The effect of signaled reinforcement availability on concurrent performances in humans.

The team worked with adult volunteers in a lab. Each person sat at two buttons. Pressing either button could earn nickels on separate VI schedules.

In one test phase, a light came on to show which button currently had money ready. In the other phase, no light signaled the money. The team compared how often people pressed each button across the two phases.

When the light signaled money on Button A, people pressed Button A less. They moved over to Button B, where no light ever appeared.

The matching equation still held, but it now tilted toward the unsignaled side. The signal acted like a mild punisher for the signaled option.

Garcia et al. (1973) first showed pigeons match response ratios to reinforcement ratios. Goldman et al. (1979) now show the same rule works in humans, but a simple signal can bend the match.

Boutros et al. (2011) later unpacked why. A single reinforcer acts like a traffic light, steering the next response toward the just-rewarded side. The 1979 finding is an early human example of that steering effect.

Szatmari (1992) saw contrast when extra reinforcers moved between schedule parts. The 1979 study flips the idea: merely telling the client which side has the reinforcer can create the same shift, without moving any extra money.

Guest et al. (2013) tested brief unsignaled delays and found rate boosts in interval schedules. Their work warns us that signaling versus not signaling is not the same as delay versus no delay; schedule type matters too.

If you use concurrent schedules in token boards or choice tasks, watch what you tell the learner. A simple green light that marks the richer side can backfire and push the learner toward the unmarked side. Try running a quick probe with and without the signal to see which way your client actually leans.