2001-2007 Peter Nordberg. E-mail: webmaster@daubertontheweb.com. Last revised: 9/17/06. See the reviews. Buy the T-shirt. Read the disclaimers. View the FAQ.


    "Discretion without a criterion for its exercise
    is authorization of arbitrariness.

    -- Frankfurter, J., in Brown v. Allen, 344 U.S. 443 (1953)

Standard of Review

    General Electric Co. v. Joiner, 522 U.S. 136 (1997), is usually cited for its holding that Daubert rulings are to be reviewed on appeal under a comparatively lenient "abuse of discretion" standard.  That nostrum, however, begs the really interesting questions.  There are different kinds of evidentiary ruling.  How should "abuse of discretion" review be conducted in these different contexts?

The Traditional Rationales for Deferential Review

    Deference to evidentiary decisions by trial courts is sometimes justified because the court must make factual determinations in order to evaluate admissibility -- e.g., whether a document is authentic, whether an expert has the necessary education or experience, etc.  The trial court's prerogative to make such determinations is sanctioned by Fed. R. Evid. 104(a), and there are plausible reasons why the appellate courts should hesitate to second-guess them (including all the usual reasons for deference to finders of fact).     

    In addition, one main traditional rationale for deferential appellate review has been that trial judges are better situated than distant appellate tribunals to strike the pragmatic balances commonly integral to evidentiary rulings.  Such rulings often involve weighing the probative value of concededly relevant testimony against (1) the cost, in time and other resources, of  hearing it, and/or (2) the danger that the testimony will unduly prejudice the jury.  Decisions like these centrally involve the exercise of practical judgment, which is characteristically exercised by splitting differences -- e.g., by allowing some of the evidence but not all of it, by controlling the time and manner of its presentation, and/or by attaching conditions (e.g., limiting instructions) to its admissibility.  Pragmatic compromise is not a universal feature of such decisions, but it is certainly a hallmark.

One Reason Why Daubert Decisions Are Different

    Daubert rulings, by contrast, often involve no element of compromise.  Either the expert's ultimate opinion is admitted, or it is excluded -- and when it is excluded, the consequences are frequently dispositive.  Often, the excluded testimony is necessary to carry the plaintiffs' burden of proof on an indispensable element of the plaintiff's claim (e.g., causation).  In such instances, the power to exclude testimony amounts, in practical effect, to the discretionary power to dismiss the plaintiff's case.

    Now it is no secret that some trial judges have long aspired to just such power, nor that one principal mission of the appellate courts, historically, has been to hold such ambitions in check.  Formerly, this battle was frequently joined in the context of summary judgment.  Trial courts would sometimes yield to the temptation, when a defendant moved for summary judgment, to weigh the plaintiff's evidence and find it wanting.  The courts of appeal would then step in to remind everyone that weighing the evidence was a jury function, and that the trial judge's only proper role, on a motion for summary judgment, was to determine whether the plaintiff's evidence would suffice to uphold a verdict in the plaintiff's favor.  The question, that is, was supposed to be not whether the trial judge did credit the plaintiff's evidence, but rather whether a reasonable jury could -- a question the appellate courts have long evaluated for themselves, supposedly with no deference to the trial court's analysis.  To be sure, these rules of engagement left a certain amount of de facto room for interstitial fact-finding by trial judges -- but only sub silentio.  Overt usurpation of the jury's fact-finding role would lead to reversal, for reasons not unrelated to the Seventh Amendment's jury guarantee. 

    Daubert, however, opened up broad new domains of factual inquiry for district courts -- in the context, now, of "reliability" rulings attendant to admissibility determinations.  Before Daubert, an award of summary judgment predicated on a judicial determination that an expert's testimony was methodologically unsound would have been highly vulnerable to reversal on de novo review.  After Daubert, the same determination, rendered in the course of granting a motion in limine and thereby precipitating a failure of proof in a concurrently pending motion for summary judgment, would be reviewed only for abuse of discretion.  In short, district courts could now attempt to immunize their own summary judgment rulings from close appellate scrutiny, in cases heavily reliant on expert testimony, by the mere act of relabeling their analysis.  

    One plausible response to this problem would have been to adopt a stricter standard of appellate review (perhaps a particularly stringent variety of abuse-of-discretion review) where the practical effect of a Daubert ruling was to precipitate the dismissal of a claim.  That is the proposal rejected by the Court in Joiner.  The standard of review for Daubert decisions, the Joiner Court held, should not vary in its strictness depending on whether the evidentiary ruling is "outcome determinative."  Nor should it "categorically" distinguish between decisions allowing expert testimony and decisions disallowing it.

