Measuring Compliance, Claiming Confession: A Meta-Analysis of Laboratory False Confession Paradigms

Abstract

In the most widely used laboratory false confession paradigm, 36% of participants “confess” even when no interrogation tactic is applied—the experimenter simply states an accusation. Taken at face value, this baseline rate alone would imply over a quarter-million false confessions annually in the United States, exceeding the cumulative historical count of known false confession exonerations by orders of magnitude. We reanalyze 99 experimental conditions drawn from 41 studies using hierarchical Bayesian estimation, random effects meta-analysis, publication-bias diagnostics, calibration simulations, and dose-response analysis. The pooled false confession rate is 43%, with a 95% prediction interval spanning 4% to 93% (I2 = 93%)—a level of instability difficult to reconcile with measurement of a stable psychological construct. Calibration simulations show that the observed laboratory means are statistically compatible only with real-world prevalence rates that no credible estimate supports. Interrogation tactics account for just 4% of the variation in confession rates; paradigm structure does the rest, with the no-tactic baseline alone explaining 60% of the false evidence ploy rate. Drawing on Meehl’s (1967) critique that paradigms structured to produce easy confirmations tell us little about the phenomena they attempt to model, we argue that the observational hurdle in this literature is set too low. These findings indicate that dominant laboratory paradigms primarily capture low-stakes compliance rather than interrogation-induced false confession, warranting substantial caution when extrapolating experimental results to legal, policy, or forensic contexts.