Stimulus Presentation Ratios And The Outcomes For Correct Responses In Signal-detection Procedures.

Researchers worked with pigeons in a signal-detection task. Birds pecked one key when they saw bright light and another key when they saw dim light.

The team changed two things across rounds: how often each light appeared and how often correct pecks earned food. They also tested extra feedback clicks on trials that were correct but did not pay off.

The birds’ choices matched the mix of lights and the mix of payoffs. More food for a correct dim choice made the pigeons peck the dim key more, even if the bright light still showed up most.

Adding clicks on non-paid correct trials did not change choice at all. The pigeons still followed the food ratios, not the extra noise.

Nevin (1967) saw the same thing years earlier: changing how often food follows a correct response alters how often the bird pecks, but the bird still sees the difference between the lights.

Todorov et al. (1984) add that frequency beats size: many small pellets steer behavior more than a few big ones. Farmer-Dougan et al. (1999) now show that even zero pellets still count in the math, as long as the food ratio across choices stays clear.

Zentall et al. (1975) found that light or sound alone can keep rats pressing when food is free. Farmer-Dougan et al. (1999) turn that idea around: once food odds are known, extra sound without food is ignored.

Your client’s choice follows the same rule: relative payoff rate drives responding. When you thin reinforcement on one program, boost the rate on another skill or the learner may drop the first task. Extra praise without tangible pay does not wreck the balance, but it will not save it either. Track your ratios first, then add praise if you want—just don’t expect praise alone to do the heavy lifting.