The Deconstruction of the Hy Scale of MMPI–2: Failure of RC3 in Measuring Somatic Symptom Expression

Abstract

The MMPI-2 (Butcher, Dahlstrom, Graham, Tellegen, & Kaemmer, 1989) Clinical Scales have a long history in psychological assessment. Recently, Tellegen et al. (2003) conducted a series of analyses to restructure the scales to reduce what they considered to be problems that limit scale functioning. In a critique of the Restructured Clinical (RC) Scales published in this issue, Nichols (2006/this issue) questions a number of aspects of the approach Tellegen et al. took including their theoretical assumptions, methods of analysis, and failures to report important information needed for scale evaluation such as relationships with existing scales. We concur with many points raised by Nichols. In our analysis of the performance of the RC3 scale, we found that it has "drifted" so far from the original Hy scale as to be a completely different measure- a scale of cynical attitudes that is already well represented in existing MMPI-2 measures. In this article, we take these concerns a step further and examine the history and construct validity of the Hy scale in evaluating the somatic expression of problems that the original authors (McKinley & Hathaway, 1944) intended. We also include new information from a medical setting, an application not represented in Tellegen et al.'s RC Scale monograph. In agreement with Rogers et al. (2006/this issue), it is our conclusion that some RC Scales do not represent the measurement domain of the original scales and should not be relied on for or used to refine traditional interpretation, particularly in medical or forensic situations (such as personal injury cases) because of their confusing and conflicting results.

What Do the MMPI–2 Restructured Clinical Scales Reliably Measure? Answers From Multiple Research Settings

“…Although the correlation between RC9 and Scale 9 are reported to be .77 and .80 for men in two of the three clinical samples in the RC 435 Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 21:45 17 November 2014 monograph, the community mental health sample of 410 men yielded a correlation of .26 for RC9 and Scale 9 (Tellegen et al, 2003). The low absolute value of the correlations between RC3 and Scale 3 for men (ranging from -.13 to -.20) and women (ranging from -.09 to .01) reported by the scale developers (Tellegen et al, 2003) and found by others (Butcher, Hamilton, Rouse, & Cumella, 2006) suggests there was no preservation of a "core" of Scale 3. Therefore, additional validation data are needed to identify the constructs measured by each of these scales.…”

Let's Not Get Hysterical: Comparing the MMPI-2 Validity, Clinical, and RC Scales in TBI Litigants Tested for Effort

“…If the conversion V is to resurface in the RC scales profile, it would be expected that RC1 (akin to the Hs scale) would be elevated, RC2 (akin to the D scale) would be moderately elevated, and RC3 (akin to the inverse of the Hy scale) would be depressed. However, Butcher, Hamilton, Rouse, and Cumella (2006) argue that the restructuring of the Hy scale into RC3 resulted in a completely different scale. In particular, Butcher et al point out that the decision to compose RC3 entirely of a cynicism factor and not of a defensive somatization factor-in order to reduce overlap with RC1-has led the scale to drift too far away from McKinley and Hathaway's (1944) original intention.…”

Personality Assessment with the MMPI‐2: Historical Roots, International Adaptations, and Current Challenges

“…Others, however, have pointed to flaws in this approach and the resulting RC Scales (e.g. Butcher, Hamilton, Rouse, & Cumella, 2006; Caldwell, 2006; Gordon, 2006; Nichols, 2006; Ranson et al., in press; Rogers & Sewell, 2006; Rogers, Sewell, Harrison, & Jordan, 2006). For example, Rogers and Sewell (2006, pp.…”

“…Methodological Issues and Findings. Despite their name, which implies a close link to the MMPI‐2 Clinical Scales, the Restructured Clinical Scales are not simply shortened or alternative forms of the Clinical Scales (Butcher, Hamilton, et al., 2006; Nichols, 2006; Rouse, Greene, Butcher, Nichols, & Williams, 2008; Simms, Casillas, Clark, Watson, & Doebbeling, 2005), even though the RC Scales replace the Clinical Scales in the new MMPI‐2‐RF (Ben‐Porath & Tellegen, 2008). As Simms and colleagues (2005, p. 357) explained:…”

“…The conclusions of Rogers and Sewell (2006) and Simms et al. (2005) that the RC Scales differ fundamentally from the Clinical Scales is most clearly illustrated by the drastic change to Hy with its “restructuring” and the resulting RC3 (Butcher, Hamilton, et al., 2006; Gordon, 2006; Nichols, 2006). The Hy scale is a well‐established construct for understanding medical/psychological conditions such as chronic pain (Arbisi & Seime, 2006).…”