View in Plain Text


Assorted topics: brief comments and opinions


Alex B. Caldwell, Ph.D.


Topics: RC, Ho, AAS and APS, GM and GF, S, F(p), and non-gendered norms


1. What clinical contributions can the RC scales make? Are there potential uses? What validation do they need? How can we respond to misleading RC testimony? I do not believe these are the "cores" of the clinical scales as claimed; (for detailed discussions of the deficiencies of the RC scales see Butcher et al., 2006, Caldwell, 2006, and Nichols, 2006). The RC scales (Tellegen, A., Ben-Porath, Y. S., McNulty, J., Arbisi, P., Graham, J. R., & Kaemmer, B. (2003). The MMPI-2 Restructured Clinical scales: Development, validation, and interpretation. Minneapolis: University of Minnesota Press) depend heavily on the face validity of the items, and the items are so transparently obvious that the "scored direction" usually is readily discernable (just as with the content scales). The one unique and clinically interesting contribution is the RC8 scale, the restructured version of scale 8-Sc. RC8 collects the reality-disturbed but non-paranoid content in 8-Sc (feelings of unreality, peculiar experiences, hearing strange things, etc.). Their subjective realness to the individual may at times bypass the often pervasive suppression of persecutory content when L is much elevated. These RC scales need to be studied by neutral researchers who have no investment in the outcome, especially in circumstances that include incentives to bias one’s responses, both self-favorably and self-negatively in different samples.


Suppose that in a custody case, for example, the opposing parent has by history a problematic conscience, e.g., financial deceptions, abusiveness, dishonesty, or etc. The opposite expert utilizes a very low score on the RC4 scale to argue that that parent does not test as psychopathic, i.e., is not a scoundrel, can be trusted to reform as promised, etc. Your MMPI-2 profile has an at least somewhat elevated score on scale 4-Pd as well as an L and/or K score that is mid-60's or higher, etc.: the parent was clearly defensive in responding to the test. The opposite expert can be presented with the RC4 items together with the parent’s corresponding responses and be asked to read the RC4 items to the court. (Note that the RC4 items are obvious as to what is the desirable response (trouble with the law, truant in school, past petty thievery, engaged in fights, etc.). After reading the items to the court, the expert can be asked such questions as, "This client has completed at least three years of college, hasn’t he/she doctor?" "Would these items not be quite obvious as to what is the more favorable response?" "Has this parent not already been shown to have been quite defensive toward the test?" Note the face valid interpretation of the RC4 items: such presentations as this can help to expose the shallowness of depending on face valid uses of item responses.


2. Where does the Ho scale come from? What can it tell us? The Cook-Medley Hostility scale was originally developed as a schoolteacher student-applicant-screening scale. Those who tended to see many of the kids as potentially ill-behaving brats often needing restraint tended to fail as teachers; those who mostly saw them as generally interested and wanting to learn much more often succeeded. This unexpectedly showed up many years later as the best predictor of atherosclerosis in the 25 year follow-up of physicians who had been University of North Carolina medical students. Such cynicism was seen as making an interesting contribution to atherosclerosis. The content scale CYN seems directly fashioned after Ho, which latter I believe is a bit more subtle and a somewhat better scale.


3. What are the utilities of the newer addiction scales AAS and APS? AAS (Addiction Admission Scale) is basically a set of critical items regarding chemical abuse, extensively overlapping the corresponding critical item sets for chemical abuse in our and other computer-generated reports. We provide AAS because it appears useful in order to pick up some test misses from the MAC-R, although it can overweight episodic abuse whereas the MAC-R appears more a predictor of daily dependence. APS (Addiction Potential Scale) is a competitor to the MAC-R. I haven’t seen much research on it since the original derivation of the scale. I think it is a bit less subtle and not as consistently effective as the MAC-R, which is one of the most non-transparent and non-obvious sets of items of which I am aware (liking the work of a forest ranger, liking to cook, etc.). We need a non-obvious set of items because of the frequency if not also the urgency of the motivation to disguise one’s problems with chemicals. I think the MAC-R just works better, and it has a far longer history and breadth of acceptance.


4. How do the GM and GF scales for the MMPI-2 fit into our thinking about gender identity and gender roles? Gender Masculine and Gender Feminine are curious scales. To a considerable degree they are the more stereotypic, expected responses such as to like hunting vs. liking to play "house" as a child, etc. What is most curious is how low the correlations are (mostly +.20s to +.30s) not only with Mf but also with each other. A person can have high scores on both scales with a diversity of gender-related interests as well as being low on both, reporting relatively few such interests in either direction. Mf contains numerous items about feelings of interpersonal sensitivity, avoidance of loud, rough play, etc., that do not appear in either GM or GF. GM and GF appear to me to be more gender role related versus Mf as predominantly gender identity determined.


