ABA Fundamentals

Preference for unsegmented interreinforcement intervals in concurrent chains.

Leung et al. (1985) · Journal of the experimental analysis of behavior 1985
★ The Verdict

Pigeons show us that simple, unsegmented schedules beat chained ones every time.

✓ Read this if BCBAs who write chained token or DRL programs for any population.
✗ Skip if Clinicians who only use immediate, unbroken reinforcement.

01Research in Context

01

What this study did

The team used pigeons in a lab chamber with two keys.

Each key led to a different schedule of food delivery.

One schedule stayed simple. The other was chopped into chained segments.

Birds could hop between keys and show which style they liked.

02

What they found

The birds almost always picked the simple, unbroken schedule.

They liked it even more when the wait time was long or when a short delay blocked quick key changes.

Breaking the interval into linked pieces made the same food feel less attractive.

03

How this fits with other research

Warren et al. (1986) ran the same comparison and saw the same tilt toward simple schedules, so the finding is solid.

Reed et al. (1988) later asked how big the chunks had to be. They showed that smaller chunks (earlier food) widen the gap, proving the 1985 result holds across different split sizes.

Alba et al. (1972) once hinted that chaining lowers choice. The 1985 study gives that idea a clear test and a clean yes.

GOLLUMIGLER (1964) saw birds pause more inside chained links. The new data say pauses are not just noise; they make the whole schedule less popular.

04

Why it matters

For you, the lesson is simple: fewer hoops, more buy-in.

When you design token boards, response chains, or wait-time schedules, keep the path to reinforcement short and smooth.

Drop extra steps or blend them together. Your client will work harder for the same payoff.

Free CEUs

Want CEUs on This Topic?

The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.

Join Free →
→ Action — try this Monday

Cut one middle step from a token chain and see if responding speeds up.

02At a glance

Intervention
not applicable
Design
single case other
Sample size
5
Population
other
Finding
positive

03Original abstract

Five pigeons were trained under concurrent-chain schedules in which a pair of independent, concurrent variable-interval 60-s schedules were presented in the initial link and either both variable-interval or both fixed-interval schedules were presented in the terminal link. Except for the baseline, one of the terminal-link schedules was always a two-component chained schedule and the other was either a simple or a tandem schedule of equal mean interreinforcement interval. The values of the fixed-interval schedules were either 15 s or 60 s; that of the variable-interval schedules was always 60 s. A 1.5-s changeover delay operated during the initial link in some conditions. The pigeons preferred a simple or a tandem schedule to a chain. For the fixed-interval schedules, this preference was greater when the fixed interval was 60 s than when it was 15 s. For the variable-interval schedules, the preferences were less pronounced and occurred only when the changeover delay was in effect. For a given type of schedule and interreinforcement interval, similar preferences were obtained whether the nonchained schedule was a tandem or simple schedule. The changeover delay generally inflated preference and lowered the changeover rate, especially when the terminal-link schedules were either short (15 s) or aperiodic (variable-interval). The results were consistent with the notion that segmenting the interreinforcement interval of a schedule into a chain lowers the preference for it.

Journal of the experimental analysis of behavior, 1985 · doi:10.1901/jeab.1985.44-89