    The merits and demerits of this holding are subject to fair debate.  On the one hand, the ruling in Joiner did little to squelch mischief at the district court level.  On the other hand, it is desirable, no doubt, to promote the rhetoric of neutrality in matters of judicial review, and a standard that expressly favored proponents of expert testimony at the expense of objecting parties might be considered insufficiently even-handed.   In practice, of course, many appellate courts will continue to examine claim-defeating Daubert rulings with special attentiveness.  They'll just be careful not to say so aloud. 

     But whatever the pros and cons of Joiner's actual holding, it is worth noticing two things that the Joiner Court did not hold.  First, the Court did not conclude, as is too often implied, that appellate review of Daubert rulings should be acquiescent to the point of flaccidity.  (Indeed, the Court's own analysis of the expert testimony in Joiner, and of the underlying epidemiological and animal studies on which the Joiner experts relied, was relatively searching.)  Nor, secondly, did the Joiner Court hold that there is simply no difference in the implementation of abuse-of-discretion review for Daubert rulings and other evidentiary decisions.

    In fact, not only is there such a difference; there has to be.

Another Reason Why Daubert Decisions Are Different

     The Joiner Court rejected the argument that judicial review of exclusionary Daubert rulings should be stricter than usual because such rulings are often especially damaging to the propounding party.  But quite apart from their frequently dispositive character, there is a second distinguishing feature of Daubert rulings, not explicitly addressed in Joiner, that also bears on the modalities of appellate review.  A Daubert analysis commonly embroils the district court in an explicit analysis of whether the witness is reasoning soundly -- an assessment, that is, of what the Court's subsequent Kumho Tire opinion calls the expert's "intellectual rigor."

    Now admittedly, Daubert analyses are not the only evidentiary context in which this ever occurs.  Most notably, under Fed. R. Evid. 701(a), lay witnesses "not testifying as an expert" may offer only such opinions as are "rationally based on the perception of the witness."  Even when a lay witness testifies to what may seem a straightforward matter of fact -- e.g., "the bus was blue" -- counsel must usually lay some foundation to show that the witness was in a position to know.  If the witness is not color blind and actually saw the bus on a clear day from a distance of fifty feet, fine.  If the witness never saw the bus but believes it was blue based on something the witness heard on a psychic hotline, less fine.

    Nor, admittedly, do all rulings under Daubert focus on the legitimacy of the expert's reasoning.  This caveat is necessary not only because of the separate problem of legal relevance, but also because some of the factors prescribed in Daubert for evaluating the validity of expert testimony are not intrinsically linked to the logical or rational validity of the expert's inferences.  It is possible, for example, to imagine Daubert decisions excluding an expert's testimony solely because the expert's methods were not followed by most other members of the relevant discipline, and/or because the expert's work had not been published in a peer-reviewed journal.  All the same, such decisions begin to seem less plausible in circumstances where the expert's reasoning is concededly unassailable.   Factors like publication and peer acceptance (or the lack thereof) are generally adduced as additional grounds for admitting expert testimony that seems otherwise defensible, or as additional warrants for skepticism where some aspect of the expert's analysis already seems problematic.

    For present purposes, however, the point is not to establish that Daubert rulings are unique, but rather to draw attention to one of their broadly differentiating features: Daubert analyses are characteristically epistemological, in the sense that they gauge the presuppositions and foundations of putative scientific or technical knowledge and the extent of its validity.  This very commonly involves assessing the rationality of a sustained argument presented in the expert's report and testimony.  To be sure, there are supposed to be limits on how deeply this assessment may probe.  Something is supposed to be left for the jury.  But it seems fair to say that if the performance of a Daubert analysis did not entail at least some significant judicial inquiry into the legitimacy of an expert's reasoning, there would no necessity for validity analyses under Daubert at all.  The Frye test, e.g., would do.

Implications for Judicial Review: Five Theses

    Based on the discussion to this point, it seems appropriate to offer several propositions for consideration.

  1. To the extent that district court "validity" rulings under Daubert are predicated on factual determinations (e.g., that the expert did indeed review the data as claimed, that the expert's techniques do enjoy currency among other practitioners of the discipline, that the lab instruments actually were in good working order, that the expert's work has in fact been published), no compelling reason appears why the trial court's factual determinations should not receive the same deference customarily afforded under Fed. R. Evid. 104(a).

