Abstract:
Davison and Jenkins (1985) suggested that behaviour allocation between concurrent alternatives was affected by the
the degree to which animals could discriminate between the
response-reinforcer contingencies associated with each of those
concurrent alternatives. The present study evaluated their
models for free-operant concurrent-schedules performance and
discrete-trial signal-detection performance.
In Experiment 1, pigeons were trained in six sets of
conditions. In each set of conditions, two intensities of white
light were used as the sample stimuli in a discrete-trial signaldetection
procedure. The relative reinforcer frequency for
correct choice responses was varied across conditions. The same
two stimuli were then arranged as the discriminative stimuli in a
switching-key concurrent schedule. The relative reinforcer
frequency was varied across conditions. Then the intensity of
one of the white lights was varied, and these procedures were
repeated for the next set of conditions. Analysis of the data
from the signal-detection procedures showed that the effect of
varying relative reinforcer frequency decreased as
discriminability between the stimuli increased. A new model of
signal-detection performance, based on the Davison and Jenkins'
(1985) model of concurrent-schedule performance, accounted for
this interaction. There was ordinal relation between measures of
stimulus discriminability obtained from the signal-detection
procedure, and measures of the effect of varying relative
reinforcer frequency obtained from the concurrent-schedule
procedures.