Preference for starting and finishing behavior patterns.

Researchers watched pigeons choose between two keys. Each key started a different pattern of pecks. The birds could see when reinforcement was coming on one schedule, but not on the other. The team wanted to know if the birds' first choice predicted whether they would finish the whole pattern.

They used concurrent schedules. That means both patterns were available at the same time. The only thing that changed was whether the upcoming food time was signaled or silent.

When the birds could not tell when food was due, the pattern they started was usually the one they finished. Their first pick predicted the full run. When a light or sound gave away the timing, that link disappeared. The birds often switched before the end.

In plain words, signaling the reinforcer broke the momentum of the pattern they had begun.

Mace et al. (1990) later showed the same rule works with adults in a workshop. Higher reinforcer rates made the workers stick with their task despite noise, just like unsignaled food kept pigeons in their pattern. The lab result translated to humans.

Rilling et al. (1969) and Hawkes et al. (1974) had already proven that pigeons match their time to reinforcement rates. Wehman et al. (1989) adds a twist: the signal itself can override that matching when it tells the bird 'wait, something better is coming.'

Adkins et al. (1997) looked at several upcoming reinforcers, not just the next one. Their math fits here too. Once the signal reveals the next food, the bird recalculates and may jump ship, breaking the start-to-finish link.

If you want a client to finish a task chain, hide the clock. Unsignaled reinforcement keeps the momentum going. If you need flexibility and want to teach shifting, add clear signals that tell when each option pays off. Use this trade-off to build either persistence or adaptive switching into your programs.