Interactive Dashboards
100% of Shark Attack Victims Were Wet
A jokes-first tour of why data selected on the outcome can describe the cases but can't tell you how risky anything is. Two silly stories (a wedding buffet and shark attacks), one switch that reveals the people who were never in the file, and one slider that moves the true risk from "safe" to "yikes" while the case-file number refuses to budge. Ends with why the same logic applies to exoneration data.
Does a Six-Hour Interrogation Carry a 10% Risk?
Smith et al. (2026) report that the risk of a false confession wrongful conviction reaches 10% at about six hours of interrogation, and 25% at twelve. Both numbers depend on choices the evidence cannot settle. Move the duration slider and watch the twelve-hour estimate swing from 26.3% to 72.6% on the choice of false confession study alone; a second figure puts the six-hour threshold anywhere from 3.9 to 14.8 hours, or nowhere at all. Both are computed live from the published duration model. Based on Mourtgos & Adams (2026), a reply to Smith et al. (2026), Journal of Criminal Justice.
What Do Laboratory False Confession Paradigms Measure?
A guided walkthrough of a calibration meta-analysis of 41 laboratory false confession studies (99 conditions, ~3,700 participants). The pooled false confession rate is 41% — and 37% of participants confess in the control condition. Interactive calibration tools show why those numbers are incompatible with any plausible real-world rate, and what the paradigms are more likely measuring: low-stakes compliance. Based on Mourtgos & Adams (2026), Journal of Quantitative Criminology.
Seeing Confessions Clearly: Diagnosticity, Risk, and Signal Detection
For two decades the interrogation literature has graded techniques with a single ratio — true confessions divided by false confessions. Move two sliders and watch that ratio swing while the method’s actual ability to tell guilty from innocent stays frozen. Reanalyzes Russano et al. (2005), where the ranking by that ratio matches the ranking by confession threshold exactly, and the ranking by discriminability hardly at all. Based on the CrimRxiv preprint by Mourtgos, Blair & Adams.
Criminogenic Entropy: Tracking the Structural Drivers of Crime
An interactive look at how 15 structural conditions — jobs, housing, mobility, trust, policing, and more — jointly track the long-run rise and fall of U.S. crime (1970–2023). Includes a what-if simulator that lets you adjust each condition and see how entropy, and therefore crime, responds, plus scenario projections to 2033. Based on the CrimRxiv preprint by Mourtgos & Adams.
Did Banning Deceptive Interrogation Tactics Affect Juvenile Case Outcomes?
A step-by-step walkthrough of what happened to juvenile case outcomes after nine states banned deceptive interrogation of minors. Across 510,582 juvenile-involved incidents (NIBRS 2021–2024), Bayesian difference-in-differences and state-specific synthetic controls show arrest clearances did not decline after adoption — estimates run null to modestly positive across twelve estimation strategies — and the drop in prosecution-declined clearances is not juvenile-specific. Based on Mourtgos & Adams (2026), Crime & Delinquency.
The Risk of False Confession Wrongful Convictions
An interactive Bayesian framework for estimating how often lawful interrogation tactics contribute to wrongful convictions of the innocent. Based on Mourtgos & Adams (2026), Journal of Criminal Justice.