Multiple schedule performance

Abstract

The keypecking of four pigeons was reinforced in each of six multiple schedules, and two measures of performance were obtained: time allocated to each component and number of keypecks in each of five successive sub-intervals of each component. Behavioural contrast of overall response rate was found, but when contrast was assessed separately in each sub-interval there appeared to be no change in magnitude of (local) contrast as a function of time since component change. Ratios of response rates in initial sub-intervals were highly sensitive to ratios of reinforcement rates, but sensitivity declined sharply over the first 50 seconds and then remained approximately constant to the end of the component. Overall time measures showed behavioural contrast, and ratios of time measures showed moderate sensitivity to ratios of reinforcement rates. Ratios of times and peckrates were found to be highly correlated. These results confirm the importance of time since component change as a determinant of sensitivity of response ratios in multiple schedules, but suggest that behavioural contrast and matching are not related in that they do not covary within components. The findings concerning time allocation are consistent with an account of behavioural contrast of absolute response rate in terms of differential time allocation, and also suggest that undermatching of response to reinforcement ratios in multiple schedules may partially result from the constraint on variance of time ratios imposed by experimenter-determined component alternation.