The purpose of this study was to construct-validate the REPI. The judgmental data of 25 content specialists were examined. The response data of 416 students were item analyzed and submitted to principal components analyses with orthogonal and oblique rotations. The items were generally representative of the content universe and 50 items optimally discriminated among the subjects. Both the REPI defined and factor derived scales were reliable. Uncorrelated factors identified the underlying philosophical dimensions, but could not be combined to fit the item grouping assumption of the REPI. A substantial amount of test variance was explained and factor invariance was demonstrated. Evidence pointed to the internal validity of the instrument. (Author)