Do Risk Measure Scores and Diagnoses Predict Evaluator Opinions in Sexually Violent Predator Cases? It Depends on the Evaluator

Objective: Field research increasingly reveals that forensic evaluators are not interchangeable. Instead, they tend to differ in their patterns of forensic opinions, in ways that likely reflect something about themselves, not just the persons evaluated. This study used data from sexually violent predator (SVP) evaluations to examine whether evaluator differences in making intermediate decisions (e.g., instrument scoring, assigning diagnoses) might explain their different patterns of final opinions. Hypotheses: Although this study was generally exploratory and not strongly hypothesis driven, we expected that there might be evidence for a simple form of bias in which some evaluators would be more likely than others to consistently "find" indications of SVP status (i.e., consistently assigning higher risk scores and more SVP-relevant diagnoses) and, therefore, be more likely to find behavioral abnormality, the legal construct qualifying someone for commitment as an SVP. Method: The study used data from 745 SVP evaluations conducted by 10 different evaluators who were assigned cases from the same referral stream. Potential evaluator difference variables included behavioral abnormality opinions, paraphilia and antisocial personality disorder diagnoses, and Psychopathy Checklist-Revised and Static-99 scores. Results: Evaluator differences explained a statistically significant (p < .001) amount of variance in behavioral abnormality opinions (17%), paraphilia diagnoses (7%), and Psychopathy Checklist-Revised scores (16%). Contrary to our expectation of a simple tendency for some evaluators to find all indicators of SVP status more often than others, evaluators differed in the ways that underlying diagnoses and scores corresponded with their conclusions. The overall pattern was one in which different evaluators appeared to base their final opinions on different factors. Conclusions: Findings reveal further evidence of substantial forensic evaluator differences in patterns of assigning instrument scores and reaching forensic conclusions. But these findings are the first to also reveal wide variability in their patterns of reaching forensic conclusions.