Development and Effects of an MMPI--A K-Correction Procedure

Abstract

Although the results of research on the effectiveness of the K-correction factor have been inconclusive, this procedure has been widely used with adult respondents to correct for defensiveness or underreporting of symptomatology on the Minnesota Multiphasic Personality Inventory. Although the K-correction procedure was incorporated into the Minnesota Multiphasic Personality Inventory-2, the Minnesota Multiphasic Personality Inventory-Adolescent (MMPI-A) was developed based exclusively on the use of non-K-corrected T scores. This study derived age-appropriate K-weights for the MMPI-A to determine the degree to which the use of this procedure could improve test accuracy in the classification of participants into normal and clinical groups. Discriminant function analyses were performed to determine the K-weight that, when combined with basic scale raw score values, optimally predicted normal versus clinical status for each of the eight basic clinical scales. Hit rate analyses were utilised to assess the degree to which K-corrected T scores resulted in improvements in classification accuracy in contrast to standard MMPI-A non K-corrected norms. Results indicate that the adoption of K-correction procedure for the MMPI-A does not result in systematic improvements in test accuracy and the current findings do not support the clinical use of a K-correction factor in interpreting MMPI-A protocols.

The Classification Accuracy of the Minnesota Multiphasic Personality Inventory–Adolescent

“…Consistent with the prior findings reported by Alperin et al (1996) and Fontaine et al (2001), reduction of the T-score cutoff from 65 to 60 to identify pathological adolescents does reduce the frequency of W-N-L profiles but does not significantly improve overall classification accuracy, either with the standard MMPI-A norms or with the modified norms. Classification accuracy as reflected in hit rate and positive predictive power was consistently higher at the T ≥ 65 cutoff level in all analyses.…”

“…Several researchers have attempted to account for the high frequency of W-N-L basic scale profiles for adolescents in clinical treatment settings. Alperin, Archer, and Coates (1996), for example, developed an MMPI-A K-correction procedure in an effort to improve test sensitivity. They found, as did Marks, Seeman, and Haller (1974) in an earlier study with adolescents using the original MMPI, that the K-correction procedure failed to improve the overall test accuracy.…”

Identifying Causes of Disagreement Between Self‐Reports and Spouse Ratings of Personality

“…But sometimes the information will be wrong, and they have no way of knowing when. Validity scales have been devised to identify responses distorted by social desirability and defensiveness, but they have not worked well (see Alperin, Archer, & Coates, 1996;Ones, Viswesvaran, & Reiss, 1996), perhaps because they have addressed the wrong sources of distortion. No validity scale has been designed to reveal whether the respondent has attended to the full range of relevant behaviors, or used the appropriate reference group-yet these are more much more common reasons for discrepancies than faking good or extreme responding.…”

“…In practice, however, this has proven to be extremely difficult. Great ingenuity and effort have gone into the design of validity scales to detect and correct response biases, but with limited success (Dicken, 1963;Alperin, Archer, & Coates, 1996). Indeed, a number of studies have shown that corrections for defensiveness or socially desirable responding sometimes have the effect of reducing rather than enhancing validity (see Barrick & Mount, 1996;McCrae & Costa, 1983;McCrae et al, 1989).…”

MMPI-A Characteristics of Male Adolescents in Juvenile Justice and Clinical Treatment Settings

“…Furthermore, these researchers observed that the MMPI-A basic and content scales generally show a much lower percentage of effective items than do corresponding scales for the MMPI-2. They also noted that the relatively low MMPI-A clinical scale mean T-scores found for clinical samples of adolescents also tended to be produced on the original form of the MMPI by adolescents in clinical samples and that efforts to resolve this problem by developing a K correction for the basic scales (e.g., Alperin, Archer, & Coates, 1996) or by lowering the T-score criterion used to define clinical range elevations (Fontaine, Archer, Elkins, & Johansen, 2001) have proved unsuccessful.…”