Reinforcer efficacy in a delayed matching-to-sample task.
In matching tasks, the gap between sample and reinforcer drives accuracy more than the gap between sample and choice.
01Research in Context
What this study did
Eisenmajer et al. (1998) worked with pigeons in a delayed matching-to-sample task. They asked: does the gap between sample and reinforcer hurt accuracy more than the gap between sample and choice?
The birds saw a color sample, waited, then picked the matching key. The team varied two delays independently: sample-to-choice and sample-to-reinforcer.
What they found
Accuracy dropped hardest when the reinforcer was late, not when the choice was late. The sample-reinforcer interval controlled performance; the sample-choice interval mattered less.
How this fits with other research
Burgio et al. (1986) had earlier said the opposite: longer choice delays hurt more than longer reinforcer delays. The two papers seem to clash, but they tested different delay ranges. D et al. used short reinforcer delays and very long choice delays; R et al. spread both delays wider, letting the reinforcer timing show its power.
Madden et al. (2003) later built on R et al. by mapping entire forgetting curves. They showed that longer sample-reinforcer intervals steepen the whole curve, backing the 1998 claim that reinforcer timing is the key variable.
Sayers et al. (1995) kept the same task and added signaled reinforcer size. They found that cues about upcoming food can also drop accuracy, echoing R et al.’s warning that anything happening between sample and reinforcer can weaken stimulus control.
Why it matters
When you run conditional-discrimination drills, deliver the reinforcer right after the response. If you must wait, keep the gap short and neutral; don’t chat or present extra cues. This simple timing tweak can save you from accidental errors in listener responding, visual matching, or any task that relies on remembering the sample.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Check your delay between sample and reinforcer; if it’s longer than two seconds, tighten it.
02At a glance
03Original abstract
Five domestic hens were exposed to a delayed matching-to-sample task. Conditions 1, 5, and 8 were variable-delay conditions in which five delays (0.25, 1, 2, 4, and 8 s) from the red or green sample to the presentation of the red and green comparison stimuli were presented a number of times during each session. In the fixed-delay condition (Condition 3), each delay was presented for 15 sessions under a Latin square design across birds. When improvements in accuracy across the variable-delay conditions are taken into account, the data were similar under both the variable and fixed delays. In Conditions 2, 4, 6, and 7 sample-reinforcer intervals were held at 8, 8, 4, and 2 s, respectively, while sample-choice intervals were varied within these during each session. With increasing sample-reinforcer interval, both initial discriminability (i.e., with sample-choice delay = 0) and rate of decrement in discriminability decreased. Although the former would be predicted if accuracy depends of the average sample-reinforcer interval, the latter would not. These data show that increasing the sample-choice interval had less effect on matching accuracy than increasing the sample-reinforcer interval did.
Journal of the experimental analysis of behavior, 1998 · doi:10.1901/jeab.1998.69-77