K-corrected MMPI-168 scales using short and long form K scales (N = 2439). The latter increased the correlation between abbreviated and standard scales in most instances. Because using the entire K scale involves the addition of only 18 items to the MMPI-168 while it generally improves the statistical validity of several scales in addition to K, this was proposed as a cost-effective means of improving the MMPI-168.
Since the introduction of the Mini-Mult by Kincannon in 1968, research on short forms of the MMPI has burgeoned; the current literature involves over 150 citations (Newmark & Faschingbauer, 1978; Stevens & Reilley, 1980). Short forms have enjoyed a modest, but by no means insignificant popularity among clinicians (Moreland & Dahlstrom, Note 1). Therefore, further attempts to improve their reliability and validity seem warranted.
Scoring most short forms involves extrapolating from short to predicted long-form raw scores via equations or tables and then proceeding to score in the usual manner (Faschingbauer & Newmark, 1978). Given that this is the case, it has seemed to this author that one of the most cost-effective ways to improve currently existing short forms would be to score the entire K scale. This presumably would yield, in addition to a perfectly accurate K scale, greater accuracy in those scales to which fractions of K are added prior to T scoring (Hs, PD, PT, Sc, and MA). The increment in accuracy should be proportional to the amount of K-correction. The accuracy of the Faschingbauer Abbreviated MMPI (FAM) already has been shown to be improved by this technique (Moreland, Note 2). The current study explores the increment in validity yielded by full K-correction of another of the most promising short forms, the MMPI-168. This potential improvement in five scales and certain improvement in one involves adding 18 items to the MMPI-168, thereby increasing its length by 1 I%. The author proposes that this revision of the MMPI-168 be referred to as the MMPI-168-E(xtended) because to call it the MMPI-186 would invite confusion.
METHOD
The MMPI-168 was extracted from the full protocols of 1181 psychiatric inpatients and 1258 normals. Hs, PD, PT, and Sc scores were transformed to their long form predictions using the equations proposed by Ward (1980). MA was extrapolated using the equation put forth by Overall, Butcher, and Hunter (1975). These equations had been shown to be the most accurate ones, of those K-corrected in the usual manner, proposed for short-to long-form prediction in a previous study of these same samples (Moreland, in press). (The equations proffered by Overall and Gomez-Mont (1974) were not considered because they predict directly to K-corrected long form scores.) These scales, corrected using the entire K scale, then were correlated with their long form counterparts for groups divided according to sex and race. T scores were used. These correlations were 'This work was supported in part by N l H M grant #3l302. Thanks are due to W.