5. What is the usefulness of the S scale, Superlative adjustment? Scale S seems to me to be confounded with at least three elements: conscious defensiveness (airline pilot training applicants), genuinely good adjustments (creeps and weirdos need not apply to be airline pilots), and an over-control quality akin to or associated with the 34/43 MMPI code that is relatively frequent among pilot applicants. Thus confounded, it contains a quite limited amount of unique information, and I find limited use for it because of the uncertainty as to which psychological aspects are contributing to the elevation. Some clinicians like the subscales of the S scale as illuminating specific attitudes on an item content basis.


6. Where does the F(p), or psychiatric F scale fit into validity assessments? This is an interesting scale. The original F scale was a set of responses that normals rarely made. F(p) is a set of responses that neither normals nor patients rarely make. The items carry a distinctly greater degree of implausibility, e.g., all laws should be thrown away, than the other rare response and malingering scales. F(p) seems to work well with severely disturbed populations, e.g., state hospital and VA patients (the latter often being (1) seriously disturbed, (2) concerned about their pensions, and (3) of lower SES levels, all three of which can simultaneously inflate both their clinical profile elevations and specifically their scale F elevations. Note that the correlation of F and the Ss or socioeconomic status scale (Nelson, 1952) is -.77 in the Caldwell Data Set, n = 52,543, a huge contribution of Ss (SES) to F. The correlation of Ss to F(p) is .59, considerably less of the F(p) variance, 35%, as compared to the 59% contribution (r at = .77) of Ss to F. In addition, the F(p) scale has a serious flaw: four items overlap the L scale, being more extreme and rarely claimed denials of less than proper behaviors. But this subset of items operates in the opposite direction. Thus, an unsophisticated, guarded subject can get an elevation on F(p) from being defensive and perhaps somewhat ill-comprehending in item interpretation without any effort to fake sick: beware F(p) and L up together. The scale needs to be re-normed without the L items.


7. What about non-gendered T scores? Clinically, I think this is psychometric bosh and nonsense. As with the non-K-correction idea, all of the useful interpretive research on the code types has been on the traditional gender-specific and K-corrected original MMPI norms. There is no body of in-depth pattern research on the MMPI-2 where, for about one-third of psychiatric patients and just over one-half of normals, the same set of raw scores will generate two different codes using the MMPI vs. MMPI-2 norms. Also, we need gender-specific norms to filter out - as best we can - gender-determined but non-pathological influences on item responding and scores. For example, averaged across women, do their menstrual cycles and associated physical vulnerabilities tend to heighten their attention to the states of their bodies? Ironically, the consequences of using non-gendered norms for anti-discriminatory reasons end up being more discriminatory against women than are the gender-specific norms. For example, any average heightening of somatic attention tends to raise the raw and T scores on scales 1-Hs, 2-D, and 3-Hy. On the average, women do answer a few more somatic items, but I believe this is primarily on a commonly shared experiential basis rather than on an individually psychopathological basis. In sum, I have yet to perceive any clinical utility for the non-gendered norms.




Butcher, J. N., Hamilton, C. K., Rouse, S. V., & Cumella, E. J. (2006). The deconstruction of the Hy scale of MMPI-2: Failure of RC3 in measuring somatic symptom expression. Journal of Personality Assessment, 87(2), 186-192.


Caldwell, A. B. (2006). Maximal measurement or meaningful measurement: The interpretive challenges of the MMPI-2 restructured clinical (RC) scales. Journal of Personality Assessment, 87(2), 193-201.


Nelson, S. (1952). The development of an indirect, objective measure of social status and its relationship to certain psychiatric syndromes (Doctoral dissertation, University of Minnesota, Dissertation Abstracts, 12, 782. Discussed in Caldwell, A. B., (1997). Whither goest our redoubtable mentor, the MMPI/MMPI-2? Journal of Personality Assessment, 68(1), 47-68.


Nichols, D. S. (2006). The trials of separating bath water from baby: A review and critique of the MMPI-2 restructured clinical scales. Journal of Personality Assessment, 87(2), 121-138. Tellegen, A., Ben-Porath, Y. S., McNulty, J., Arbisi, P., Graham, J. R., & Kaemmer, B. (2003). The MMPI-2 Restructured Clinical scales: Development, validation, and interpretation. Minneapolis: University of Minnesota Press.