Parisien, M. (2021). Empirical vs. factorial validity in personality inventories: The MMPI-2 and the restructured RC scales. The Quantitative Methods for Psychology, 17(3), 329-344.
- Robert M. Gordon
- Jun 22
- 2 min read
Abstract
Since 2003, several American publications have appeared on “Restructured Clinical Scales RC”. These scales were constructed in contradiction with the empirical psychometric strategy adopted from the outset for the MMPI (1943) and MMPI-2 (1989). Rather, the RC scales are based on a theory-factor strategy, which long-time MMPI/MMPI-2 experts disavow because it does not fulll its promise of validity. In 2003, the RC scales (Tellegen et al., 2003) were rst arbitrarily grafted to the set of empirical MMPI-2 scales. Then, in 2008 (see Tellegen and Ben-Porath), they were instead integrated as an asset to a brand-new instrument based on factor analysis, called MMPI-2-RF on the sole basis of borrowing 338 of the 567 items of the MMPI-2. The background to this article is therefore the possible adulteration of the most important personality test used in forensic expertise and in clinical contexts. Our goal is to review both the psychometric foundations of the MMPI-2 and the value of these RC scales (Tellegen et al., 2003) in the now open confrontation with the MMPI-2.
This RF form should therefore be considered as a new instrument and not as an update of the MMPI-2 as was the case when the MMPI was changed to MMPI-2. In short, the MMPI-2-RF is a new psychometric instrument and was not validated on several aspects. It lacks the empirical support (with attendant credibility) and the interpretive data of the MMPI-2 (Friedman et al., 2015). The designation MMPI-2 (RF) is even considered to be undue, even illegitimate. To add insult to injury, the University of Minnesota Press publishers and their distributors Pearson Assessments recently released an "MMPI-3" (emerged from MMPI-2-RF), which basically has nothing to do with MMPI/MMPI-2. For the moment, the RC scales, resulting from a theoretical and factorial strategy, are no match for the empirical psychometric strategy that made and still makes the MMPI/MMPI-2 instruments the nec plus ultra of personality inventories used in clinical and psycho-legal contexts. While, in principle, it is interesting for a personality inventory to shed light on the factorial constructs that constitute it for the user, it aims first and foremost to produce convincing data and reliable indications for practice, either clinical or psycho-legal, as the empirical approach of the classic MMPI/MMPI-2 allows. The identification of a syndrome or a clinical disorder from an objective external criterion is here the essential element of the empirical psychometric strategy. Nowadays, it typically proceeds by discriminant function analysis, giving rise to a form of validity which is also called "discriminant" or “criterion” validity (McLachlan, 2004). Discriminant function analysis makes it possible, by association of items on identified, ‘typed’ groups, to capture the configuration of “facets” (Guttman, 1954; Guttman & Greenbaum, 1998) which constitutes each type and is likely to trigger or reveal psychopathology. It is the correlations of the items with an external criterion that mattermost, not the correlations between items. If one undertook a re-validation of the clinical scales of MMPI/MMPI-2 by forming new clinical groups for comparison with a normal sample, which has not been undertaken since the 1940s, perhaps it would represent the solution for the future. In the meantime, a possible demise of MMPI-2 would be the result of poorly thought-out psychometrics and questionable editorial maneuvers, with disastrous clinical repercussions.
Comments