Internal consistency of the Group Conformity Rating of the Rosenzweig Picture-Frustration Study.

Abstract

T HE Rosenzweig Picture-Frustration Study (referred to herein as the P-F) yields, among other measures, a Group Conformity Rating (GCR) [10, p. 201]. Rosenzweig states that the GCR "forms one basis for judging social adjustment" [11, p. 168], is indicative of "capacity for conformity with conventional behavior under stress" [11, p. 172], and may reveal whether a subject is "capable of conventional responses in social situations and . . . probably . . . , at least superficially, better adjusted" [11, p. 180].Capacity or capability, of course, can only be inferred, since neither is necessarily ever revealed in overt behavior. Rosenzweig's concept of "group conformity" may be said to be at the level of a hypothetical construct, and subject to experimental verification against an external criterion, but not subject to disproof by the same method. At least one corollary can, however, be deduced from the concept, and furnish a hypothesis which can be tested against a criterion internal to the GCR.For those unfamiliar with the mechanics of the P-F and GCR, a brief description, paraphrased from the manual [10, pp. 165-166, 201], may serve to clarify some of the arguments to follow.Description of the GCR. The items are cartoonlike pictures, each depicting two persons involved in a mildly frustrating situation of common occurrence. One person is shown saying certain words which either help to describe the frustration of the other individual, or which are themselves actually frustrating to him. The subject is instructed to write the first appropriate reply which enters his mind. 1This study was conducted at the Allegheny Vocational Counseling Center with the assistance of a grant from the Buhl Foundation.

Evidence for the validity of the children's form of the Picture-Frustration Study.

“…It is similar in this respect to the GCR derived from the adult form of the P-F. The Kuder-Richardson reliability of the adult form for a sample of 130 adult males was -.04 (12). In the present study the comparable reliabilities are .14 for the Lows and -.60 for the Highs.…”

“…In the present study the comparable reliabilities are .14 for the Lows and -.60 for the Highs. As Taylor and Taylor (12) point out, a negative reliability coefficient obtained by means of the Kuder-Richardson formula indicates that the true reliability probably does not differ significantly from zero.…”

Internal consistency of the scoring categories of the Rosenzweig Picture-Frustration Study.

“…There are certain peculiarities of the P-F which, regardless of population, may seriously limit its reliability. Large "Items" variances militate against high reliabilities here as with the GCR [9]. It may not have been the intention to select stimuli for the P-F which would strongly favor responses in one scoring category on one item, in a second category on a second item, and so on, but that is the effect.…”

“…In the analysis of variance missing responses were treated as zero scores without loss in degrees of freedom [9]. When degrees of freedom for total and remainder were reduced by the number of missing responses, the resulting rtt's were slightly reduced, but only one sample coefficient was as much as .02 smaller.…”

Retest Reliability of the Rosenzweig Picture-Frustration Study and Similar Semiprojective Techniques

“…Taylor (1952) and Taylor and Taylor (1951) evaluated the consistency of the P-F scoring categories and GCR by analysis of variance methods. They reported little reliability for the GCR and a range of reliability for the scoring categories from .58 to .lo.…”