Multiple and concurrent schedule performance: independence from concurrent and successive schedule contexts.
Multiple schedules produce stronger undermatching than concurrent ones, and the pattern sticks no matter what schedule came first.
01Research in Context
What this study did
Researchers placed pigeons in a chamber with two keys.
Birds earned grain for pecking on two different variable-interval schedules.
The team compared two set-ups: multiple schedules (one key at a time) and concurrent schedules (both keys present).
They also switched the order of sessions to see if context changed how birds divided their pecks.
What they found
Undermatching was stronger on multiple schedules than on concurrent schedules.
When birds switched from concurrent to multiple, the way they split time between keys stayed the same.
Context did not matter: allocation patterns were set by the schedule type, not by what came before.
How this fits with other research
Tracey et al. (1974) warned that short components in multiple schedules can fake matching.
The new data agree: undermatching, not perfect matching, is the norm on multiple schedules.
Jarrold et al. (1994) later saw the same undermatching with kids doing math for nickels, showing the pattern holds outside the lab.
Bruder (1986) extended the finding to four-peck sequences, proving the rule works even when the response is complex.
Why it matters
When you use multiple schedules in teaching or treatment, expect learners to undermatch reinforcement rates.
Do not assume perfect matching; check actual time or response splits and adjust schedule values accordingly.
If you switch between concurrent and multiple arrangements, know that the learner’s allocation rule stays stable, so you can plan transitions without extra warm-up.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Track exact time or responses on each alternative; if splits drift below the reinforcement ratio, tighten the richer schedule or add brief changeover delays.
02At a glance
03Original abstract
Six pigeons were trained on multiple variable-interval schedules and performance was measured in the presence or absence of another variable-interval schedule (the common schedule) arranged concurrently with both components. Manipulations included varying the rate of reinforcement on the common schedule, leaving the common schedule unchanged while the components of the multiple schedule were varied, varying the multiple schedule components in the absence of the common schedule, and varying one component of the multiple schedule while the other component and the common schedule were unchanged. The normal rate-increasing and rate-decreasing effects of reinforcement rate increase were found, except that changing one multiple schedule component did not affect the response rate in the successively available common schedule component. Both concurrent and multiple schedule performance undermatched obtained reinforcement-rate ratios, but the degree of undermatching in multiple schedules was reliably greater. Allocation of responses between multiple schedule components was unaffected by the concurrent availability of reinforcement, and allocation of responses between concurrent schedules was unaffected by the successive availability of different reinforcement rates.
Journal of the experimental analysis of behavior, 1977 · doi:10.1901/jeab.1977.28-27