<p>The Personality Assessment Inventory (PAI) is a widely used broadband personality assessment embedded with validity scales capturing over-reported pathology. This meta-analysis examines the utility of the PAI over-reporting scales as measured by mean differences on standard (NIM, MAL, RDF) and supplemental (NDS, MFI, HMI, CBS, CB-SOS) scales. These comparisons are made across simulation and criterion studies to compare scale efficacy and effectiveness, respectively. 6,451 participants across 43 studies were analyzed using a series of random and fixed effect meta-analyses. We calculated general (e.g., detection effectiveness) and specific (e.g., simulation vs criterion) effect sizes as well as summarized classification statistics and other contextual information (e.g., criterion groups) observed across the literature. Results demonstrate moderate to large effect sizes across most standard (<i>g</i> = .99-1.50) and supplementary scales (<i>g</i> = .84-1.81), consistent with expected ranges. Compared to criterion studies (<i>g</i> = .17-.92), simulation designs (<i>g</i> = 1.16-2.27) were more effective (<i>g</i>_{<i>differences</i>} = .57 – 1.60). Published studies also produced lower effects than unpublished (<i>g</i>_{<i>differences</i>} = -.21—-.59). Our findings generally support the efficacy of the PAI’s over-reporting scales, as well as their effectiveness in non-simulation (e.g., criterion-based) designs. However, RDF does not effectively measure over-reporting in criterion groups and should not be used for those decisions at present. Implications and future directions for the PAI and over-reporting are discussed.</p>