Concept formation & Stimulus eq Flashcards
Herrnstein & Loveland (1964)
Multiple VI-60s schedule EXT – half the time reinforcement was available and half the time it was not. Displayed 80 photo slides - some were + instances of the concept person, and some were negative (roughly half). Responses to + instances were reinforced, whilst those to - were not.
Graph - found higher response rates for + slides compared to -.
BUT – did not use novel slides - hence could just be learning individual contingencies.
Herrnstein, Loveland & Cable (1964)
Repeated but used novel slides.
Tree, water, person = concepts.
Showed less distinct evidence of concept formation (overlap) BUT was more reliable as removed possibility that the birds were just learning individual contingencies due to the use of novel responses.
BUT
argument only applies in early testing as reinforcement were given for responding to + slides during testing – hence may be using reinforcers are signals….
Herrstein (1979)
-mean positive rank
REPEATED using novel slides + only reinforced following the pigeon response +showed it could be reversed
==overall good evidence for concept learning by the pigeons
Wright et al
Arranged a matching to sample task.
Control: got standard MTS training that used two training stimuli pictures
Experimental: trial unique-stimulus training with 76 slides.
–> Prediction was that the control group would just learn the two stimulus-specific rules but the experimental group would learn the matching concept.
Control: accuracy reached 75% after 16 sessions = 1200, whereas the experimental group only needed 360 sessions… Good evidence + when tested with novel stimuli, control group was unable to match but the experimental group were.
Sideman & Cresson (1973)
OTM training.
Names read out to child –> Names written
Names read out to child –> Pictures
found that not only did picture names written relationship emerge, but the names were also spoken by the child.
