On the functions of the changeover delay.
Changeover delay style tweaks local hopping but leaves the matching relation intact.
01Research in Context
What this study did
The team ran rats on two levers at once. Each lever paid off on its own clock.
A changeover delay (COD) made the animal wait after it switched sides. They tried two COD styles.
One style reset the wait clock on every switch. The other style let the clock run once per visit.
What they found
Both COD styles kept the big picture the same. The animals still matched their time to the payoff rates.
But the small stuff changed. Reset clocks made more hops back and forth. Run-once clocks made longer stays.
When total time on each side ended up equal, the hop rate looked the same no matter how they got there.
How this fits with other research
Nakamura et al. (1986) showed the matching law works across many set-ups. This study says the law still holds even when you tweak the switch rule.
Rider (1983) found micro-patterns inside single payoff periods. Here, micro-patterns show up between two payoff streams.
Killeen (2015) argues matching is just logistic regression driven by the Law of Effect. The steady matching seen here fits that view; only the local dance changes, not the final choice curve.
Why it matters
If you run concurrent schedules in a classroom or clinic, the COD style you pick will shape how often the client flips between tasks. Pick reset CODs when you want quick, frequent checks. Pick run-once CODs when you want longer, deeper engagement. Either way, the overall time balance should still line up with the payoff rates, so you can trust the matching law to guide your data check.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Pick one COD style and stick with it for a week; graph hops per minute to see the local effect.
02At a glance
03Original abstract
The function of changeover delays in producing matching was examined with pigeons responding on concurrent variable‐interval variable‐interval schedules. In Experiment 1, no changeover delay was compared to two different types of changeover delay. One type, designated generically as response‐response but in the present example as peck‐peck, was timed from the first response on the switched‐to key; the other, designated generically as pause‐response but in the present example as pause‐peck, was timed from the last response on the switched‐from key. High changeover rates occurred with no changeover delay. Peck‐peck and pause‐peck changeover delays produced low and intermediate changeover rates, respectively. In Experiment 2, pause‐peck and peck‐peck changeover delays were compared across a range of relative reinforcement rates. Similar matching relations developed despite differences in the changeover rates and local response patterns as a function of the type of changeover delay. In Experiment 3, both types of changeover delay yielded similar changeover rates when their obtained durations were equal via yoking. The results suggest that changeover delays function to separate responses on one key from reinforcers on the other or to delay reinforcement for changing over. In addition, the distribution of responding during and after the changeover delay may vary considerably without affecting matching.
Journal of the experimental analysis of behavior, 1998 · doi:10.1901/jeab.1998.69-141