Viewing the MMPI–2–RF Structure Through the Personality Psychopathology Five (PSY–5) Lens

Abstract

Ben-Porath and Tellegen (2008) recommend organizing MMPI–2–RF scale interpretive information around 3 broad topics, emotional/internalizing dysfunction, thought dysfunction, and externalizing/behavioral dysfunction, and 3 additional topics labeled somatic complaints, interpersonal functioning, and interests. That organization is based primarily on structural analyses of the Restructured Clinical (RC) scales. This study reviewed the MMPI–2–RF's scale structure when the Personality Psychopathology Five (PSY–5) scales are included. Principal axis factor analyses with oblique rotation were conducted on the Restructured Clinical, PSY–5, and Special Problem (SP) scales in 2 samples, by gender. One sample was an outpatient community health center, the other a large, metropolitan inpatient psychiatric facility. The 6-factor solution evidenced each of the PSY–5 constructs plus a general somatic concerns factor. Implications of this solution in comparison to the 3-factor organizing structure recommended by Ben-Porath and Tellegen are discussed.

Acknowledgments

We wish to thank Beverly Kaemmer and the University of Minnesota Press for permission to use the MMPI–2 Normative data set, and Yossef Ben-Porath, John Graham, and Paul Arbisi for permission to use the two study data sets. We also thank Martin Selbom for his recommendations and assistance on methodological issues.

Notes

Participant level of education in the inpatient sample was coded based on year of schooling achieved. For example, a grade school level of education was coded as a 1, and college graduate education as a 13. Rough conversions were calculated to approximate actual years of education.

Tellegen's correction formula is:

where C is the correction to be applied to the original correlation between scales X and Y, S_x and S_y are the standard deviations of the original scales, S_s ² is the variance of the composite formed by items shared between the original scales that are scored in the same direction, S_o ² is the variance of the shared item composite of items scored in opposite directions, and R_ss’ and R_oo’ are reliability estimates for the item composites scored in the same or opposite directions, respectively.

Adjustment components and original and adjusted correlation matrices are available from the first author, as are the structure and factor correlation matrices, for all four study subsamples.