Countering perfectionism in research on clinical practice II: Retrospective analysis of treatment progress

Abstract

A simple correlational study is described which has methodological advantages, when compared with an equivalent manipulative (i.e., quasiexperimental) research design, for examining exploratory questions about treatment outcome. These advantages stem from the fact that retrospective designs, when compared with prospective ones, are much less likely to involve ethical problems due to offering relatively ineffective or unnecessary treatment, and less likely to involve confounds from newly hypothesized differential expectations about treatment. Based on Millon Clinical Multiaxial Inventory (MCMI) scale scores, clients with high levels of histrionic or narcissistic (and possibly also those with antisocial, compulsive and paranoid) personality traits seemed to profit more from RET than those without such traits. Clients with high levels of schizoid, avoidant, and schizotypal traits (and probably also clients with strong passive-aggressive traits, and possibly also clients with strong dependent and borderline traits) seemed to profit less from RET than those without such marked traits. We have previously shown that the relatively “healthy” personality traits and disorders are associated with global self-upping and the relatively “sick” ones with global self-downing, as well as contrasting patterns of positive and negative reactions to stressful events. Our findings here support the concept that clients who consistently handle life situations well will tend to progress most rapidly in therapy, and those who generally handle life situations badly will tend to progress slowly. Personality traits may be especially significant markers for therapeutic attention to problems stemming from global self-rating. Expectations based on the severity of global self-upping and/or self-downing may be good predictors of therapeutic progress. In contrast, expectations about RET outcomes based on severity of distress or “neediness” will probably be less accurate.