The Judicial Gatekeeping Project
Home | About the Gatekeeping Project | Judicial Gatekeeping Book | Educational Resources |
| Message Board | News & Developments | Subscribe to the Judicial Gatekeeping Newsletter |
| Site Map | Search | Contact Us | Credits | Back

& National Conferences

New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Rhode Island
South Carolina
South Dakota
West Virginia


Ruling on Reliability and Relevance in a Daubert Hearing:
The Methodology-Conclusion Debate and Other Issues

by Eric D. Brown, Blake Snider, and Victor Svilik - Harvard Law School '99

The Gatekeeping Role
     The Daubert opinion fundamentally altered the structure of judicial decision-making in regard to the admissibility of expert testimony. Under the Frye general acceptance test, which Daubert superseded (or, more precisely, according the ruling of Federal Rules of Evidence superseded), the judge did not, herself, assess the reliability of an expert's opinion. Instead she judged whether the expert was qualified and the opinion relevant an only touched upon the reliability through inquiring whether others in the expert's field considered the methodology used by the expert to be reliable. Daubert changed this by placing the responsibility to make a reliability determination squarely on the shoulders of the trial judge. In a post-Daubert setting, the Frye inquiry is merely one component to which a judge may look in making a decision whether or not an expert's methodology is reliable and his opinion ultimately admissible. This critical structural change is what places judges in a gatekeeping role.

     The most commonly heard objection from judges to this role is that it forces them to perform a function for which they are not well suited. After all, many contend, judges are not amateur scientists. While this is a valid objection to some extent, it may be overstating the case. In fact, Daubert does not ask judges to cull through scientific opinions and pronounce a select few to be "good science." Rather, the gatekeeping framework asks judges to filter out testimony that falls outside the bounds of acceptable reliability. Certainly Daubert requires more of a judge than the Frye test did, but the judicial role becomes less daunting when it is seen as requiring an exclusionary function, as filtering out "bad science" rather than anointing "good science." In this sense, the judge need only determine whether there are holes in the expert's reasoning, a task that can be carried out successfully by an informed non-scientist.

     Taking this as the task of a judge when confronted with expert testimony, how should she approach the two-pronged Daubert test of reliability and relevance? The first prong concerns the "scientific...knowledge" component of Federal Rule of Evidence (FRE) 702 and asks judges to examine the methodology employed by the expert witness.(1)The second prong, often referred to as "fit," requires that the expert's testimony "assist the trier of fact to understand the evidence or to determine a fact in issue."(2) Both reliability and relevance must be established for expert testimony to be admissible.

Assessing Reliability
     At the heart of a Daubert hearing is the question of the reliability of the expert's scientific testimony. To aid judges in making this determination, Daubert sets forth a non-exhaustive list of factors to be used by judges to determine the reliability of an expert's methodology. In addition to the familiar Frye acceptance analysis, the Court also notes that important inquiries include (1) whether the methodology has been tested, (2) whether it has been subjected to peer review and publication, and (3) the known or potential rate of error during the study.(3) Courts in various jurisdictions have added to this list or focused on one or more components of the analysis. In any case, noncompliance with one or more of these (or other court-established) factors does not necessarily invalidate the expert's methodology, so long as the court is satisfied that the overall processes constitute reliable grounds upon which the results can rest.

The Methodology-Conclusion Debate
     Ruling on the reliability of an expert's opinion presents difficulties when the so-called methodology-conclusion debate is taken into consideration. Justice Blackmun's opinion in Daubert asserts that when determining the scientific validity of proffered testimony, "the focus, of course, must be solely on principles and methodology, not on the conclusions that they generate."(4) On its face, this analysis seems to suggest that a trial judge should not base her decision on an evaluation of the conclusions proposed by an expert. This may seem clear cut, but in reality differences of opinion exist in where to draw the line between the methodology used by an expert and the conclusions reached. The methodology-conclusion debate takes on three different forms:

     (1) In one version, a judge struggles with where her gatekeeping inquiry should fall along the line of generality/particularity. At one extreme, the argument is that judges should inquire only at a general level about the reliability of the expert's methodology. Such an inquiry would be satisfied, for example, by evidence that the expert used methodology generally accepted in the field of epidemiology. But does this collapse into the defunct Frye test? Doesn't Daubert seem to require a deeper inquiry? At the opposite pole, a judge's inquiry would entail examination of every step in the expert's process of gathering data and reasoning to a conclusion. This very particular examination of an expert's methodology would look at every step leading up to, but not including, the conclusion. Such an examination would mirror the evaluative inquiry made by the most demanding peer review journals. Should judges and scientific journals utilize the same standards in reviewing studies?

