Effects of reinforcer duration on responding in two-link chained interval schedules.

Lendenmann et al. (1982) asked a simple question: if you make the reinforcer last longer, do pigeons peck more?

They set up a two-link chain. In the first link, birds pecked on a variable-interval schedule. Completing that link moved them to a second link where pecks produced food.

The team varied how long the grain feeder stayed on—short, medium, or long—while everything else stayed the same.

Longer feeder time pushed response rates up in both links, but the first link jumped the most.

In plain words, making the payoff bigger made the birds work harder, especially early in the chain.

Landon et al. (2003) saw the same pattern in concurrent schedules: bigger reinforcers created bigger “preference spikes,” a direct echo of the 1982 duration effect.

Oliver et al. (2002) added the next layer: rate and magnitude add together. They showed that preference and resistance to change grow when both are high, extending the 1982 finding to choice situations.

Attwood et al. (1988) seems to disagree. They found that delaying all reinforcers to the end of the session hurt response rates. The studies differ because 1988 removed immediacy while 1982 kept it and only changed duration—timing still mattered in both.

When you build token boards, task chains, or DRL programs, remember that bigger or longer payoffs boost early responding most. If you need more work at the start of a chain, increase reinforcer size first. If behavior stalls, check whether the delay—not the size—has crept up, and keep the delivery tight.