Intrinsic motivation, the preferred facilitator of performance, may be a relatively stable trait or specific to a given task. This study compared the computer task performances of 207 children in two schools, on the basis of their teachers' reward practices and the experimental reward conditions. Parents' reward practices, teachers' reward practices, and children's trait intrinsic motivation were measured. Baseline task performance scores and the chosen level of difficulty were statistically higher for children who were higher in the Judgment subscale of trait motivation than for those who scored lower on the Judgment subscale. The trait measure was positively related to most of the game scores and difficulty levels of the task motivation. Higher parent reward usage was related to lower SES and to lower achievement. Higher teacher reward usage was positively related to Grade 1 and to higher levels of difficulty. Children who had usually received rewards or who had not usually received rewards, according to a teacher survey of rewarding attitudes and behaviors, were given a challenging task with (a) no mention of rewards or (b) the promise of a reward. Experimental reward conditions consonant with reward experiences related to higher game scores, especially in Grade 1. Experimental reward conditions which differed from reward experiences related to lower game scores after the experimental condition. All scores were higher for Grade 2, except the number of minutes played. The subjects' choice of level of difficulty tended to increase throughout the three trials. The number of minutes played tended to increase during trials in Grade 1 and to decrease in Grade 2. Affect for the task was higher For Grade 2, higher For girls, and higher for Grade 2 children who were lower on the internal Judgment subscale of trait intrinsic motivation.