The effects of brief rest pauses on performance and well-being were evaluated for a highly repetitive, data entry task. Experienced data entry operators (N = 20) performed the task in a two-day experiment in a simulated office environment. Each day was divided into six, 40-min work periods. Subjects took a brief rest pause at the workstation (microbreak) in the middle of each work period. Subjects were instructed to terminate this microbreak when ready to resume work. Keystroke rate, error rate, correction rate, heart rate and heart rate variability were scored for each half of the work period. In addition, mood states before and during the work period were assessed. Microbreaks were found to average 27.4 s in duration. High ratings of fatigue and boredom during the work period were associated with longer microbreaks, suggesting that the break period was self-adjusted relative to mood state. In addition, correction rate and heart rate were lower following long microbreaks, implying that the degree of recovery was linked to the length of the microbreak. Comparison of keystroke output and correction rate before and after the microbreak, however, revealed that performance worsened after the microbreak, suggesting that subjects terminated microbreaks before complete recovery could occur.