First-year graduate students were asked to indicate through their responses to a BIB which ability measure was the one best indicator of how well they would do in graduate school. The sample was then divided into two parts, those who felt tests were the best indicators of success (test choosers) and those who felt that some other means of assessment was the best for them (non-test choosers). Within-group regressions were then computed and compared using path analysis techniques. The obtained empirical least squares weighting system gave support to the possibility that graduate students could identify those predictors which would yield minimum errors of prediction for them. Indications of the importance of motivational measures as predictors for "non-test" chooser were suggested.