Discretion may be the better part of valor. In dispensing justice, though, it can be the wild card. And that’s not a good thing.
The range of judicial discretion is almost always defined as broad, its exercise reversible only when abused. On occasion, an appellate court might reverse a discretionary decision, but usually the loser has no recourse. And the nature of discretion means that, most of the time, the judge could have decided for the other party, with equal freedom from reversal.
If the issue is material to the outcome, then the outcome doesn’t depend on the facts or the law; rather, it depends on the luck of the draw, on which judge got the case.
To boot, the border line between a permissible and impermissible exercise of discretion is fuzzy. Cases use vague words—plain error, clear error, arbitrary and capricious, irrational, not justified by the evidence. That leaves even appellate judges disagreeing over whether a trial court abused its discretion.
In Koon v. United States, 518 U.S. 81 (1996), five justices found that the trial court abused its discretion in departing downward from sentencing guidelines; four justices disagreed. If reasonable justices can differ, doesn’t that suggest that the ruling below was within the judge’s broad discretion?
Perhaps. But maybe they disagreed because, with such an elusive standard, appellate reviewers just follow their gut reactions, either rubber-stamping discretionary decisions or second-guessing them.
The categories of matters left to judicial discretion are as varied as the discretion is broad. They include license over what discovery is allowed; when to rule on motions; how much time each side has to put on its case; what evidence is relevant; whether to limit or exclude impeaching, corroborating, or expert evidence; what the jury is told about the law and how to decide the case; how much to award in alimony; whether to order a remittitur, additur, or new trial; how long to sentence a defendant. The list goes on.
Those three features of judicial discretion—its breadth, its fuzzy boundaries, and its ubiquity—lead not only to predictability problems but also to an uneasy feeling that we aren’t governed by laws after all. That’s a far cry from the justice system Chief Justice John Marshall envisioned in Osborn v. Bank of the United States, 22 U.S. (9 Wheat.) 738, 866 (1824):
Judicial power, as contradistinguished from the power of the laws, has no existence. Courts are the mere instruments of the law, and can will nothing. When they are said to exercise a discretion, it is a mere legal discretion, a discretion to be exercised in discerning the course prescribed by law; and, when that is discerned, it is the duty of the court to follow it. Judicial power is never exercised for the purpose of giving effect to the will of the judge, always for the purpose of giving effect to the will of the legislature; or, in other words, to the will of the law.
Today’s discretion is different. Consider the Daubert trilogy: In Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993), the Supreme Court held that scientific expert testimony needn’t rest on a technique that the relevant scientific community generally accepts as reliable. The Court said that such a requirement would contradict the “liberal thrust” of the evidence rules, which “relax[ed] the traditional barriers to opinion testimony.” While the Court noted that judges should exercise “some degree of regulation of the subjects and theories about which an expert may testify,” it said nothing about judicial discretion. Instead, it said that cross-examination, contrary evidence, and jury instructions are the “appropriate means of attacking shaky but admissible evidence.”
Discretion crept into that jurisprudence in General Electric Co. v. Joiner, 522 U.S. 136 (1997). There, the Court held that a judge’s order either excluding or allowing scientific expert testimony is reviewed for abuse of discretion. Rather than explain why, the Court simply held it to be so, citing two cases in which the Court, analyzing other rules of evidence not involving expert testimony, had reviewed certain evidentiary rulings under an abuse of discretion standard.
Then, in Kumho Tire Co. v. Carmichael, 526 U.S. 137 (1999), the Court held that Daubert’s preliminary screening requirement—gatekeeping—and General Electric’s holding that those decisions are discretionary and reviewable only for abuse apply to all expert testimony, not just scientific testimony. And it said that trial judges have “broad latitude” in how to test an opinion’s reliability, but as in General Electric, the Court didn’t say why.
So how does that play out in practice? Consider expert testimony that is experience-based. Rule 702’s committee notes state that experience alone may provide a sufficient basis for expert testimony and, in “certain fields, experience is the predominant, if not sole, basis for a great deal of reliable expert testimony.” But problems arise when judges assess experience-based opinions against scientific standards, often leading to the exclusion of essential experts.
Look at United States v. Frazier, 387 F.3d 1244 (11th Cir. 2004). There, the trial court barred a well-credentialed defense expert, who had worked on 150–250 sexual assault cases, from testifying that when rapes occur in the way the alleged victim described, typically some hair or seminal fluid from the defendant is found on the victim, on the victim’s clothing, or in the car. (None was found in Frazier.) The testimony was essential to the defense because the case otherwise consisted of only an accusation and a denial.
The judge excluded the opinion because the expert had no scientific data on how often, in those situations, those hairs or fluid is found. Then the judge let two government witnesses testify that the lack of a defendant’s hair or seminal fluid didn’t mean that sexual contact had not occurred.
The jury convicted. A life sentence followed.
On appeal, a 2–1 majority vacated the conviction, finding that the exclusion was an abuse of discretion. On en banc review, a different majority found no abuse of discretion and, over a stinging dissent, reinstated the outcome. The discretion pendulum swung back and forth. But how can we know whether the trial judge and en banc majority got it right or whether an innocent defendant is spending the rest of his life in prison because they got it wrong?
Compare that with United States v. Brown, 7 F.3d 648 (7th Cir. 1993), in which a government expert was allowed to offer an experienced-based opinion that, in light of paraphernalia and behaviors typically associated with distributors and not with mere users, a defendant intended to distribute cocaine, rather than personally use it.
Why should a government expert be treated more indulgently than a qualified defense expert? Why are empirical data excused in one case but not the other? Is there a de facto double standard—as many scholars have concluded—that favors opinions from government experts over those from the defense?
