Maximizing present value: A model to explain why moderate response rates obtain on variable-interval schedules.
Holding every reinforcer until the end drains VI response speed because quick inter-response times lose their timing advantage.
01Research in Context
What this study did
The team used pigeons on a variable-interval food schedule.
They paid every response after the interval, but held all food until the session ended.
Birds could peck fast or slow; the delay stayed the same no matter what.
What they found
When food was moved to the very end, the birds slowed their pecking on VI.
The same delay trick did not hurt rates on a ratio schedule.
Short, quick responses lost their payoff edge, so birds quit rushing.
How this fits with other research
Dukhayyil et al. (1973) saw the same rule with kids: choice plus quick payoff kept them in the game longer.
Shimp et al. (1974) ran the opposite move—adding a signal that food was coming—and birds flocked to that side.
Together the three papers show the same VI schedule can look tasty or dull, depending on how soon or how clearly the reinforcer shows up.
Why it matters
Your client’s “payday” may be too far away. If tokens, praise, or breaks all come at 3 p.m., you have built a session-end VI. Expect slower, scattered work until the clock strikes three. Instead, deliver small payoffs right as the target responses happen. You will keep the brief, efficient bursts that VI schedules are meant to maintain.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Break the end-of-session lump: give one token, point, or bite of snack within three seconds of the target response.
02At a glance
03Original abstract
In Phases 1 and 3, two Japanese monkeys responded on a multiple variable-ratio 80 variable-interval X schedule, where the value of X was adjusted to ensure equal between-schedule reinforcement rates. Components strictly alternated following the delivery of a food pellet, and each session ended following 50 components. Phase 2 differed from the others only in that the 50 pellets previously earned during the session were delivered together at session's end. Variable-ratio response rates did not decrease across phases, but variable-interval response rates decreased substantially during the Phase 2 procedure. This rate decrease was attributed to the food-at-session's-end manipulation removing the greater immediacy of reinforcement provided by short interresponse times relative to long interresponse times. Without this time preference for short interresponse times, the variable-interval interresponse-time reinforcement feedback function largely controlled response emission, dictating a response-rate reduction. This result was explained in terms of the economic notion of "maximizing present value."
Journal of the experimental analysis of behavior, 1988 · doi:10.1901/jeab.1988.49-331