     In Hall v. Baxter Healthcare Corporation, Judge Jones of Oregon acknowledged that Daubert limits the judge's evaluation to the methodologies used by an expert, yet he later stated that there appears to be no real demarcation between the scientific methodology used and the conclusions that subsequently arise.(5) Jones ultimately determined that all steps of an expert's reasoning process must be evaluated, which includes the formulation of conclusions.(6) Jones' decision appears concerned with situations in which the underlying methodology does not support the expert's offered conclusions. An example of a conclusion that does not follow from a given methodology arose in In re Paoli R.R. Yard PCB Litigation, in which the court excluded an expert's testimony because the conclusions proffered were beyond the scope of the methodology.(7) In this case, where the expert's methodology required her to perform physical examinations of the patients, she only examined two of them, yet tried to apply her conclusions to all of the plaintiffs.(8)

     (2) A second version of the controversy involves looking at the conclusions reached not for their merit but as a means of judging how particular the examination of an expert's methodology should be. If an expert offers a conclusion that is at odds with the overwhelming majority of similar studies in a field, can the judge consider this fact to be a reason to take an especially careful look at the expert's methodology? In this respect, a judge can only look at an expert's conclusion in order to focus the inquiry into methodology, but not to throw the testimony out solely because of the outlying conclusion.

     (3) The third version of the debate involves actually looking at the conclusion as a part of the methodology. The argument is that conclusion is inseparable from methodology in that an expert's choice of methodology and application of it determines the expert's conclusion. According to this view, an expert who offers a conclusion that is at odds with the weight of opinion in his field must explain why this is so as part of the testimony offered. If this explanation is missing, the testimony can be deemed unreliable. As such, an un-reconciled outlying conclusion could automatically invalidate expert testimony. The Ninth Circuit in Lust v. Merrell Dow Pharmaceuticals, Inc. indicated that, at the very least, the proponent of expert testimony bears the burden of proving admissibility, and, if few others share an expert's conclusions the court can use that fact as a basis for excluding the testimony.(9) Is this crossing Justice Blackmun's boundary between methodology and conclusion?

Reliability Determination by the Judge or the Jury?
     If a judge in a Daubert hearing makes reliability determinations, a jury will never see the evidence offered. Should a judge withhold testimony from the jury because an expert's conclusion is inexplicably at odds with other studies in the same field or if the expert's methodology is "questionable"? Or should the jury be allowed to consider this testimony given complete information, for example with a caveat that an expert reached a conclusion different from the thirty other studies in the expert's field? One argument is that a conclusion that seems unpersuasive or novel does not necessarily indicate that the expert's methodology was flawed in some respect; such an evaluation, in any case, should be a question for the jury, not the judge.(10) The language in Daubert notes that a methodology accepted by even a minority of the scientific community may be admissible, as long as it has some demonstrable basis of reliability. In the recent Oregon case Hall v. Baxter Healthcare Corporation (involving claims against breast implant manufacturers for injuries allegedly caused by silicone gel breast implants), Judge Jones recognized that, in making an assessment of reliability, the fact that an expert's conclusions differ from those of other experts is not a sufficient basis alone for determining that the testimony is inadmissible.(11)

     Those who wish to curtail the gatekeeping role of judges often speak in terms of protecting the province of the jury. If, they argue, a judge admits testimony over the objections of the other side, opposing counsel still has adequate safeguards, such as cross-examination and contrary evidence, to protect against conclusions perceived to be in error.(12) Those who construe Justice Blackmun's delineation between methodology and conclusion as inviolable note that by limiting review to methodology, the Daubert inquiry eliminates the temptation for judges to take an expert's credibility into account, leaving this determination properly in the hands of the jury.(13) According to this interpretation, the focus on methodology acts as a check on the judge's gatekeeping power, ensuring the accepted division of authority and protecting the plaintiff's right to his day in court.

Assessing Relevance
     Daubert's second prong, relevancy, has received considerably less attention from commentators than the question of reliability. Nonetheless, this dearth of scholarly treatment does not indicate that the second prong is of any lesser importance, for the evaluation of fit, like the reliability inquiry, requires a high degree of judicial scrutiny. A determination of fit requires that a judge decide whether or not proffered testimony will assist the trier of fact in drawing a conclusion. This necessitates a determination as to whether or not testimony produced is relevant to the facts at issue in a case. Justice Blackmun's opinion in Daubert provides a useful example of how the question of relevance comes into play. According to Justice Blackmun's hypothetical, while expert testimony about the phases of the moon may be applicable in determining how dark it was on a given night, the very same testimony should be inadmissible if offered to demonstrate the psychological state of a defendant.(14) Thus, an expert's opinion can be methodologically flawless yet it can still be dismissed for failure to reach a reasonable conclusion. A judge, who chooses to review how an expert reached his conclusion via his methodology, as informed by the methodology-conclusion debate discussed above, could well throw out on the grounds of reliability, testimony about the moon aimed at proving psychological state. This ruling could be based on a belief that the expert improperly reasoned to his conclusion, and underlying this would be an assumption that this reasoning is part of the expert's methodology. But what if the expert doesn't offer a conclusion but merely lays out scientific evidence about the phases of the moon? While the expert himself draws no conclusion about the effect of the phases of the moon on the psychological state of the defendant, the lawyer for the defense will attempt to use the testimony to this effect. A judge in her gatekeeping role can throw out this testimony for a failure to "fit" the causation requirements of the case at hand.

