Consistency of feigning measure scores over time: A study of SIMS scores across five administrations


Feigning of mental health and neurocognitive symptoms is a serious concern, particularly in forensic settings. While researchers have developed several assessment tools to detect feigning, little research exists regarding the pattern of scores on these measures over time. This dissertation involved secondary data analyses using the publicly available dataset from the study entitled, “Evaluation of the Psychological Effects of Administrative Segregation in Colorado, 2007-2010” (O’Keefe & Klebe, 2014). Participants were 270 adult male inmates who were administered the Structured Inventory of Malingered Symptomatology (SIMS; Widows & Smith, 2005) along with measures of psychopathology at five time points (baseline and 3, 6, 9, and 12 month follow-ups). This dissertation sought to investigate patterns of scores on the SIMS across the five administrations in order to examine the consistency of inmates’ reporting over time. Additional research questions included comparing two different recommended cut-scores for the SIMS among inmates with and without mental illness and investigating other factors that may influence evaluees’ SIMS scores over time. At all time points, inmates with mental illness received significantly higher mean SIMS Total Scores compared to inmates without mental illness. Indeed, 75.5% of inmates with mental illness scored in the feigning range at least once, compared to 35.7% for inmates without mental illness. Test-retest reliability for the SIMS Total Score at all combination of time points revealed moderate to high reliability (r ranging from .61 to .83). McNemar’s tests indicated similar proportions of participants scored in the feigning range at all combinations of time points, with the exception of the Time 1 and Time 5 comparison. Repeated measures MANOVAs were used to compare subgroups of inmates (i.e., Never, Sometimes, or Always Feigning) on measures of psychopathology (i.e., the PAS and BPRS). Results revealed inmates in the Always Feigning group tended to receive higher mean scores on the PAS and BPRS compared to inmates in the Sometimes Feigning Group, and both groups scored higher than inmates in the Never Feigning group. Scores on these other measures also tended to decrease from the baseline administration to later time points. Regression models revealed that changes in scores on the PAS and BPRS predicted changes in scores on the SIMS; 25.4% of the variance in the model was explained by these two predictors, with the PAS having the stronger influence. Neither housing placement nor mental health needs significantly predicted changes in SIMS scores; however, a moderate-strength correlation suggested regression to the mean explained some of the change in SIMS scores over time.



Feigning, Malingering, SIMS, Structured Inventory of Malingered Symptomatology, Consistency, Cut scores