A comparison of pecking generated by serial, delay, and trace autoshaping procedures.
Stimulus-food timing drives how fast new responses appear — keep gaps tiny in your own sessions.
01Research in Context
What this study did
Glover et al. (1976) worked with pigeons in a lab.
They compared three ways to teach key pecking.
Serial, delay, and trace autoshaping were tested side-by-side.
What they found
Serial autoshaping won.
Birds pecked fastest when the light and food overlapped.
Long trace gaps slowed learning and changed how the birds moved.
How this fits with other research
Barnard et al. (1977) extended the idea.
They added omission schedules and showed that even after pecking starts, a four-second food delay kills the response.
Frost et al. (1996) kept the same setup but swapped pellet size.
Bigger pellets made the birds peck faster and open their beaks wider, proving reinforcer size still matters after the response is born.
Corrigan et al. (1998) later showed that unsignaled delays push birds into hopper-watching instead of pecking, linking response-reinforcer gaps to the same contiguity story.
Why it matters
The lesson for you: contiguity is king.
Keep the reinforcer close in time to the response or the stimulus.
When you shape a new skill, deliver praise, tokens, or snacks within seconds.
If you must use a delay, signal it clearly and keep it short.
These pigeon data remind us that basic conditioning rules still guide human teaching.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Time your reinforcer delivery — aim for 1 s or less after the target response.
02At a glance
03Original abstract
Pigeons were exposed to serial, delay, and trace autoshaping procedures. In Experiment I, all conditioned stimuli (CSs) were changes in illumination of the response key. The number of trials to acquisition of the keypeck increased from serial, to 4-sec delay, 8-sec delay, and 8-sec trace procedures, in that order. In Experiment II, which used a longer intertrial interval, trials to criterion increased from 8-sec delay, to 28-sec delay, 8-sec trace, and 28-sec trace procedures, in that order. In Experiment III, two groups received serial procedures in which the first CS was either a tone or a houselight, and the second was a keylight. The tone group acquired the key peck more rapidly than the houselight group. Early in conditioning in these experiments, and when the conditioned stimulus was a change in the keylight, there was a short latency to the onset of pecking and pecking was directed at the CS. After extensive conditioning, or when the CS was relatively diffuse, pecking still occurred, but had a longer latency and was not reliably directed toward the conditioned stimulus.
Journal of the experimental analysis of behavior, 1976 · doi:10.1901/jeab.1976.25-227