The primary purpose of this study was to examine, using meta-analytical techniques, the differential effects of differing intensities of acute exercise on speed and accuracy of cognition. Overall, exercise demonstrated a small, significant mean effect size (g = 0.14, p less than 0.01) on cognition. Examination of the comparison between speed and accuracy dependent variables showed that speed accounted for most of the effect. For speed, moderate intensity exercise demonstrated a significantly larger mean effect size than those for low and high intensities. For speed of processing during moderate intensity exercise, central executive tasks showed a larger effect size than recall and alertness/attention tasks; and mean effect size for counterbalanced or randomized studies was significantly greater than for studies in which a pre-exercise followed by during or post-exercise protocol was used. There was no significant difference between mean effect sizes when testing took place post-exercise compared to during exercise for speed but accuracy studies demonstrated a significantly larger mean effect size post-exercise. It was concluded that increased arousal during moderate intensity exercise resulted in faster speed of processing. The very limited effect on accuracy may be due to the failure to choose tests which are complex enough to measure exercise-induced changes in accuracy of performance. (Contains 2 tables.)