Observational learning of two visual discriminations by pigeons: a within-subjects design.

Scientists worked with eight pigeons. Each bird had to peck the correct picture out of two to earn grain.

Before the birds tried alone, half watched another pigeon already doing the task. The other half only saw the pictures or got grain with no model.

The team timed how fast each bird learned the right choice. They repeated the test with a second new picture pair.

Birds that watched a peer reached 80 % correct in about half the trials. Birds without a model needed almost double the tries.

Seeing a successful pigeon worked better than just seeing the pictures or getting free grain. Watching gave a clear learning boost.

Shvarts et al. (2020) later used the same pigeon setup and added kids with autism. They showed that pairing a praise light with food also helps, but the effect is smaller and jumpy.

van Laarhoven et al. (2003) asked whether showing many examples fast speeds learning. They found cross-trial examples help speed, while within-trial examples help generalization. Siegel et al. (1986) adds that a live model beats both schedules.

Syriopoulou-Delli et al. (2012) taught identity matching to a child with autism using blocked trials and saw fast success. The pigeon study seems opposite because K used trial structure, not a model. The clash fades when you see K worked with a boy who had autism, while B used healthy birds. Different brains, different tools.

You now have lab proof that simply letting a learner watch a skilled peer can cut training time in half. Try seating a new client beside a proficient partner before direct trials. Keep the session short, let the eyes do the first pass, then run your usual teaching. You may save precious therapist hours and reduce client frustration.