Abstract
Distributions of all possible split-half correlations for the Harris and Lingoes, Wiggins, and Serkownek Scales of the MMPI that contained twelve items or fewer revealed significant negative skewness that resulted in spuriously lowered estimates of reliability. Corrections based on the optimized r11 increased the average Subscale split-half correlation from an alpha of .409 (or the average odd-even split of .411) to a maximized r11 of .604. Generally, reliability of scales, with the maximized r11 increased to levels acceptable for research purposes. It was concluded that research scales of the MMPI show split-half distributional anomalies which should dictate the use of maximized r11 to estimate internal consistency. It is suggested that these findings might be taken into consideration in the evaluation of the new MMPI.
Get full access to this article
View all access options for this article.
