Three consecutive experiments examined the effect of offering college students performance-contingent points toward their final course grade under various combinations of knowledge of correct response and rules with examples of how to perform the task. Students who were offered points and received rules with examples scored lower on an immediate post test than students who received rules with examples but were not offered the points. The situation was reversed for students who did not receive rules with examples: post test scores were higher when points were offered. This disordinal interaction between performance-contingent points and rules with examples was significant (p < .05) in all three experiments. The main effect of rules with examples was significant (p < .001) across all three studies. Neither points nor knowledge of correct response produced significant main effects. (Author/MV)