Recently, Hartmann and Hartmann (2014) found that psychiatric outpatients, both with and without access to Internet-based information about the Rorschach Inkblot Method (RIM; Weiner, 2003 ) and the MMPI-2 (Butcher, Dahlstrom, Graham, Tellegen, & Kaemmer, 1989 ), were unable to imitate healthy test performance on these tests. We replicated the study by administering the RIM and the MMPI-2 to 63 incarcerated violent offenders using similar testing conditions. As in the previous study, comparisons were made not only among the 3 subgroups of incarcerated offenders, but also between these offender groups and the group of nonpatients examined in the previous study. On the RIM, Internet-coached and uncoached "faking good" offenders produced records with significantly higher F% and X-% and significantly lower M, m, SumC, X+%, P, AG, and COP than nonoffenders under standard instructions (effect sizes between d = 0.24 and d = 2.39). For AgC, AgPot, AgPast, and TCI% there were no significant differences between the faking offenders and the nonoffenders under standard instructions. On the MMPI-2 clinical scales, there were no significant differences between the faking good groups and the nonoffenders under standard instructions, except on Hs, Pd, and Sc. Both faking groups were identifiable by their high L scale scores. Although both faking groups managed to avoid giving responses with aggressive and generally psychopathological content on the RIM, they were unable to produce test profiles demonstrating healthy test performance on any of the tests; nevertheless, Internet-based test information might weaken test validity.
