An Examination of the Bilevel Dimensionality of Lower and Higher Mental Processes in Probability Achievement

Abstract

The purpose of this study was to examine the bilevel dimensionality of probability achievement, as operationalized by Levin (1975), through using three psychometric procedures: (1) Rasch latent trait analysis; (2) factor analysis; and (3) classical item analysis. It was found that seven items exceeded the criterion fit for the unilevel latent trait model. The remaining items defined a unidimensional trait. The factor structure lent support to the concept of unidimensionality. Further, the point biserial coefficients indicated increased homogeneity among only Rasch fitting items compared with that in the original test.