Real-time detection of orientation during negative behavioral contrast with key pecking and a turning response.

Shah (1992) filmed pigeons while they pecked a key for food. A computer tracked head angle in real time. When the bird turned away from the key, the screen flashed and gave a brief food hopper.

The birds first pecked on a rich schedule. Then the schedule thinned to extinction in one component. This drop causes negative contrast: less pecking in the now-poor component.

Negative contrast still happened even though the birds could earn food by turning away. The turning response rose, but peck rate still fell in the poor component.

Orientation tracking showed the turn was a side effect, not the engine of contrast. The drop in key pecks drove the contrast, not avoidance of the key.

Van Hemel (1973) and Green et al. (1975) saw contrast in moment-to-moment peck rates. Shah (1992) adds live head-angle data, showing those rate drops are not caused by turning away.

Davol et al. (1977) cut negative contrast by reinforcing long pauses. Shah (1992) tried the opposite: reinforce turning away. Both studies prove you can tweak the response class without erasing contrast.

Gentry et al. (1980) shaped tiny changes in peck duration. Shah (1992) shows you can also shape where the head points, yet baseline reinforcement still rules overall peck rate.

If a client looks away during hard tasks, you might think the glance causes the drop in correct responses. This study says watch the main response first. Reinforce looking back, but know the true problem is likely the reinforcement rate for the target skill. Track both behaviors in real time; fix the schedule, not just the gaze.