Choice behavior of rats in a concurrent-chains schedule: Amount and delay of reinforcement.
Equal longer delays boost preference for larger reinforcers in rats, following a predictable power function.
01Research in Context
What this study did
The team put rats in a two-lever cage. Pressing one lever gave one food pellet right away. Pressing the other gave three pellets after a delay.
Every condition kept the delay ratio the same. Only the absolute length of both delays grew. The researchers watched how often the rats picked the bigger snack.
What they found
When both delays got longer together, the rats chose the three-pellet side more often. Their preference followed a neat power curve.
The result backed up Fantino’s delay-reduction model and gave cleaner numbers for it.
How this fits with other research
Two years earlier Sanford et al. (1980) saw mixed results with fixed ratios. The new study shows the model works once you stretch both delays equally.
Higgins et al. (1992) later added ‘income.’ They cut the total trials and saw rats become more patient. That finding builds on the 1982 curve by showing context matters.
Haynes et al. (2022) ran a fresh rat study. They used rising delays within one session and still saw the switch to smaller-sooner rewards. Their data conceptually replicate the 1982 curve with a new twist: practice trimmed the flip-flops.
Ghaziuddin et al. (1996) mapped the exact indifference point. Six pellets lose to one pellet once the delay hits about 55 s. That successor paper quantifies the trade-off the 1982 study first refined.
Why it matters
You now have a rule of thumb: longer equal waits push learners toward bigger later rewards. Use this when you design token boards or chained schedules. Start with short waits, then stretch both sides together to keep the learner choosing the richer option.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Add two seconds to both the small and large reinforcer delays in your token board and watch if the learner still picks the bigger stack.
02At a glance
03Original abstract
Rats were exposed to concurrent-chains schedules in which the terminal links were equal fixed-interval schedules terminating in one or three food pellets. Choice proportions for large reward increased with increases in delay intervals programmed on fixed-interval schedules and supported the predictions derived from a general choice model originally formulated by Fantino and later developed by Navarick and Fantino. In addition, a functional equivalence of two alternatives was established by increasing delay intervals with large reward, whereas delay intervals for small reward were held constant. Functionally equivalent delay intervals with large reward, for each delay interval with small reward, can be described by a power function with exponent smaller than 1.0. A better prediction of choice proportions resulted when this function was used to derive predicted choice proportions.
Journal of the experimental analysis of behavior, 1982 · doi:10.1901/jeab.1982.37-383