In 2001, Foster, Hackenberg, and Vaidya replicated these effects by measuring key
pecking in pigeons. Responses on the right side key were reinforced with an illuminated token.
Exchange periods were made available when the red center key was illuminated. During this
time, a peck on the exchange key produced access to food. In their first experiment, Foster et al.
alternated the exchange-production schedule between fixed-ratio and variable-ratio schedules
with the exchange period becoming available after 1-8 responses. For the higher ratio
requirements (e.g. FR 8 or VR 8), the variable-ratio schedules had produced higher response
rates. More importantly, as ratio requirements increased, responding decreased under the fixedratio
schedule and the variable-ratio schedule remained unaffected. Finally, for most pigeons,
preratio pausing was shorter under a variable-ratio exchange schedule when compared to a fixedratio
exchange schedule of the same value. In their second experiment, Foster et al. manipulated
both the token-production schedule and the exchange-production schedule. The exchange- and
token-production schedules were either a FR production schedule with a VR exchange schedule,
or a VR production schedule with a FR exchange schedule. The token-production requirement
was always 50 (either an average of 50 responses when it was a VR exchange condition, or a
SCHEDULES OF TOKEN REINFORCEMENT 11
fixed requirement of 50 responses when it was an FR exchange condition). The exchange
schedule requirement varied from 1-8 tokens. Results of the second experiment supported
results from the first experiment. When the exchange schedule was fixed-ratio, rate of
responding decreased as the token-production ratio schedule increased (e.g. from VR 4 to VR 8).
However, when the exchange schedule was variable-ratio, response rates were unaffected as the
token-production schedule increased (e.g. from FR 4 to FR 8). In addition, more pausing
occurred when the FR token-production schedule was increased than when the VR tokenproduction
schedule was increased. Rate of responding was steady throughout the session in the
VR schedule, meaning that after reaching an exchange opportunity, response rate in the next
token-production segment was similar to the previous one. However, responding increased
throughout the session in the FR schedule, meaning that after receiving reaching an exchange
opportunity, response rate in the next token-production segment was higher than the previous
one. Foster et al. demonstrated that responding under schedules of conditioned reinforcement is
similar to responding under simple schedules of reinforcement.
The purpose of the current study was to examine if schedules of conditioned
reinforcement is similar or distinct from simple schedules of reinforcement. In particular, the
current study examined if fixed vs. variable ratio exchange schedules of token reinforcement will
produce behavior similar to what is found under fixed and variable ratio simple schedules of
reinforcement.