Situational descriptions of behavioral procedures: the in situ testbed.
The In Situ testbed turns old pigeon curves into a free proving ground for new learning models.
01Research in Context
What this study did
Richman et al. (2001) built a free tool called the In Situ testbed.
It lets you drop any learning model onto real pigeon data.
The data cover four classic schedules: continuous reinforcement, extinction, fixed ratio, and fixed interval.
No live birds needed; the records are ready on the web.
What they found
The testbed ran three published models side-by-side.
All models fit the same four sets of cumulative records.
The paper shows how to spot which model tracks real behavior best.
How this fits with other research
Older pigeon work becomes new again. REYNOLDS (1964), Dodd (1984), and Striefel et al. (1974) collected the very curves now hosted in the testbed.
These studies are predecessors: they gathered the raw data that the tool recycles for model testing.
Peterson et al. (2024) looks like a contradiction at first. They test humans, not pigeons, and use conjugate reinforcement to study automatic behavior.
The clash is only surface-deep. Both papers build lab models for later use—one for clinical assessment, one for computer benchmarking.
Same blueprint, different questions.
Why it matters
If you tweak or code learning models, the testbed gives instant feedback.
You can swap in your own equations and see within minutes if they beat the classics.
No new animals, no new hardware—just download, run, and revise.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Download the testbed, plug in one published model, and compare its output to the supplied FR-20 curve.
02At a glance
03Original abstract
We demonstrate the In Situ testbed, a system that aids in evaluating computational models of learning, including artificial neural networks. The testbed models contingencies of reinforcement rising an extension of Mechner's (1959) notational system for the description of behavioral procedures. These contingencies are input to the model under test. The model's output is displayed as cumulative records. The cumulative record can then be compared to one produced by a pigeon exposed to the same contingencies. The testbed is tried with three published models of learning. Each model is exposed to up to three reinforcement schedules (testing ends when the model does not produce acceptable cumulative records): continuous reinforcement and extinction, fixed ratio, and fixed interval. The In Sitt testbed appears to be a reliable and valid testing procedure for comparing models of learning.
Journal of the experimental analysis of behavior, 2001 · doi:10.1901/jeab.2001.75-135