The presence of significant amounts of double alternation in a 70:30 binary choice problem is demonstrated among first-grade children as a function of positive reinforcement and of instructions that stress only correct responses. Evidence supports the notion that Ss understood the response behavior required of them, but nevertheless persisted in this stereotyped response. Like single alternation, double alternation poses problems for various learning theories. However, the hypotheses advanced to account for single alternation do not appear to explain double alternation. A partial and cognitive explanation that assumes that Ss impose complexity on what they perceive to be a boring task is offered.