Dependency, temporal contiguity, and response-independent reinforcement.

Decasper et al. (1977) worked with pigeons in a small lab chamber.

The birds pecked a key under three different grain schedules.

One schedule paid only for pecks (VI). One paid no matter what (VT). A third mixed both (tandem VI-FT).

Birds pecked fastest when every reinforcer hinged on a peck.

They pecked slowest when grain came no matter what.

The mixed schedule sat in the middle, proving timing still matters even when the response is required.

Madden et al. (2003) later repeated the idea with college students pressing a button. Target presses only rose when free points came right after the button press, extending the pigeon finding to humans.

Meuret et al. (2001) used fixed-time schedules to cut baseline behavior. They showed the drop gets bigger when the free-food rate clearly differs from the old rate, building on the VT slowdown seen here.

Macht (1971) had already shown that simply making grain last longer pulls birds to stay in that side of the chamber. The 1977 paper adds that when the grain comes, not just how long it lasts, also steers responding.

For BCBAs, the lesson is to watch the clock, not just the contingency. A reinforcer that is "earned" but delivered late can still thin out the behavior. Check how quickly praise, tokens, or edibles follow the target response. If the gap grows, expect the rate to fall even though the plan says the kid "gets it for doing it." Tighten the delay and responding should bounce back.