A Cross-Cultural Evaluation of the MMPI-2-RF in Sexual Violence Risk Assessment

Jamie L. Anderson, Sam Houston State University
Robbi Brockhaus, Institut fur Psychologische Diagnostik und Begutachtung
Julia Kloefer, Correctional Facility (JVA) Ludwigshafen and Private Practice
Martin Sellbom, University of Otago

The prediction of sexual violence recidivism is an important societal concern and has been the frequent subject of psychological research. Although research has supported the use of the MMPI-2-RF in violence risk assessment, there is a paucity of research related to sexual violence risk assessment. The present study evaluated the convergent validity of the MMPI-2-RF with measures used in sexual violence risk assessment evaluations (i.e., the Static 99, SVR-20, and PCL:SV) in a sample of German inmates in a psychological treatment setting. Analyses generally showed an expected pattern of results, with MMPI-2-RF externalizing scales having the strongest associations with measures of violence and sexual violence risk assessment. Indeed, although smaller associations were found with Static-99 scales, there were moderate to large associations between SVR-20 and PCL:SV scores and MMPI-2-RF externalizing scales. Taken together, these findings suggest that, although the MMPI-2-RF does not appear to be associated with specific static risk factors (e.g., previous victim qualities, type of sexual offenses), the externalizing scales on the MMPI-2-RF are useful in examining more dynamic factors (e.g., those assessed on the SVR-20) in conjunction with other dynamic and static risk tools (e.g., SVR-20, Static-99) in sexual violence risk evaluations.

MMPI-R-RF, sexual violence, risk assessment, psychopathy, SVR-20

“Risk assessment is an important clinical task in forensic psychological settings, including the assessment of sexual violence recidivism…Actuarial risk assessment instruments (ARAIs) have dominated the sexual violence risk assessment literature…with static factors such as previous sexual and/or violent offending…representing robust predictors of future sexual violence recidivism…However, dynamic risk factors, such as current mental health symptoms and personality factors, are also important to assess…such as social support, substance use, occupational instability, and emotional problems…” (p.403).

“The Minnesota Multiphasic Personality Inventory (MMPI; Hathaway & McKinley, 1942) and its subsequent iterations, the MMPI-2 (Butcher et al., 1989/2001) and the MMPI-2-Restructured Form (MMPI-2-RF; Ben-Porath & Tellegen, 2008/2011) have a long history of use and acceptance in forensic settings, including in risk assessment contexts…Importantly, although no MMPI scale explicitly measures sexually deviant behavior, the MMPI has long been used to assess risk factors related to both physical and sexual aggression…Despite evidence suggesting that the MMPI-2-RF may be useful in assessing both risk and psychopathic traits, minimal research has examined the use of this instrument specifically in sexual violence risk assessment…” (p.404).

“The current study aimed to examine the convergence between relevant MMPI-2-RF scales and measures frequently used in sexual violence risk assessment evaluations (i.e., the Static 99; SVR-20; and PCL:SV). Importantly, the current study used a sample of German inmates in a psychological treatment setting, providing one of the few cross-cultural risk assessment studies using the MMPI-2-RF…we aimed to take a targeted approach by selecting specific scales that are most likely to be relevant in the assessment of risk…the majority of our hypotheses include associations between specific risk factors and externalizing scales…” (p.404-405).

“…Generally speaking, the MMPI-2-RF showed utility in assessing many risk factors, particularly those associated with dynamic personality factors and externalizing behaviors. More specifically, the current results added to the already growing support for the utility of the MMPI-2-RF in assessing various aspects of psychopathy…as expected, the majority of meaningful associations were with externalizing scales…although several interpersonal scales on the MMPI-2-RF showed negative associations with PART 1 as anticipated, we hypothesized that several internalizing scales would also show associations with Part 1 [of the PCL:SV] and its facets. These hypotheses were not supported and AXY [anxiety scale] showed an unexpectedly positive correlation with Part 1…” (p. 409-410).

“Unsurprisingly, there were minimal associations between the MMPI-2-RF scales and the static risk factors associated on the Static-99-R. Although the total score was associated with broad externalizing psychopathology…there were few associations with specific Static-99-R items. The exception was Prior Sentencing Dates, suggesting that externalizing psychopathology scales on the MMPI-2-RF capture a history of criminal offending…the MMPI-2-RF’s coverage of areas such as impulsivity, behavior problems, substance abuse, and social support appeared to overlap with several areas covered on the SVR-20, particularly Psychosocial Adjustment factors…Taken together, the current study suggests that there is substantial overlap in many of the constructs being evaluated in sexual violence risk assessment. Not surprisingly, the externalizing and interpersonal scales on the MMPI-2-RF may be particularly useful in these evaluation contexts. Of note, however, its utility appears relatively limited to the assessment of dynamic, rather than static, risk factors. Indeed, although externalizing scales were associated with a history of previous violence, the majority of static factors showed minimal associations with any MMPI-2-RF scales…” (p.411).

“…we do not assert that the MMPI-2-RF should serve as a replacement for measures commonly administered in sexual offender risk assessment contexts. Importantly, the MMPI-2-RF did not appear to provide adequate coverage of all relevant risk factors assessed in a sexual violence risk evaluation. However, based on the current findings, we propose that the MMPI-2-RF would make a useful addition in these evaluations. Given the high-stakes involved in such clinical decisions, it seems prudent to collect information from a variety of sources, including self-reported symptoms of current psychological functioning. As noted, some scales on the MMPI-2-RF provide coverage of constructs that are not otherwise adequately assessed elsewhere…the MMPI-2-RF in conjunction with standard dynamic and static risk tools is likely to be useful in sexual violence risk evaluations” (p.411-412).

“There were several surprising findings that should be explored. For instance, associations between the SVR-20 Major Mental Illness item and scales on the MMPI-2-RF were not consistent with hypotheses. This may have occurred due to the historical nature of the SVR-20 and the present context of the MMPI-2-RF…Therefore, it is possible this suggests a limitation in the assessment of major psychiatric illness using standard risk assessment instruments…there were surprising findings between the SUI scale and several factors on the SVR-20…the Suicidal/Homicidal Ideation scale on the SVR-20 showed stronger associations with aggressive behavior (i.e., AGG) rather than over[t] suicidal content (i.e., SUI)…we hypothesize that higher scores on the SVR-20 may suggest an increased vulnerability in this population. Similarly, low level of hopelessness (HLP) was associated with supervision failure, suggesting that individuals who fail in community supervision may report greater perceived wellness and ability to succeed without supervision…” (p.411).

As always, please join the discussion below if you have thoughts or comments to add!