Answers are grounded in this paper and cite their sources.
Polygraph validity is often summarized by classification accuracy, potentially obscuring whether a procedure produces a definitive decision and how errors differ for guilty and innocent examinees. This meta-analysis compared the Comparison Question Test (CQT) and Concealed Information Test (CIT) using harmonized methods across laboratory mock-crime studies. Classification analyses included 35 CQT effects comprising 2,127 participants and 22 CIT effects comprising 1,081 participants. For All-Tested Accuracy, which retained inconclusive outcomes in the denominator, pooled estimates were 71.6% for the CQT and 86.2% for the CIT, a significant difference. For Conclusive-Decision Accuracy, estimates were 82.7% and 86.6%, respectively, and did not differ significantly. Decision yield was 84.7% for the CQT and 99.8% for the CIT. Within-test analyses indicated higher guilty- than innocent-participant accuracy for the CQT and higher innocent- than guilty-participant accuracy for the CIT; cross-test status differences were not tested directly. Scoring method did not significantly moderate either outcome after adjustment for multiple testing. Nine AUC studies were analyzed separately; the pooled CIT AUC was .895, whereas only one CQT AUC study was available. These findings show that comparative evaluation of polygraph procedures should consider not only accuracy among conclusive classifications, but also decision yield and diagnostic profile.
No takes yet. Share an insight, caveat, or question.
Ruscio et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: