Molar optimization versus delayed reinforcement as explanations of choice between fixed-ratio and progressive-ratio schedules.
Choice is driven by delays across upcoming trials, not by the average payoff of the current task alone.
01Research in Context
What this study did
Researchers let pigeons pick between two keys. One key always needed the same number of pecks. The other key needed more and more pecks each time.
The birds could switch keys at any moment. The team watched which key the bird chose next and how long it waited.
What they found
Birds cared less about the work on this trial. They cared more about how soon food would come across many trials.
If choosing the fixed key made later food arrive faster, birds picked it. Simple averages could not explain this.
How this fits with other research
Hassin-Herman et al. (1992) later added a reset rule. When the bird picked the fixed key, the rising count on the other key dropped back to one. Birds switched even sooner, showing they track future work across trials.
Quilitch et al. (1973) said choice follows delay on this trial only. The new data do not erase that rule; they stretch it. The 1987 study shows delays on future trials matter too.
Davis et al. (1994) ran a similar test with variable work. They also found that timing, not just rate, guides choice. All three papers point the same way: organisms think ahead more than old molar models said.
Why it matters
When you build a token board, DR schedule, or work-break plan, think about what the client gains on the next three tasks, not just this one. A child may accept five math problems now if it means a quick game follows. If the game keeps getting delayed, problem behavior may rise even though the single-task payoff looks the same. Build chains so the smarter, longer path still feels short.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Plot the next three reinforcer delays before you raise the response count; keep that total wait under the client’s known tolerance.
02At a glance
03Original abstract
In a discrete-trials procedure, pigeons chose between a fixed-ratio 81 schedule and a progressive-ratio schedule by making a single peck at the key correlated with one or the other of these schedules. The response requirement on the progressive-ratio schedule began at 1 and increased by 10 each time the progressive-ratio schedule was chosen. Each time the fixed-ratio schedule was chosen, the requirement on the progressive-ratio schedule was reset to 1 response. In conditions where there was no intertrial interval, subjects chose the progressive-ratio schedule for an average of about five consecutive trials (during which the response requirement increased to 41), and then chose the fixed-ratio schedule. This ratio was larger than that predicted by an optimality analysis that assumes that subjects respond in a pattern that minimizes the response-reinforcer ratio or one that assumes that subjects respond in a pattern that maximizes the overall rate of reinforcement. In conditions with a 25-s or 50-s intertrial interval, subjects chose the progressive-ratio schedule for an average of about eight consecutive trials before choosing the fixed-ratio schedule. This change in performance with the addition of an intertrial interval was also not predicted by an optimality analysis. On the other hand, the results were consistent with the theory that choice is determined by the delays to the reinforcers delivered on the present trial and on subsequent trials.
Journal of the experimental analysis of behavior, 1987 · doi:10.1901/jeab.1987.48-251