Free-operant choice behavior: A molecular analysis.
A one-second blackout after switching keys stops useless hopping and gives cleaner choice data than the usual changeover delay.
01Research in Context
What this study did
Pigeons pecked two keys in a lab. Each key led to a short chain of schedules. A switch from one key to the other could trigger a brief blackout or a changeover delay. The team tracked every peck to see how these small time gaps shaped switching.
The birds worked under concurrent-chain schedules. Researchers added or removed a 1.5-s blackout and a 3-s changeover delay across conditions. They recorded minute-by-minute patterns to spot superstitious switching.
What they found
When only the changeover delay was present, birds switched more, not less. Adding the short blackout cut most of this extra switching. The blackout cleaned up the data and made choice patterns easier to read.
Switching dependencies flipped once the blackout joined the delay. Blackout also trimmed repetition biases that had clouded earlier sessions.
How this fits with other research
Arnett (1972) saw a similar clean-up effect seven years earlier. His pigeons also faced blackouts that cut preference for the key producing them. Both studies show blackout beats plain delay at controlling unwanted switching.
Barrett et al. (1987) later showed the flip side. They inserted silent 3-s gaps between chain links and watched response rates crash. Together the papers warn that any dead time—whether blackout, delay, or inter-link pause—can reshape behavior.
Cohen et al. (1990) extended the same concurrent-chain method to sub-optimal choice. They proved that long terminal links plus signals can reverse preference. The 1979 blackout trick helps keep those links clean so true preference shows through.
Why it matters
If you run concurrent schedules in skill-acquisition or preference assessments, add a brief blackout after every switch. One second is enough. It stops superstitious hopping between options and gives you clearer data on what your learner actually prefers. Try it next time you probe choice for tasks, reinforcers, or break activities.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Program a 1-s blackout after every changeover in your concurrent reinforcer assessment and watch switching drop.
02At a glance
03Original abstract
Pigeons' pecks to two concurrent initial-link stimuli occasionally produced one of two mutually exclusive terminal links. Further responding to the terminal-link stimulus produced food under fixed-interval or fixed-ratio schedules. In such concurrent chained schedules, investigators rarely use a changeover delay to control superstitious switching, although it is customary to use a changeover delay in simple concurrent schedules in which choice responses produce food directly. When terminal-link fixed-interval schedules were equal or similar in duration and no changeover delay was employed, conditional probabilities of choice for a schedule were found to be lower if the last choice was for that schedule than if the last choice was for the other schedule ("switching dependency"). Imposition of a changeover delay with equal or unequal terminal links produced the opposite pattern: conditional probabilities of choice for a schedule were higher if the last choice was for that schedule than if the last choice was for the other schedule. Turning off all chamber lights during the changeover delay interval attenuated these "repetition dependencies." The results indicate that excessive switching can complicate the interpretation of data from concurrent chains much as from simple concurrent schedules, and that using blackouts to control switching may be preferable to using a changeover delay.
Journal of the experimental analysis of behavior, 1979 · doi:10.1901/jeab.1979.32-213