Abstract
Sampling distributions are obtained for scores on the grid form of the Role Construct Repertory Test (Rep Test) using a null hypothe sis of random behavior. Three different scoring rules are considered; the method of tied-ratings based on 6-point scales (Bieri et al., 1966), the method of tied-ratings based on 2-point scales (Scott, 1962) and Fiedler's most preferred person score (Fiedler, 1967). In all three cases the mean and variance of the sampling distribution under Ho are given exactly and the distribution form is shown to be approxi mately normal. The sampling theory is compared to empirical re sults reported by Schneier (1979) for Bieri's rule. The discrepancies between theory and behavior form the basis for the discussion of a few test design issues.
Get full access to this article
View all access options for this article.
