The token economy is a widely used and versatile motivational system within applied behavior analysis. Moreover, token reinforcement procedures have been shown to be highly effective in the treatment of children with autism spectrum disorder (ASD). In the experimental analysis of behavior, token reinforcement contingencies are conceptualized as three interconnected schedule components: (1) the token-production schedule, (2) the exchange-production schedule, and (3) the token-exchange schedule. Basic work with nonhuman subjects has demonstrated that the exchange-production schedule is the primary driver of performance in these arrangements, and that variable-ratio exchange-production results in reduced pre-ratio pausing and greater overall rates of responding relative to fixed-ratio exchange-production schedules. However, little applied research has been conducted to assess the generality of these findings within applied settings. The purpose of this study was to determine if fixed- and variable-ratio token exchange-production schedules would exert differential effects on pre-ratio pausing and overall rates of responding for three children with ASD during a free-operant sorting task. The results showed that pre-ratio pausing and overall rates of responding were not differentially effected by the fixed- and variable-ratio exchange-production schedules. Discrepancies between the experimental work and the current study are discussed along with additional limitations.