An examination of naive and role informed faking on Rorschach performance

This study was undertaken to determine if non-psychotic adults could simulate a psychotic presentation, malinger, on the Rorschach Inkblot Test, as evaluated using Exner's Comprehensive System. All Ss were volunteers from an undergraduate psychology class and were screened for the absence of psychosis with the Minnesota Multiphasic Personality Inventory. Ten males (ages 18-35) and 20 females (ages 18-35) were randomly assigned to three equal groups. Males and females were distributed on an equal basis to each group.

Each group was administered the Rorschach by examiners who were naive regarding the experimental conditions. The control group was tested according to standard procedure. A second group of naive fakers (NF) was told to imagine that they were being tested following a judicial request and that they wanted to appear to be "crazy" to avoid incarceration. A final group of role informed fakers (IF) was given specific instructions regarding how to malinger prior to testing. Following test administration both treatment groups were debriefed regarding their strategies for malingering.

Three Multivariate Analyses of Variance (MANOVAs) were employed to evaluate the 36 dependent variables under study; two MANOVAs had 13 variables and the other had 10. Individual Analyses of Variance (ANOVAs) were employed to follow up the two MANOVAS which had significant F-ratios (p > .05). Eleven ANOVAs displayed significant treatment effects (p > .05). Next, Tukey "B" Comparisons (α < .05) were employed to compare group means. The IF and NF groups differed from the controls on nine and four variables. Two and four variables were shown to be significantly different across and between treatment groups, respectively.

Although multiple treatment effects were identified, reliable indicants of malingering were not found. The NF group produced test results which were more indicative of actual psychosis than did the IF group who produced many results which were too pronounced to be believed. The behavioral presentation of both groups was inconsistent with psychosis. They remained in relationships which were complementary with the examiner irrespective of whether they were oppositional or cooperative. It was suggested that malingering could best be identified if the test results were examined in the context of the extra-test behavior, history, and possible motivation of the S.