Jack Moshman will be granted; defendants' motion in limine to exclude the expert testimony of Mr. William Hagerty will be denied; and plaintiffs' motion to admit the database and motion for admission of the narrowed database will be denied without prejudice to plaintiffs' submission of a Redacted Database within thirty (30) days of today's date which, if so constructed, will be admissible upon the issue of notice of alleged ignition switch fires, but not upon the issue of causation of those fires or of the disputed fires in this case. The accompanying Order is entered.

Trucks

Ford Service Recall Bulletin, Def.'s Br. Ex. B.