  2. But evaluation of the soundness of an expert's reasoning is not a factual inquiry in anything like the customary sense, and so many of the traditional grounds for deference to fact-finders (e.g., their primacy in evaluating credibility) are of questionable applicability in this context.

  3. Likewise, a finding that an expert's reasoning is valid or invalid does not depend on judgments about courtroom practicalities of the sort that trial judges are uniquely situated to make.

  4. District judges are not possessed of any other special qualifications or attributes, nor privy to any special information, that would render them better able than appellate judges to assess the soundness of expert reasoning (any further than they are in the context, e.g., of summary judgment).

  5. In any event, and whatever one may believe about propositions (1) through (4): Insofar as district court validity analyses under Daubert are predicated on evaluations of expert reasoning, there is virtually no plausible method by which meaningful appellate review of such decisions can generally proceed except via evaluation of the validity of the district court's own reasoning.  That is, the legitimacy of an evidentiary decision founded on a rational critique of expert testimony depends radically on whether the judicial critique can itself withstand (some level of) rational scrutiny.

    If proposition (5) is correct, then the "abuse of discretion" standard may leave less room for variability of outcome in appellate review of district court validity analyses than the nominally identical standard leaves in other situations.  As to preliminary questions of fact resolved in the course of evidentiary decisions, no one would normally dream of reversing a district court's determination except for procedural abuse or clear error.  Likewise, if a district court refuses to admit fifty gory color photographs but allows three black-and-white images, there is no well-defined pragmatic calculus that can be invoked to compel the conclusion that the trial judge has violated settled principles of sound practical judgment.  The canons of logical, mathematical, and scientific reason, although not without pliancy, are less forgiving.  Scientists may or may not have better criteria than the rest of us for truth, but they do often have clearer criteria, amenable to independent appellate investigation, for what constitutes a mistake.  Perhaps it is correspondingly easier for district judges to make what will count as one.

    Proposition (5), if correct, also counsels in favor of relatively full exposition in district court validity analyses.  If appellate review of validity decisions necessarily involves appraising the soundness of a district court's analysis, then the actual analysis must be stated in sufficient detail to permit such review.  This does not seem an unfair demand.  After Daubert, the experts themselves are required to state the bases for their views in frequently excruciating detail, to facilitate independent review of their methods and reasoning.  The same reasons warrant comparable transparency in district court opinions.  Moreover, providing a full and detailed explanation for evidentiary rulings imparts an especially salutary discipline in the Daubert context, where judges are passing on the methodological validity of complex technical work performed by scientists in whose fields the judges themselves have no special training.  Not only does such discipline help prevent unnecessary mistakes; it also helps keep all the participants intellectually honest. 


    A litigant who treats Joiner as announcing a safe haven for intellectually sloppy Daubert rulings does so at the litigant's peril.  In close cases, appellate review is frequently fairly searching, especially where important and creditable evidence has been excluded.  The best way to interpret Joiner's "abuse of discretion" standard may be to see the Court as having been primarily concerned to rule out the only clear viable alternative: de novo review.  Even that interpretation could be misleading, because appellate scrutiny has often been essentially indistinguishable from de novo review when matters of recurring significance are presented on a clean record (e.g., whether epidemiological support is needed for medical causation testimony).

    In any event, no litigant should accept, or be asked to accept, a Daubert ruling not supported by an opinion giving clear and intellectually legitimate reasons for the ruling.  And very arguably, appellate scrutiny of such opinions should be at least as demanding as the courts' own scrutiny of the underlying testimony.

Coda: Some Potential Glosses on "Abuse of Discretion"

    It is possible to imagine verbal formulations that would specify the meaning of "abuse of discretion" more precisely in the Daubert context.  A far from exhaustive list of candidates would include:

bulletrulings founded on demonstrably erroneous (or unsupported) factual or scientific premises;
bulletrulings weighing competing expert testimony to determine who has the better of the scientific argument on the merits, rather than focusing on the "validity" of the proponent's testimony;
bulletrulings evaluating testimony under a sufficiency standard rather than a validity standard;
bullet"validity" rulings evaluating expert testimony on credibility rather than methodological grounds;
bulletrulings excluding testimony based on methodological errors committed by the expert but not material to the expert's conclusions;
bulletrulings where a fair opportunity for proponent and objecting party to be heard has not been afforded;
bulletrulings evincing plain misunderstanding or bias.

Kumho Tire