In order to examine the mechanism(s) responsible for response reductions fr
om noncontingent reinforcement (NCR), multiple magnitudes and densities of
NCR were superimposed on a variable ratio (VR) 3 schedule of contingent rei
nforcement for the performance of an arbitrary manual response. Data were c
ollected on responding that occurred during access to the reinforcer separa
tely from responding that occurred between reinforcer access intervals (i.e
., when the participant did not have access to the reinforcer). Higher magn
itudes and denser schedules of NCR produced greater reductions in respondin
g than did lower magnitudes and leaner schedules. Within-session response p
atterns suggested that decrements in responding were primarily a function o
f the increased amount of reinforcer access time associated with higher mag
nitudes and denser schedules of NCR, That is, it appeared that the particip
ant consumed reinforcers (regardless of whether they were delivered conting
ently or noncontingently) when they were available and responded for contin
gent reinforcers primarily when reinforcers were absent. (C) 2001 Elsevier
Science Ltd. All rights reserved.