Effects of percentage of reinforcement and number of reinforcements in S+ on discrimination learning in the runway

Abstract

Williams (1989) reported that discrimination learning and its reversal in the operant chamber are regulated by the number of reinforcements to the S+ alternative independently of reinforcement schedule, a finding Williams called the invariance effect. To determine whether there is an invariance effect in the runway, we trained four groups of rats on a successive brightness discrimination problem and its reversal. The experimental design was a factorial combination of two reinforcement schedules associated with S+ (partial reinforcement, PRF, vs. consistent reinforcement, CRF) and the number of trials per day to S+ (five vs. seven). The design provided one PRF group and one CRF group given the same number of reinforcements to S+. It also provided for a comparison between two PRF groups given different percentages of reinforcement in S+, and for PRF and CRF groups given the same number of trials in S+. PRF impaired both original discrimination learning and reversal learning whether PRF and CRF groups were equated for number of trials or for number of reinforcements, and learning was not affected by the percentage of reinforcement difference between the two PRF groups. Thus, no invariance effect was obtained.