Choice as a dependent measure in autoshaping: sensitivity to frequency and duration of food presentation.
Side-by-side choice reveals reinforcer value that rate and latency hide.
01Research in Context
What this study did
The team worked with pigeons in a lab.
Two colored keys lit up at the same time.
Pecks on one key sometimes led to food.
The other key gave food more or less often, or for shorter or longer times.
They tracked which key the birds picked.
They also counted how fast and how often the birds pecked.
What they found
Choice told the whole story.
Birds shifted their picks when the odds or the food time changed.
Old measures like peck rate and reaction time stayed flat.
Only the choice data showed the reinforcer differences.
How this fits with other research
Cicerone (1976) had already shown that birds peck more when food odds rise.
But that study looked at single keys shown one at a time.
Skrtic et al. (1982) proves that choice, not peck count, is the sharper tool when both keys are on together.
Iwata (1993) later kept the same two-key setup and asked if birds care about the overall odds or just the last few rewards.
They found birds use the big picture when both choices are in view.
Pilgrim et al. (2000) then asked if longer food times also make choices stickier under stress.
They saw that constant-length rewards were picked more and held up better when food was given early.
All three later papers keep the core method but add new questions, showing the 1982 finding is solid and stretchy.
Why it matters
If you want to know what a client truly values, give them two clear choices at once.
Watch where they go, not how fast they move.
This works for picking AAC devices, leisure items, or work tasks.
A simple side-by-side test beats guessing every time.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Place two potential reinforcers in front of your client at the same time and count which one they pick first across ten trials.
02At a glance
03Original abstract
Previous investigations have shown that rate, latency, and percentage of trials with at least one response are somewhat insensitive measures of the strength of autoshaped responding. In the present studies, these measures were contrasted with the allocation of responding during simultaneous choice tests, a measure of response strength frequently used in operant paradigms. In two experiments, nine pigeons were exposed to a forward pairing autoshaping procedure. Training sessions consisted of the successive presentation of three stimuli, each followed by food on either 100%, 50%, or 0% of the trials. Choice testing involved the simultaneous presentation of the three stimuli. In Experiment I, all pigeons consistently directed their initial choice responses and the majority of subsequent responses to the stimulus always followed by food, despite the fact that during training sessions the response rates of most birds were highest in the presence of the stimulus followed by food on 50% of the trials. In Experiment II, rate, latency, and percentage of trials with at least one response did not change appreciably as a function of duration of feeder presentations. However, choice responding was lawfully affected by duration of feeder presentations. These data suggest that choice is perhaps a more sensitive measure of the strength of autoshaped responding than other, more commonly employed, indices.
Journal of the experimental analysis of behavior, 1982 · doi:10.1901/jeab.1982.37-393