One central issue when evaluating the validity of the MCMI is the extent to which va- lidity studies on previous versions can be generalized to the newer versions. With appro- priate caution, some transferability is probable because the correlations between the MCMI-II and MCMI-III scales are moderately high. Specifically, the correlations range from a high of .94 for Debasement to a low of .59 for Dependent, with 12 of the 25 scale comparisons above .70. Comparisons between the Depressive and Posttraumatic Stress Disorder scales could not be made because they were uniquely developed for the MCMI-III. However, Marlowe, Festinger, and Kirby (1998) found much lower correla- tions and little comparability for code types. This should be balanced with their conclu- sion that the MCMI-III did provide comparable clinical information. It should also be noted that they used a relatively small sample of persons who had a quite specific disor- der (cocaine dependence). It would thus be important for future studies to determine the extent to which the MCMI-III is actually comparable to the MCMI-II. Some researchers consider the MCMI-III to be sufficiently different to be considered a separate instrument (Marlowe et al., 1998; R. Rogers, Salekin, & Sewell, 1999). Rogers et al. (1999) have even suggested that, in contexts requiring maximum accountability (forensic), it might be preferable to use the MCMI-II until the validity of the MCMI-III becomes better docu- mented. Given these considerations, the following selective and representative overview of validity draws on material from both the more recent MCMI-III and the MCMI-II.