PARTIAL REINFORCEMENT (ACQUISITION) EFFECTS WITHIN SUBJECTS.
Partial reinforcement speeds the early links of a response chain but can slow the final one.
01Research in Context
What this study did
WERTHEIWENZEL et al. (1964) worked with pigeons in a straight runway.
Each bird first got food every time it finished the run.
Later the same bird got food only after some runs.
The team timed three parts: start, middle run, and final goal entry.
What they found
Partial pay made the birds start faster and run faster.
Yet those same birds slowed at the very last step.
Continuous pay kept the final step quick but the early parts slower.
Where you measure in the chain changes the story.
How this fits with other research
Jones (1969) extended this idea into extinction.
Birds on partial pay kept the whole chain longer when food stopped.
Together the papers show: partial pay speeds early links and shields the chain later.
Weinsztok et al. (2023) looked at kids in DTT, not birds.
Toys and snacks beat praise for fast skill gain.
The two studies seem opposite: partial lost at the final step, tangible won.
The gap is in the final response.
A et al. measured the last step into the food box.
Weinsztok measured how fast kids mastered a new task.
Different endpoints, different answers.
Why it matters
Check where in the chain you take data.
If you want a brisk start, thin the schedule.
If the last step must stay quick, keep every response paid for now.
Try alternating: continuous for the final target, partial for the warm-up links.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Keep reinforcement dense for the very last step of the chain; thin it for the early steps to build speed.
02At a glance
03Original abstract
Acquisition performance of 22 rats in a straight alley runway was examined. The animals were subjected to partial reinforcement when the alley was black (B+/-) and continuous reinforcement when it was white (W+). The results indicated (a) higher terminal performance, for partial as against continuous reinforcement conditions, for starting-time and running-time measures, and (b) lower terminal performance under partial conditions for a goal-entry-time measure. These results confirm within subjects an effect previously demonstrated, in the runway, only in between-groups tests, where one group is run under partial reinforcement and a separate group is run under continuous reinforcement in the presence of the same external stimuli. Differences between the runway situation, employing a discrete-trial procedure and performance measures at three points in the response chain, and the Skinner box situation, used in its free-operant mode with a single performance measure, are discussed in relation to the present findings.
Journal of the experimental analysis of behavior, 1964 · doi:10.1901/jeab.1964.7-135