Previous
Generalization: the pigeon experiments
We're taking a brief detour from signal detection now to delve into comparative psychology and stimulus discrimination. The connection between these two fields will soon become clear.
In
1956, two researchers trained pigeons to peck at a single color. It has
become customary to call this training color (or any positive stimulus)
"S+."
They then presented the trained pigeons with a range of colors, including the S+ color. The pigeons pecked most at the S+ color, but they also pecked at similar colors surrounding the S+ color.
Redrawn from Guttman and Kalish (1956) (Full
reference)
This is called generalization: the pigeons generalized from the training color to other, similar colors. In other words, they guessed that although in the past they found a reward with one color, other similar colors might also yield a reward.
What happens if you train the pigeons
to peck at one color, but to avoid pecking
at a similar color?
Top
of page