Abstract
This study evaluated ethnic differences on the Minnesota Multiphasic Personality Inventory-2 (MMPI-2) in 126 Caucasian and 106 Native Hawaiian and Other Pacific Islander (NHOPI) mental health center outpatients. MMPI-2 Restructured Clinical (RC) scale scores were first evaluated to determine the extent or influence that the demographic variables of race, gender, age, and education have on identified group MMPI-2 RC scale differences. When assessing for racial group mean differences, five scales studied were identified to be of significance. When assessing for age mean differences, two scales were observed to have been of significance, although there was no significant interaction between race and age for either of these scales. No differences in gender were observed and there were no gender by race interactions. The findings suggest that gender and age had less effect on how a participant would respond to MMPI-2 questions than race. Because preliminary analyses identified a racial group difference by education, this variable was controlled for through a MANCOVA. With the exception of RC scale RC8, racial group differences were not confounded by education. These results were generally consistent with the observed racial group differences on five of the MMPI-2 RC scales and the available extratest criterion data. To assess for test bias, a step-down hierarchical multiple regression procedure was used in which relevant clinical criteria were predicted from the RC scales with race as a potential moderator. Of the 19 scale criterion comparisons, nine demonstrated statistically significant intercept bias of which five achieved a small effect size. Two scales from the regression analyses revealed statistically significant slope bias, both of which demonstrated a small effect size. Although several of the MMPI-2 RC scales suggested some overprediction and underprediction of psychopathology in Native Hawaiian and Other Pacific Islanders, it was determined that racial group membership was not a substantive moderator between MMPI-2 scales scores and clinically relevant phenomena because the effect size of race on prediction was uniformly small. It was recommended that the MMPI-2 RC scales not be interpreted differently for Caucasian and Native Hawaiian and Other Pacific Islanders, as the magnitude of the predictor error did not warrant separate interpretive guidelines.
|
