Replicating a 1978 study of methods of setting cut-off scores in standardized tests for student placement, the Rasch model was used to estimate ability levels and item difficulty for use in determining the probability of course success. Community college students were administered the New Jersey College Basic Skills Placement Tests and items were calibrated by the BICAL program. Item difficulties were analyzed using a three-course mathematics sequence, a remedial mathematics course, and a three-course English sequence. Success was generally defined by earned grade. The comparisons of different grade cut-offs showed the course grades can be transformed into dichotomous success "items" which fit the Rasch model. (Auth/CM)