Changeover delay effects on topographically tagged discriminative behavior.

Pigeons pecked a key for food on a VI schedule.

A second key lit up as a signal.

The birds started pecking this signal key too, even though it never gave food.

The team added a 2, 3, or 4-second changeover delay.

They wanted to see if the pause would stop the extra pecking.

Longer pauses cut signal-key pecks in half.

The 4-second delay almost wiped the extra behavior out.

The pause broke the accidental chain: peck signal → move to food key → get food.

Without the chain, the signal key lost its magic.

Finney et al. (1995) later tested hens with 5–15 second delays and saw sharper choice matching.

Their longer delays helped, not hurt, because they measured different things.

Davis et al. (1972) and Corrigan et al. (1998) both showed that any delay cuts response rates.

These studies line up: gaps weaken behavior, whether the delay is signaled or not.

Barnard et al. (1977) found the same 4-second cliff in autoshaping, backing the idea that immediacy matters.

If a client keeps pressing a non-functional button before using the real one, add a 2–4 second pause after the switch. The brief wait can break the accidental chain and clean up the response pattern.