Incremental validity of the Ego Impairment Index: It's fine when it's there.

Abstract

The results published by W. Perry (2001) demonstrate that with a different criterion more incremental validity can be achieved with the Ego Impairment Index than R. M. Dawes (1999) found using the criterion available to him. Attempting to find statistical rationale for various Rorschach characteristics is, indeed, desirable, especially as a basis for the statistical use of the test. The degree of predictability indicated by most of these statistics, however, still presents a sizable gap between well established validity and much practice, particularly in legal settings. Further, both W. Perry (2001) and D. J. Viglione and M. J. Hilsonroth (2001) raise the issue of evaluating incremental validity in a context where there is some degree of multicolinearity and nesting. The logical equivalence of a nested with a non-nested analysis is established algebraically, and it is noted that the standard concerns about multicolinearity involve the estimations of the coefficients of predictive models--not the degree of predictability per se.