Algorithmic shaping and misbehavior in the acquisition of token deposit by rats.
Computer-run shaping teaches rats as fast as people do, but it still creates the same sticky misbehavior.
01Research in Context
What this study did
Fifteen rats learned to drop tokens into a slot for food.
A computer watched each rat through a camera.
The machine gave food every time the rat moved closer to the slot.
No human hand moved the food hopper.
What they found
Thirteen rats learned the token drop in about the same time as hand-shaped rats.
Two rats never caught on.
Almost every rat also started chewing the slot, the walls, or the token.
The chewing stayed even after the rats earned all their food.
How this fits with other research
Mace et al. (1990) got kids to talk about their own response rates and saw the rates change.
Both studies show shaping works, but species and response differ.
Navarick et al. (1972) saw rats bite the lever during shock avoidance.
Like Davison et al. (1989), the biting looked like misbehavior, not the target response.
Together the papers warn: unwanted topographies pop up in rats across very different tasks.
Why it matters
If you use automated shaping software or tablet apps, watch for odd side responses.
Record them early; they can stick even after the skill is mastered.
Build in brief human checks to catch chewing, tapping, or swiping that the program might reinforce by accident.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Run a one-minute probe: watch the learner and list any extra movements the software might be feeding.
02At a glance
03Original abstract
In two experiments, rats were trained to deposit ball bearings down a hole in the floor, using an algorithmic version of shaping. The experimenter coded responses expected to be precursors of the target response, ball bearing deposit; a computer program reinforced these responses, or not, according to an algorithm that mimicked the processes thought to occur in conventional shaping. In the first experiment, 8 of 10 rats were successfully shaped; in the second, 5 of 5 were successfully shaped, and the median number of sessions required was the same as for a control group trained using conventional shaping. In both experiments, "misbehavior," that is, excessive handling and chewing of the ball bearings, was observed, and when the algorithmic shaping procedure was used, misbehavior could be shown to occur in spite of reduced reinforcement for the responses involved.
Journal of the experimental analysis of behavior, 1989 · doi:10.1901/jeab.1989.52-27