← All Posts

Juvenile Deception Interrogation Bans and Case Clearances: Preliminary Interactive Results

juvenile justice interrogation deception bans NIBRS difference-in-differences

Note: These are preliminary results from an ongoing study. They are shared for grant sponsor review and are subject to revision. Please do not cite or distribute.

Overview

Between 2022 and 2024, seven U.S. states---Illinois, Oregon, Utah, Delaware, Nevada, Indiana, and Colorado---enacted legislation restricting or prohibiting law enforcement from using deceptive interrogation tactics when questioning juvenile suspects. A central policy concern is whether these restrictions reduce case clearance rates.

Using National Incident-Based Reporting System (NIBRS) data covering 510,582 juvenile incidents from 2021 through 2024, we employ a staggered difference-in-differences design with 44 never-treated control states to estimate the causal effects of these bans on arrest clearance rates, exceptional clearance rates, and overall case clearance rates.

Key Finding

The pooled national effect is null. Randomization inference yields p = 0.558 for arrest clearance and p = 0.906 for any clearance. However, the null average masks substantial heterogeneity: Oregon’s any-clearance rate increased by 4.4 percentage points (p < 0.001) while Colorado’s decreased by 2.1 pp (p = 0.003). This heterogeneity is systematically related to legislative design---Oregon has the strongest evidentiary standard (clear and convincing), while Colorado has the broadest definition with additional compliance mandates.

Interactive Dashboard

The full results---including state-by-state DiD estimates, distributed lag dynamics, offense-specific effects, exceptional clearance subcategory analysis, sex crime deep dives, Bayesian multilevel models, agency-level heterogeneity, and robustness checks---are available in an interactive dashboard.

Open Interactive Results Dashboard

The dashboard allows you to explore all results interactively---filter tables, hover over plots for detailed estimates, and compare across states and outcome measures.

Methods Summary

  • Estimators: Two-way fixed effects DiD (state-by-state), Sun & Abraham (2021) interaction-weighted estimator, Callaway & Sant’Anna (2021) group-time ATTs, randomization inference (500 permutations), and Bayesian multilevel binomial models (brms/cmdstanr)
  • Outcomes: Arrest clearance rate, exceptional clearance rate, any clearance rate (constructed from NIBRS administrative and arrestee segments)
  • Robustness: Placebo-in-time tests, pre-trends validation, offense-specific estimation, agency-level ATT distributions, incident volume composition checks