The performance of children on concurrent variable-interval variable-time schedules

Six 5-year-old children were reinforced on concurrent variable-interval variable-time schedules for a button pressing task. Reinforcers consisted of candies that could either be consumed or traded for other items or activities. The proportion of the total reinforcement available on the two component schedules was varied and the effect of this on relative time and response allocation between the two components was assessed. It was found that relative time allocation was a direct linear function of relative reinforcement. When greater relative reinforcement was scheduled on one component, relatively more time was allocated to that component. However time allocated to the richer component was not proportionally as great as the reinforcement received on that component, i.e. undermatching was found. This relationship was not adequately described by the matching principle in its simplest form. An even greater degree of undermatching was found when relative response rates were considered as a function of relative reinforcement. The variable-interval and the variable-time components were equally preferred in terms of time allocation. In terms of response allocation two children showed a fairly strong preference for the variable-interval component, and one child showed an almost complete preference for that component. The other subjects showed a slight preference toward responding during the variable-time components. Both relative and absolute local response rates were found to be negative functions of relative reinforcement rates. Using only post-changeover-delay response and time measures did not particularly improve matching. This experiment indicated that response dependent and response independent reinforcement are equally effective in producing time allocation matching. The results of this experiment were compared to the results of a number of other experiments that also used human subjects responding on concurrent schedules. It was concluded that the degree of matching characteristic to human responding is more variable than that found in lower organisms.