Empirically derived effect sizes benchmarks in criminology and criminal justice: Distributional evidence and publication bias

Authors

• Nan Li, The University of Texas Rio Grande ValleyFollow

Document Type

Article

Publication Date

6-2026

Abstract

The current study had two objectives: (1) to propose empirically derived benchmarks for three commonly used effect size metrics—correlation coefficient, standardized mean difference, and odds ratio—within criminology and criminal justice; and (2) to assess the extent to which publication bias, including the file drawer problem and p-hacking, might distort the construction and interpretation of these benchmarks. We extracted 6054 primary effect sizes (2343 correlation coefficients, 1978 standardized mean differences, and 1733 odds ratios) from 94 meta-analyses published in 11 criminology and criminal justice journals. Tertiles were used to classify effect magnitudes as small (< 33rd percentile), medium (33rd–67th percentiles), and large (> 67th percentile). Robust methods were conducted to examine the associations between publication status and effect size, as well as between sample size and effect size. The following benchmarks were proposed: for correlations, values less than 0.130 are small, 0.130 to 0.290 are medium, and 0.290 or greater are large; for standardized mean differences, values less than 0.210 are small, 0.210 to 0.550 are medium, and 0.550 or greater are large; for odds ratios, values less than 1.300 are small, 1.300 to 2.164 are medium, and 2.164 or greater are large. However, evidence of publication bias suggests that these benchmarks may be inflated relative to the true underlying distribution of effects. Empirically derived benchmarks grounded in the observed distribution of effect sizes within criminology and criminal justice offer a discipline-specific reference point that can serve as an informed starting point for interpretation.

Comments

Original published version available at https://doi.org/10.1016/j.jcrimjus.2026.102664

Recommended Citation

Li, N. (2026). Empirically derived effect sizes benchmarks in criminology and criminal justice: Distributional evidence and publication bias. Journal of Criminal Justice, 104, 102664. https://doi.org/10.1016/j.jcrimjus.2026.102664

Creative Commons License

This work is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 4.0 International License.

Publication Title

Journal of Criminal Justice

DOI

10.1016/j.jcrimjus.2026.102664