Control of responding by stimulus duration.
Differential reinforcement can create stimulus control by duration, but maintaining that control requires careful programming.
01Research in Context
What this study did
Dardano (1971) asked a simple question. Can we make an animal respond only when a light stays on for a set time?
Pigeons pecked a key. Short flashes paid nothing. One long flash paid food. The birds had to learn the cutoff.
The team then ran probe tests. They wanted to see if the birds would keep the rule when pay changed.
What they found
The schedule worked. Birds pecked fast after long flashes and rarely after short ones.
When extra tests kept coming, the tidy difference started to blur. Control slid without new teaching.
How this fits with other research
Harrison et al. (1975) picked up the same birds four years later. They showed the fix: let the response turn the light on and start pay. That kept the timing rule alive.
Cicerone (1976) used a similar logic with DRL. Only long pauses paid off. Again, short pauses stayed wild. The two studies line up—timing rules work, but only the reinforced slice stays clean.
Rapport et al. (1982) seemed to clash. They found pay rate did not change how well pigeons saw duration. The trick is they never used differential pay. F used clear on/off payoff, D kept pay equal. Same birds, different question—no real fight.
Why it matters
You can build duration control fast with differential pay, but you must plan for upkeep. Tie the target length to the start of reinforcement, not just to its presence. If you probe too long without new teaching, the rule drifts. Check Harrison et al. (1975) for a ready maintenance plan.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Program the long-duration response to turn on the stimulus and start reinforcement—do not rely on simple presence of the cue
02At a glance
03Original abstract
Pigeons were trained on a procedure in which the key was white for 30 sec, alternating with periods of darkness, or timeout. In a nondifferential training procedure, timeout duration was held constant at either 9 or 21 sec for different animals, and pecks on the white key were reinforced on a variable-interval 36-sec schedule. After 30 sessions an extinction generalization test was conducted where the duration of the timeout was varied from 3 to 27 sec. This test showed no differences in responding following timeouts of different durations. In a differential training procedure, timeout durations of either 9 or 21 sec were randomly scheduled for each animal. The variable-internal schedule was in effect following the same timeout duration as in the prior nondifferential procedure. No pecks were reinforced after the other timeout duration. In 40 sessions, differences in response rates following the two durations gradually developed. A maintained generalization procedure was then imposed in which timeout durations were varied from 3 to 27 sec, with the variable-interval schedule in effect following only the same duration as in the previous procedures. The first maintained generalization session showed that the prior differential training had established control of the animals' behavior by the timeout duration. In continued training on the maintained generalization procedure, control by the timeout duration decreased.
Journal of the experimental analysis of behavior, 1971 · doi:10.1901/jeab.1971.16-81