Retest reliability of the Rosenzweig Picture-Frustration Study and similar semiprojective techniques

Surveys the evidence for the reliability of the Adult, Adolescent and Children's Forms of the Rosenzweig Picture-Frustration (P-F) Study. Analysis of variance and split-half methods, which assume item homogeneity and internal consistency, may be appropriate to psychometric tests, but they are not apposite to projective and semi-projective techniques. The implications of applying such criteria are critically examined, with the P-F Study as an example to demonstrate their nonapplicability. Investigations on retest (total test) reliability from the laboratory of the senior author and from those of other investigators demonstrate statistically significant consistency for the main scoring categories of the P-F, especially those relating to directions of aggression. Implications for other semiprojective techniques are expounded.