Combinations of response-reinforcer dependence and independence.

Pigeons pecked a key for food on a variable-interval schedule. The researchers mixed in different amounts of free food that had nothing to do with pecking.

They wanted to see how extra, non-contingent food changed the birds’ key pecking.

The more free food the birds got, the less they pecked. Even small doses of response-independent food weakened the link between pecking and pay-off.

The study showed that keeping the response-reinforcer bond tight matters for maintaining behavior.

Kuroda et al. (2018) later proved that the correlation itself, not just timing, drives behavior. They built on the 1974 warning by showing that even unseen correlations can strengthen or weaken responding.

Coe et al. (1997) looked closer at the ratio of free food to earned food. They found that timing and context can make non-contingent food suppress or boost responding. This explains why some early studies saw mixed effects.

Iwata (1993) reframed the drop in pecking as an economic choice: when free food is scarce, working for food is worth it. Together these papers tell us to guard both contingency and context.

If you add extra reinforcers that aren’t tied to the target response, you risk weakening the behavior. Keep reinforcers contingent in therapy and classroom token systems. When you must use non-contingent rewards, track response rates to be sure the skill doesn’t fade.