Unsignalled delay of reinforcement in variable-interval schedules.
Unsignalled reinforcement delays can either raise or lower response rates depending on delay length.
01Research in Context
What this study did
Hamm et al. (1978) worked with pigeons pecking a key on a variable-interval schedule.
Sometimes the grain came right after the peck. Sometimes it came a few seconds later with no light or sound to mark the wait.
The team wanted to see how these silent delays changed how fast the birds pecked.
What they found
Tiny delays of half a second made the birds peck faster.
Longer delays of several seconds made the birds slow down and then stop.
The same delay could help or hurt depending only on its length.
How this fits with other research
Guest et al. (2013) copied the half-second boost in interval schedules but showed it vanishes in ratio schedules. The lesson: check your schedule type before you bank on a delay.
Schaal et al. (1990) added a short light during the delay and still saw faster pecking. Their signal removed the guess-work yet gave the same lift, so the effect is not just about uncertainty.
Wolchik et al. (1982) zoomed in and found the half-second delay fills the data with rapid bursts. The birds did not peck more often by chance; they switched to machine-gun style.
Why it matters
If you use brief delays to stretch reinforcement budgets, keep them tiny—under a second—and stay on interval schedules. Watch for burst responding; speed is not always skill. Long or unknown delays will punish the very behavior you want.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →If you must delay a reinforcer, cap the wait at half a second and count bursts for the first five minutes.
02At a glance
03Original abstract
Three pigeons responded on several tandem variable-interval fixed-time schedules in which the value of the fixed-time component was varied to assess the effects of different unsignalled delays of reinforcement. Actual (obtained) delays between the last key peck in an interval and reinforcement were consistently shorter than the nominal (programmed) delay. When nominal delays were relatively short, response rates were higher during the delay condition than during the corresponding nondelay condition. At longer nominal delay intervals, response rates decreased monotonically with increasing delays. The results were consistent with those obtained from delay-of-reinforcement procedures that impose either a stimulus change (signal) or a no-response requirement during the delay interval.
Journal of the experimental analysis of behavior, 1978 · doi:10.1901/jeab.1978.30-169