Behavioral contrast without response-rate reduction.
Behavioral contrast can emerge even when the alternate schedule almost never gets a response.
01Research in Context
What this study did
Researchers worked with pigeons on a two-key setup. Birds pecked one key for food on a steady schedule.
The team then added a second key. In some rounds that key never gave food, yet birds hardly pecked it.
They watched if pecks on the first key still jumped up, even though the new key caused almost no drop in overall responses.
What they found
Contrast still showed up. When the no-food key appeared, rates on the food key rose, even though birds barely touched the empty key.
The jump happened without any big dip in total responses, proving contrast does not need earlier suppression.
How this fits with other research
Kodera et al. (1976) later copied the effect. They used slow stimulus steps and still saw contrast, backing the idea that error-free shifts can produce the same bounce.
Blue et al. (1971) had said contrast needs an extinction part. This study agrees, but adds a twist: almost zero pecks in that part are enough.
Rodewald (1974), Mulvaney et al. (1974), and Terrace (1974) all worked the same year. They split pecks into elicited, auto, and active non-pecks. Together they show contrast is built from many small response classes, not just a single rate drop.
Why it matters
You can stop worrying that clients must first fail or stop responding before gains appear elsewhere. When you thin reinforcement or introduce an S-delta, watch for jumps in the still-reinforced skill even if errors stay low. Use slow stimulus steps and you may still get contrast, so plan for it rather than surprise.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Track the target skill rate while you introduce a new no-reinforcement task; expect a bounce even if the new task gets zero responses.
02At a glance
03Original abstract
Behavioral contrast was obtained in two experiments, which both employed a standard free-operant successive discrimination (a multiple variable-interval extinction schedule), without the occurrence of reductions of response rate in the extinction component. In Experiment I, one group of four pigeons was trained on a multiple schedule in which one stimulus was associated with a variable-interval schedule and the second stimulus with response-independent reinforcement on a free variable-interval schedule. Though by the end of this training three pigeons were responding very little to the second stimulus, when this stimulus was associated with extinction all subjects showed a contrast effect. In Experiment II, eight pigeons were trained extensively to respond to a single stimulus on a variable-interval schedule, before a second stimulus associated with extinction was introduced. This second stimulus was dissimilar to the initial stimulus and five pigeons never responded in its presence. Nevertheless, all pigeons showed a contrast effect and there was no evidence that the effect was smaller in errorless subjects or smaller than in a subsequent discrimination where all subjects made many errors. Both experiments indicated that response reduction in one component of a multiple schedule is not a necessary condition for the occurrence of contrast.
Journal of the experimental analysis of behavior, 1974 · doi:10.1901/jeab.1974.22-453