Webbe and Malagodi (1978) published an early article that compared fixed-ratio and
variable-ratio exchange schedules of token reinforcement. In their experiment, rats were trained
to press a lever and received a marble for every 20 responses (FR 20: TOKEN). During
baseline, the rats were provided an opportunity to exchange their marbles after six completions
of the schedule (FR 6: EXCHANGE) and each marble produced a food pellet (FR 1: FOOD).
After the response rate was stable, the exchange schedule was changed from a fixed-ratio
SCHEDULES OF TOKEN REINFORCEMENT 10
schedule to a variable-ratio schedule (VR 6: EXCHANGE), and the number of marbles delivered
until reinforcement varied from 1-14. A reversal to FR 6: EXCHANGE was implemented after
response rate was stable. There were higher rates of responding under the VR condition as
compared to the FR condition. There was also less variability in rate of responding between
sessions under the VR schedule. Finally, post-reinforcement pausing after token delivery was
much shorter in the VR condition as compared to the FR condition, and pausing in the second FR
condition was much longer than the first FR condition, suggesting that exposure to a VR
schedule increased pausing in the FR schedule.