Transfer of oddity-from-sample performance in pigeons.
Oddity training creates true concept learning that transfers to never-before-seen items.
01Research in Context
What this study did
Carroll (1977) worked with pigeons in a three-key chamber.
The birds learned to pick the one color that did NOT match the sample.
After training, new colors appeared to see if the birds really understood "odd."
What they found
The pigeons picked the odd new color right away.
This showed they had learned a concept, not just memorized old colors.
How this fits with other research
Schmick et al. (2018) later used the same idea to teach autistic teens to name emotions.
LaFond et al. (2021) used it to teach preschoolers their parent’s phone number.
Both studies kept the core method: train a relation, then test with brand-new items.
Why it matters
If a pigeon can learn an oddity rule and use it on new colors, your learner can too.
Use brief equivalence training, then probe with totally fresh stimuli.
You will see right away if the child owns the concept or just the toy you used.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →After teaching same/different or emotion matching, swap in new pictures and see if the learner still picks the odd one out without extra teaching.
02At a glance
03Original abstract
Four pigeons were trained on a modified three-key oddity-from-sample task in which an observing response to the sample (center-key) stimulus lighted a single comparison (side-key) stimulus. If the comparison stimulus was different from the sample stimulus, a single peck to the lighted comparison was reinforced. If the comparison and sample stimuli were identical, the pigeons had to refrain from pecking the comparison for 4.6 seconds to terminate the matching comparison and to produce immediately a nonmatching comparison on the remaining side key. Each peck to the matching comparison reset the 4.6-second delay interval. Three hues were used during acquisition. During tests for transfer of the oddity performance, two novel hues were substituted either individually or together for one or two of the original training hues. For three birds, latencies to novel nonmatching hues were identical to baseline nonmatching latencies. Latencies to novel matching hues were shorter than baseline matching latencies but were consistently longer than novel nonmatching latencies. These transfer data demonstrate that the pigeons learned the oddity concept.
Journal of the experimental analysis of behavior, 1977 · doi:10.1901/jeab.1977.27-195