Influence of the retest artefact on outcome mesurement in psychotherapy - the Inventory of Interpersonal Problems and the Eating Disorder Inventory

Repeated administration of self-report scales is the method of choice for observing changes over time in many fields of psychology, including psychotherapy research. However, methodological questions arise due to the problem of retest artifact. Retest data of the German versions of the Inventory of Interpersonal Problems (IIP) (N=241) and the Eating Disorder Inventory (EDI) (N=307) with four consecutive assessments were analyzed using the General Linear Model. Only the scores for the EDI-scale "drive for thinness" were found to change significant over multiple administrations. A decrease from the first to the second administration was observed. The estimated effect size was 0.13, the rate of reliable improvement 6%. A correction based on the expected change distribution and the use of a dummy test design are recommended to handle the retest artifact.