Abstract
The purpose of this study was to expand the empirical basis for interpretation of the Minnesota Multiphasic Personality Inventory–Adolescent (MMPI–A; Butcher et al., 1992). Participants were 157 boys from a forensic setting and 197 girls from an acute psychiatric inpatient setting. Criterion variables were identified from sources such as psychiatrist report, parent report, and psychosocial history. Results generally support the construct validity of MMPI–A scales. Scales measuring internalizing problems were more highly correlated with criterion measures of internalizing behaviors than measures of externalizing behaviors, whereas scales measuring externalizing problems were more highly correlated with externalizing variables than with internalizing criteria. Implications of this study include an expanded empirical foundation for interpretation of the MMPI–A, greater understanding of the constructs it measures, and evidence supporting the generalizability of these constructs across settings.
Acknowledgments
Support for this study was granted in part by the University of Minnesota Press, publisher of the MMPI–A. We thank Christa Veltri, Stephanie Miller, Wendy Dragon, and Jill Barnes for their contributions to this study. Portions of this article were presented in 2005 at the 40th Annual Symposium on Recent Developments in the Use of the MMPI–2/MMPI–A in Fort Lauderdale, Florida and the 2006 Annual Meeting of the Society for Personality Assessment, San Diego, California.
Notes
a The base rate of that criterion variable in that sample was < 10% or > 90%; thus, correlations may be attenuated because of range restriction.
b The interrater reliability was low or unavailable; thus, correlations may be attenuated because of poor reliability. The em dash (—) indicates that a correlation could not be calculated because the base rate for the RRF variable was 0%. Coefficients in bold indicate that p < .002.
a The base rate of that criterion variable in that sample was < 10% or > 90%; thus, correlations may be attenuated because of range restriction.
b The interrater reliability was low or unavailable; thus, correlations may be attenuated because of poor reliability. The em dash (—) indicates that a correlation could not be calculated because the base rate for the RRF variable was 0%. Coefficients in bold indicate that p < .002.
a The base rate of that criterion variable in that sample was < 10% or > 90%; thus, correlations may be attenuated because of range restriction.
b The interrater reliability was low or unavailable; thus, correlations may be attenuated because of poor reliability. The em dash (—) indicates that a correlation could not be calculated because the base rate for the RRF variable was 0%. Coefficients in bold indicate that p < .002.