Beyond Rule 702
     Unfortunately, questions of relevance rarely present themselves in a form as simple as Justice Blackmun's example. Courts must struggled to delineate how closely proffered scientific testimony must be related to the facts at bar. Although Daubert focuses on Rule 702 as the "primary locus" of a court's gatekeeping role, judges are not precluded from looking to other rules for guidance.(15) Some judges have relied on Rule 403 in determining what evidence will "assist the trier of fact." Rule 403 establishes a balancing test weighing the probative value of the evidence against "the danger of unfair prejudice, confusion of the issues, or misleading the jury...." Going back to the phases of the moon example, such testimony could be excluded on Rule 403 grounds that it has a high potential to mislead the jury even if the expert did not offer a specific conclusion.

The Judge-Jury Question
     Implicit in an admissibility decision is the potential danger that by excluding the evidence, the judge may intrude into the province of the jury. In exercising his gatekeeping role, the judge has predetermined that this evidence cannot--under any circumstances--support a verdict for the plaintiff. Is such a consideration proper, or should the judge allow the evidence to reach the jury, trusting that the opposing party's ability to impeach the testimony through cross-examination will preclude an unjustified verdict? In deciding what testimony could "assist the trier of fact," judges may choose, however, to look beyond the simple question of whether or not the testimony fits the facts. If a scientific study which indicates that the causal agent at issue keeps the relative risk less than two (double the risk), a judge may choose to exclude the study on grounds that such evidence would meet a "more likely than not" standard of proof.(16) A problem with exclusion on this basis is that while such a study alone may not meet the requisite level of legal proof, it may be valuable in conjunction with other evidence in the aggregate to meet that standard. In excluding testimony about tests that fall below a certain level of probability, a judge may risk denying a jury the opportunity to consider important parts of a causation puzzle.

     Judges acting as evidentiary gatekeepers are bound by the two-pronged test set out in Daubert. In examining expert testimony for reliability, judges first face the hurdle of distinguishing between methodology and conclusion and choosing how far to extend their reliability inquiry. Beyond this, even if testimony passes the reliability test, a judge may exclude it for failing to meet the relevancy standard. Analysis of fit gives judges a degree of leeway in assessing the validity of applying an expert's conclusion to the facts at hand. The challenge, then, of conducting a Daubert hearing lies in balancing an understanding of the Daubert holding, a court's interest in judicial efficiency, the parties' rights to due process, the jury's role as fact-finder, and the judge's role as evidentiary gatekeeper.


1. Daubert v. Merrell Dow Pharmaceuticals, Inc. 113 S.Ct. 2786, 2795 (1993).
2. Fed. R. Evid. 702.
3. 113 S.Ct. at 2797.
4.113 S.Ct. 2786, 2797.
5. 947 F.Supp. 1387, 1387 (1996).
6. Id.
7. In re Paoli R.R. Yard PCB Litigation, 35 F.3d. 717 (3rd Cir. 1994).
8. Id. at 766.
9. Lust v. Merrell Dow Pharmaceuticals, Inc., 89 F.3d 594, 598 (9th Cir. 1996).
10. Kenneth J. Chesebro, Taking Daubert's Focus Seriously: The Methodology/Conclusion Distinction, 15 Cardozo L. Rev. 1745, 1749 (1994).
11. 947 F.Supp. at 1387 (1996).
12. 113 S.Ct. at 2798.
13. Bert Black, Science and the Law in the Wake of Daubert: A New Search for Scientific Knowledge, 72 Tex. L. Rev. 715, 749-750 (1994).
14. 113 S.Ct. at 2796.
15. 113 S.Ct. at 2795.
16. David E. Bernstein, The Admissibility of Scientific Evidence After Daubert v. Merrell Dow Pharmaceuticals, Inc., 15 Card. L. Rev. 2139, 2170 (1994).

Page Last Modified on April 25, 1999 by Dan Fridman - Copyright 1999
All materials are the property of the Berkman Center for Internet & Society and Harvard Law School. 
Materials may be reproduced, distributed, or quoted as long as appropriate credit and citation is given.