These concerns aren’t confined to criminal cases. Last year, PwC released a 19-year study on how often courts excluded testimony of financial experts, mainly in civil cases. The study looked at over 2,800 challenges to financial experts and found that the rate of either partial or entire exclusions was about 45 percent, with wide variations among the circuits. In the First and Third Circuits, exclusion rates were 35 percent and 37 percent; in the Second and Tenth Circuits, 50 percent and 58 percent.
Yes, it might have been that the expert opinion in each case was so lacking in intellectual rigor that it would have been a miscarriage of justice to let a jury hear it. But the same study found that, on appeal, 37 percent of the decisions that entirely excluded an expert’s testimony and 27 percent of the partial exclusions were reversed. Given that the standard for reversal is abuse of discretion, those rates suggest that trial judges are abusing their discretion too often.
What the data don’t capture is the percentage of exclusions affirmed on appeal that would also have been affirmed had the judge’s decision gone the other way. Because judges have wide discretion, at least some, and maybe most, of the excluded testimony fell into a middle ground, where the judge could have ruled either way without abusing discretion.
If justice is the goal, then those middle ground cases show why wide discretion undermines achieving it: No matter how the judge rules, there’s no wrong answer.
Given the importance of expert testimony, Daubert motions give judges enormous power to shape a case’s outcome. A decision to exclude won’t likely be reversed, though it hobbles the party that proffered the opinion. That’s a bitter pill.
Some of the justices in Daubert, General Electric, and Kumho Tire had reservations about how much gatekeeping responsibility a judge should have. In his separate partial dissent in Daubert, Chief Justice Rehnquist commented that scientific knowledge, method, and validity are “matters far afield from the expertise of judges.” He cautioned that judges’ gatekeeping responsibility doesn’t impose on them “either the obligation or the authority to become amateur scientists in order to perform that role.”
In a partial dissent in General Electric, Justice Stevens explained how he envisioned judges should exercise their Daubert discretion: “An example of ‘junk science’ that should be excluded under Daubert as too unreliable would be the testimony of a phrenologist who would purport to prove a defendant’s future dangerousness based on the contours of the defendant’s skull.” He apparently saw Daubert as keeping out only opinions that are far-fetched.
But in practice, Daubert has let in farfetched opinions. Consider Stevenson v. Texas, 304 S.W.3d 603 (Tex. App. 2010), in which a video analyst was allowed to testify that a shadowy image in a video of a robbery was money and another was a gun. His methodology? Watching the video repeatedly until he figured it out. And Daubert has excluded opinions that are near-fetched, as in Frazier, where the sexual assaults expert was barred from offering an opinion that logically should have produced reasonable doubt.
On a systemic level, broad discretion is the opposite of equal treatment under the law. It fosters subjective decision-making, ungoverned by objective criteria consistently applied from judge to judge. The irony is that everything rides on one arbitrary predicate event: the assignment of the case to a judge who later rules one way, instead of to a judge who would have ruled the other.
If we could rewrite the standard, what’s better? Narrow discretion? Defined discretion? Articulable standards for its exercise? Ground rules? A set of principles and guardrails? What if, instead of wide discretion, the Court had said this:
Many litigable issues cannot be resolved by lay testimony alone. Experts are needed.
Educating a jury in an area of relevant specialized knowledge is a welcome advancement in dispute resolution. But juries shouldn’t be confused by phony opinions masquerading as expert ones. The resulting prejudice when a jury hears bogus opinions requires that judges perform some gatekeeping role.
Therefore, on a party’s request, a judge should preview and evaluate proposed expert testimony. If that evaluation shows that no reasonable jury could credit the opinion, then the judge must exclude it and explain that ruling so that it can be meaningfully reviewed on appeal.
The presumption, though, should favor admissibility. If there is a genuine dispute about the validity of the opinion and reasonable minds could differ, then the opinion should be admitted. The jury will then decide if the opinion is to be believed and what weight to give to it, just like any other litigable issue.
In allowing judges to exclude invalid opinions, we decline to commit such decisions to a judge’s broad discretion. Broad discretion would mean that one judge might exclude an opinion, while another might allow it; except in the clearest of cases, neither judge would be making a reversible error. That isn’t logical, evenhanded, or fair to the litigants. Admissibility shouldn’t turn on the random assignment of the case to one judge instead of another.
Our ruling today strikes the right balance. If the jury decides against the opinion’s proponent, no error will have been made in allowing the jury to hear it. If the jury decides in favor of the proponent, but the judge concludes that the opinion was material to the verdict, that no reasonable jury could have credited it, and that the opinion should have been excluded, then the judge has remedial tools to correct the improvident verdict, including directing a verdict, ordering a new trial, or ordering additurs or remittiturs.
A ruling like that would have established a measurable and reviewable standard, promoted consistency and reliability, given litigants confidence that a judge won’t be arbitrary or pick sides, and subjected judges who do it differently to accountability and reversal on appeal.
To be sure, different issues might call for judges to have different ranges of discretion. A decision to exclude expert testimony might warrant well-defined standards and little elbow room; a decision to grant a continuance, take witnesses out of order, or suspend a sentence might entail giving judges more discretion.
Discretion, after all, isn’t inherently bad. Sometimes it’s essential. But giving judges broad discretion in every matter that requires some exercise of judgment risks transferring too much power from the law to the person who happens to wear the robe and be assigned to that case.
That’s exactly what Chief Justice John Marshall decried. At worst, it’s a license to put a judicial thumb on what should be an evenly balanced scale.
