A re-examination of local contrast in multiple schedules.
Pigeons surge late in a rich component when extinction is next—proof that contrast can hide inside one schedule piece.
01Research in Context
What this study did
Green et al. (1975) watched pigeons peck a key during a two-part schedule. Part one paid food on a VI 3-min. Part two paid nothing (extinction). The parts swapped every 60 s.
The team counted every peck in 5-s bins. They wanted to know if birds speed up right before the switch to extinction.
What they found
Birds almost doubled their speed in the final 20 s of the VI part. This late surge happened every session and every bird.
Old theories said contrast only lifts the whole VI rate. The surge shows contrast can live inside one schedule piece.
How this fits with other research
Van Hemel (1973) saw the same birds run faster right after food, not before the switch. The two papers look opposite, but they timed different spots. Early bursts follow food; late bursts forecast extinction.
Reiss et al. (1982) later showed short VI parts can fake a rate jump if they hold richer pay. Our 1975 surge stayed even when pay and length stayed flat, so the jump is not just a duration trick.
Davol et al. (1977) used the same fine-grain method to shrink negative contrast by paying short pauses. Together the trio proves zooming in on moment-to-moment data, not just totals, reveals levers you can pull.
Why it matters
If a client stalls before a known work-break, do not blame motivation—check the upcoming shift in reinforcement. Insert a brief high-pay moment just before the lean component or fade the signal so the transition is less sharp. Measuring responses in seconds, not sessions, lets you spot and fix these mini-contrasts early.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Graph client responses in 10-s bins; if rate spikes just before a lean component, add a quick reinforcer or stretch the rich component longer.
02At a glance
03Original abstract
Pigeons were presented with multiple schedules of alternating 90-sec components. When components in which grain was never presented alternated with components in which grain was presented on a variable-interval schedule, the average rate of responding in the variable-interval components increased, showing overall positive behavioral contrast. Unlike previous reports, this study found that the response rates for all birds increased toward the end of the variable-interval components as training proceeded. This increase in local response rate disappeared when the multiple schedule was composed solely of variable-interval components and reappeared when the variable-interval components were again alternated with extinction. This finding cannot be predicted or explained by recent theories of behavioral contrast based on autoshaping, and thus questions their sufficiency. We suggest that this local response-rate increase results from the predictable change from high to low density of reinforcement at the end of the fixed-duration component. Thus, the present effect apparently illustrates a different type of interaction between components of a multiple schedule than that described by previous theories of contrast. In a given procedure, either or both types of interaction may occur; neither provides a complete account of behavioral contrast.
Journal of the experimental analysis of behavior, 1975 · doi:10.1901/jeab.1975.24-291