The relationship between teachers' and students' personal constructs regarding intrinsic motivation in the mathematics class were examined in this study. Participants were six middle-school mathematics teachers and 30 students from 5 classes. Videotape, direct observation, individual interviews, and repertory grid tasks focused on the ways in which teachers attempted to build their students' motivations into their lessons, and the belief systems of teachers vs. students. Recent mathematics activities for each class served as elements for construct elicitation in the repertory grid task. Teachers and students were presented with random pairs of activities and were asked to determine what made one activity more fun than the other. Responses (constructs) were entered into a computer program that paired each activity with each construct and asked participants to rate how well each construct described each activity. Results revealed that the individuals studied whether teachers or students were similar in their construct systems. Despite the similarities, the differences that were apparent seem to be problematic to the extent to which teachers can anticipate the motivation of their students. Teachers did pay attention to motivating their students in developing their lesson plans, but the ways in which they attempted to build motivating exercises seemed to be more dependent upon the teachers' personal conceptions of intrinsic motivation than their beliefs about their students. Most of the studied teachers had little notion of the motivational beliefs of their students. Teachers' and students' cognitive organization of construct supported the model proposed by Middleton, Littlefield, and Lehrer (1992). Both students and teachers tended to stress the interrelationship between arousal and control levels in determining the intrinsic motivation of mathematics activities. Results are examined in relation to the need to inform teachers regarding the dynamics of student motivation, and to pay particular attention to the individual differences in students' motivational beliefs. In general, results indicate that when teachers are able to predict their students' beliefs, they are better able to fine tune their instruction to turn kids on to mathematics. (Contains 24 references.) (Author)