Key pecking of pigeons under variable-interval schedules of briefly signaled delayed reinforcement: effects of variable-interval value.
Keep delayed reinforcement short relative to the schedule’s natural wait time or you will lose response speed.
01Research in Context
What this study did
Scientists watched pigeons peck a colored key. The birds earned grain on a variable-interval schedule.
Sometimes the grain came right away. Other times a short light flashed first and the grain arrived 5 or 20 seconds later.
The team changed how long the birds waited between meals. They wanted to see if the signaled delay still helped when the schedule itself was fast or slow.
What they found
Brief signaled delays kept the birds pecking fast. Longer signaled delays slowed the birds down.
The same 20-second delay hurt most when the schedule already made the birds wait 20 seconds between meals. The delay effect is relative, not absolute.
How this fits with other research
Schaal et al. (1990) showed that a longer signal itself boosts pecking. The new study adds that the schedule speed sets how long is "too long."
Hamm et al. (1978) used no signal and saw rates fall as delays grew. Adding a brief signal flips the early part of that curve upward.
Guest et al. (2013) later tested brief delays under ratio schedules and found no boost. Together the papers warn: the payoff schedule type decides whether a short delay helps or harms.
Why it matters
When you stretch reinforcement timing, pair the delay with a quick bridge or token and keep the delay shorter than the usual wait time. On dense schedules you can go a little longer; on lean schedules keep it tight. Check the learner’s pace—what feels "brief" for one program may already be too long for another.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Time your current inter-reinforcement interval, then set any delay signal to half that length or less.
02At a glance
03Original abstract
Key pecking of 4 pigeons was maintained under a multiple variable-interval 20-s variable-interval 120-s schedule of food reinforcement. When rates of key pecking were stable, a 5-s unsignaled, nonresetting delay to reinforcement separated the first peck after an interval elapsed from reinforcement in both components. Rates of pecking decreased substantially in both components. When rates were stable, the situation was changed such that the peck that began the 5-s delay also changed the color of the keylight for 0.5 s (i.e., the delay was briefly signaled). Rates increased to near-immediate reinforcement levels. In subsequent conditions, delays of 10 and 20 s, still briefly signaled, were tested. Although rates of key pecking during the component with the variable-interval 120-s schedule did not change appreciably across conditions, rates during the variable-interval 20-s component decreased greatly in 1 pigeon at the 10-s delay and decreased in all pigeons at the 20-s delay. In a control condition, the variable-interval 20-s schedule with 20-s delays was changed to a variable-interval 35-s schedule with 5-s delays, thus equating nominal rates of reinforcement. Rates of pecking increased to baseline levels. Rates of pecking, then, depended on the value of the briefly signaled delay relative to the programmed interfood times, rather than on the absolute delay value. These results are discussed in terms of similar findings in the literature on conditioned reinforcement, delayed matching to sample, and classical conditioning.
Journal of the experimental analysis of behavior, 1992 · doi:10.1901/jeab.1992.58-277