Richards, Gaver, and Golicz (1984) found that, in contrast to peers whose grades were accurately predicted from performance on aptitude tests, both extremely underachieving and extremely overachieving fourth-graders had negative academic attitudes. The present study aimed to replicate and extend these findings. Subjects in the replication study were predominantly from lower-class, rural homes, differing from the suburban, middle-class sample of the original study. In the replication, 30 fourth-grade, 71 fifth-grade, and 80 sixth-grade students were administered the Elementary Form of the Estes Attitude Scales and the Science Research Associates battery. Also, end of year mathematics, reading, and science grades were obtained. Results of regression analyses and one-way analyses of variance did not replicate the findings of Richards and colleagues, as far as overachievers were concerned: at every grade level extreme underachievers displayed the most negative attitudes and extreme overacheivers displayed the most positive. The discrepancy between findings of the original and those of the replication study is speculatively attributed to demographic differences in the samples. (RH)