College students' responding to and rating of contingency relations: The role of temporal contiguity.
Reinforcers and punishers must arrive within seconds, not minutes, to shape human behavior.
01Research in Context
What this study did
College students pressed a telegraph key. Each press could earn money or lose money.
The researchers changed only one thing: how fast the outcome arrived. Sometimes the money came right away. Sometimes it came after a long delay.
Students also rated how much they felt their pressing caused the outcome.
What they found
When the outcome arrived right after the press, students pressed more and said "I caused it."
When the outcome was delayed, pressing dropped even if the chance of money stayed the same.
This happened for both winning and losing money. Time, not probability, ruled both behavior and belief.
How this fits with other research
Barnard et al. (1977) saw the same drop in pigeons when food was delayed more than four seconds. The 1986 study shows humans follow the same rule.
Van Hanegem et al. (2014) moved the idea into therapy rooms. They found that escape from tasks works better when given right after compliance, not on a fixed timer.
Pritchard et al. (1987) seemed to disagree. They said contingency still matters even with delays. The key difference: their longest delay was only 8 s, still inside the "immediacy window." The 1986 study used much longer gaps, showing the real cutoff is tighter than we thought.
Why it matters
If you want a behavior to stick, deliver the reinforcer or punisher within seconds. A long delay, even with perfect correlation, kills the effect. Check your DTT or token board timing today.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Set a 3-second timer on your phone and deliver every token or praise within that window.
02At a glance
03Original abstract
Two experiments investigated the role of temporal contiguity in college students' responding to and rating of contingency relations during operant conditioning. Schedules were devised that determined when but not whether appetitive or aversive events would occur. Subjects' reports concerning the schedules were obtained by means of a 200-point rating scale, anchored by the phrases "prevents the light from occurring" (-100) and "causes the light to occur" (+100). When tapping a telegraph key advanced the time of point gain, responding was maintained or increased and subjects gave positive ratings. When tapping a telegraph key advanced the time of point loss, subjects also gave positive ratings, but responding now decreased. When key tapping delayed the time of point gain, responding decreased and subjects gave negative ratings. When key tapping delayed the time of point loss, subjects also gave negative ratings, but responding now increased. These findings implicate response-outcome contiguity as an important contributor to causal perception and to reinforcement and punishment effects. Other accounts-such as those stressing the local probabilistic relation between response and outcome or the molar correlation between response rate and outcome rate-were seen to be less preferred interpretations of these and other results.
Journal of the experimental analysis of behavior, 1986 · doi:10.1901/jeab.1986.46-15