Effects of alternative reinforcement: does the source matter?

Davis et al. (1972) asked a simple question: does it matter where the extra treats come from? They placed pigeons in a box with one key. Pecks on that key paid off on a variable-interval (VI) schedule.

While the bird worked, the researchers pumped in "free" reinforcers from three different sources: signalled, unsignalled, or only after the bird paused. They wanted to know if the source changed how much the bird pecked.

More free treats always meant fewer key pecks. The drop followed the same matching-law curve no matter how the extras arrived.

Signalling, not signalling, or tying extras to a pause made no difference. The birds simply matched their behavior to the overall rate of payoff.

Rose et al. (2000) later showed the story is trickier. When they changed sucrose strength, the theoretical ceiling k in Herrnstein’s equation moved with it. That breaks the fixed-parameter assumption Davis et al. (1972) relied on.

Wilkie et al. (1981) added that smaller reinforcer volumes shift the whole response curve rightward. You need more frequent payoffs to keep the same rate if the treat is tiny.

Bradshaw et al. (1978) found birds peck keys more than they press levers even when payoff rates are equal. Response type itself adds bias, another factor the 1972 paper did not explore.

For BCBAs, the core message is still useful: extra reinforcement dilutes baseline responding in a predictable way. But don’t treat the matching equation as carved in stone. Reinforcer size, type, and response form all nudge the numbers. When you add outside rewards—token boards, edible praise, iPad time—expect the client’s main response rate to drop and plan your schedule adjustments accordingly.