Some trial errors are obviously harmful. Others are obviously harmless. The hardest cases fall in between.
Arizona v. Fulminante is one of those cases. The U.S. Supreme Court held that the admission of Fulminante’s coerced confession was subject to harmless error analysis, but that the error was not harmless in his case. The decision became a landmark because it confirmed that even coerced confessions can be reviewed for harmlessness, while also recognizing that a confession may have a powerful effect on the jury.
My research uses Fulminante as a case study to test whether the method developed in Measuring Fairness can correctly analyze a difficult trial-error problem. The goal was not simply to show that coerced confessions matter. The harder question was whether the method could identify how much this confession likely affected the probability of conviction, and whether that effect crossed the legal threshold for harmful error.
Measuring the Effect of the Coerced Confession
The study compared two versions of the case. In the actual-trial condition, respondents reviewed the case with the challenged confession included. In the hypothetical fair-trial condition, they reviewed the case without that confession. The analysis then used the difference in individual verdict preferences to estimate the difference in the probability that a jury would convict.
This is the basic logic of fair trial analysis. The actual trial is compared to a legally proper hypothetical trial. The difference between the two estimated outcome probabilities measures the effect of the error.
The Fulminante study helps validate the method because the Supreme Court decision provides a benchmark. The Court concluded that the coerced confession was not harmless. The empirical analysis points in the same direction, but it also shows why the case is close. The effect is meaningful, but not so overwhelming that the answer is obvious without careful analysis.
That makes Fulminante especially useful. A method that only works in easy cases is not very helpful. Courts need better tools for hard cases—where the evidence is mixed, the error is serious, and the legal conclusion depends on how the comparison is framed.
Why Precision and Context Matter
The Fulminante case study shows why trial-error analysis must be precise. It matters who has the burden of proof. On direct appeal, the prosecution must usually show that the error was harmless beyond a reasonable doubt. In post-conviction proceedings, the defendant often bears the burden of showing prejudice. The same estimated effect can have different legal consequences depending on which party must prove its position and how much certainty is required.
It also matters where the case was tried. A jury pool in one jurisdiction may respond differently from a jury pool in another. Local attitudes, demographics, and background assumptions can affect how jurors evaluate confessions, circumstantial evidence, and credibility. Estimating the effect of an error requires attention to the relevant community, not an abstract national audience.
State jury rules matter as well. The relationship between individual juror preferences and final verdict probabilities depends on the size of the jury, unanimity requirements, and other rules governing jury decision-making. A shift in individual preferences may have one effect under a unanimous twelve-person jury rule and a different effect under another system. That is why Measuring Fairness does not stop with individual survey responses. It translates those responses into estimated verdict probabilities under the governing jury rules.
The broader lesson is straightforward: precision and detail matter. It is not enough to say that a confession is powerful, or that the remaining evidence was strong, or that the case was close. A careful framework must specify the actual trial condition, the hypothetical fair-trial condition, the relevant jury pool, the applicable jury rules, the burden of proof, and the legal threshold for harm.
That is the kind of work Fair Trial Analysis was built to do. The organization’s empirical trial-error analysis reflects core values of rigor, transparency, and practical usefulness. The goal is not to label every error harmful or harmless in the abstract. The goal is to measure whether a specific error, in a specific case, under the applicable legal rules, likely affected the outcome.
Fulminante shows why that matters. In close cases, justice may depend on getting the framework right.