The purpose of this study was to measure nonverbally the preference of alternative responses when the net probability of being rewarded was the same. A demonstration of preference under these circumstances would suggest the ability to control or maintain behavior without explicit administration of a reinforcing agent. Head Start children were used as subjects. They were provided, in the experimental situation, with a two-button console. The right button, when pushed, resulted in the illumination of a yellow light and the dispensing of a penny for every second illumination (a consistent reward schedule). The pushing of the left button would result in a 50 percent chance of the illumination of a red light, which was never followed by reward, and a 50 percent chance of the illumination of a green light, which was always rewarded. Thus, whichever button was pushed, there followed a net 50 percent chance of reward. However, only the right button provided a consistent 50 percent reward. The results indicate that children prefer a consistent reward situation to a reward uncertainty situation. (WD)