Positive conditioned suppression: Transfer of performance between contingent and noncontingent reinforcement situations.
Free reinforcement after contingent work can suppress responding and the suppression sticks around.
01Research in Context
What this study did
Researchers worked with pigeons pecking two keys. One key gave food only after a set number of pecks. The other key gave food for free at random times.
They switched the birds back and forth between these two keys. They wanted to see if the birds' speed of pecking changed when the rules changed.
What they found
When the contingent key paid off, the birds pecked faster. When they moved to the free-food key, most birds slowed down. The slow-down lasted longer than the speed-up.
A few birds kept pecking fast on the free key, but most did not. The drop in speed was the stronger and longer-lasting effect.
How this fits with other research
Evans (1963) showed that more frequent free food weakens suppression. Locurto et al. (1980) now shows that even a little free food can suppress responding if the bird first worked for it. The two ideas click: dense free food protects performance, sparse free food after work hurts it.
Rogers-Warren et al. (1976) saw that extra free food flattens the scalloped FI pattern. Locurto et al. (1980) adds that the same extra food can also drop the overall rate, especially right after a work requirement. Together they warn that non-contingent rewards can quietly cut responding.
Davol et al. (1977) found negative contrast when a rich schedule turns lean. Locurto et al. (1980) shows a cousin effect: after earning every reinforcer, even neutral free food feels lean and the bird slows down. Both papers tell us that past payoff rules shadow future performance.
Why it matters
If you run a DRO, token board, or any mixed schedule, watch what happens when the client suddenly gets free tokens or social praise. The study says responding may dip and stay low longer than you expect. Track the dip and be ready to re-prompt or thin the free reinforcement quickly so the skill does not stall.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Count responses for five minutes after you hand a free token; if rate drops more than a large share, return to contingent reinforcement immediately.
02At a glance
03Original abstract
Five homing pigeons were trained on concurrent variable-interval schedules. A fixed-duration stimulus was occasionally presented on one key; and, in various conditions, this stimulus terminated (a) without reinforcement, (b) in noncontingent reinforcement, (c) with reinforcement contingent on a response on the key on which the stimulus was presented, and (d) with reinforcement contingent on a response on the key on which the stimulus was not presented. Initially, a stimulus terminating in noncontingent reinforcement generally produced decreased response rates on both keys during the stimulus. Contingencies, however, reliably produced increased rates during the stimulus on the key on which the contingency was arranged, relative to the rate on the concurrently available key. Contingency conditions were followed by noncontingency conditions in which the separation of rates caused by contingencies was maintained. When rates during the stimulus were compared with response rates on the same keys in the absence of the stimulus, contingency-caused rate increases and decreases were again found, but only the rate decreases were maintained in subsequent noncontingency conditions. Further data suggested that the contingency-caused rate changes were not maintained when the stimulus terminated without reinforcement, and that they were unaffected by a threefold decrease in the reinforcement rate provided by the baseline schedules. The results support the suggestion that performance in the positive conditioned suppression procedure results from concurrent and multiple schedule interactions. They further suggest that the production of either acceleration or suppression is dependent on adventitious and historical contingencies.
Journal of the experimental analysis of behavior, 1980 · doi:10.1901/jeab.1980.33-51