B – 133 Domain-specific or Non-Discriminant? Divergent Validity of the MMPI-2-RF Over-Reporting Scales.

Abstract Objective Over-reporting theories emphasize deception and intent dimensions, as well as distinct somatic, cognitive, and psychological domains. Sharf and colleagues (2017) meta-analyzed the Minnesota Multiphasic Personality Inventory-2-Reformed Form (MMPI-2-RF) over-reporting scales and concluded the scales assessed their intended symptom domains (e.g., Fs assesses somatic symptoms). However, statistical analyses demonstrating discriminant performance were not conducted and effect sizes were generally equitable. Relatedly, emerging work suggests over-reporting scales share most of their variance. Method We (1) re-analyzed Sharf et al’s (2017) findings, computing Cohen’s q effects to compare differences between the effect sizes of each scale and (2) conducted a meta-analysis of over-reporting scale inter-correlation to assess for divergent validity. Results The meta-analytic re-analysis found minimal to small differences across the over-reporting scales (q = .03−.27), suggesting a non-discriminant performance between scales of different symptom domains. The correlation meta-analysis (k=14, n=7,984) includes data from the MMPI-2-RF Technical Manual (k=7) and published studies (k=7). We observed largely similar consistent intercorrelations (r =.43−.84; M=.63, SD=.10). Consistent with the meta-analytic re-analysis, these effects were mostly (60%) negligible (q =.03−.25). Assessment setting moderated some correlations whereas sex did not. Conclusion In both meta-analytic analyses, over-reporting appears a more general task, rather than one in which specific differences emerge reliably or consistently across symptom-focused domain constructs. Results are consistent with recent scale development efforts (e.g., Ingram et al., 2024). In light of recent concerns about symptom validity (Leonhard, 2023), we will discuss the implications of these findings for validity assessment theory and practice guidelines (Rogers & Bender, 2018; Sweet et al., 2021).