Effects on preference of reinforcement delay, number of reinforcers, and terminal-link duration.

Pigeons pecked two keys in a two-part choice game. First, each peck on a key started a brief "link" that always ended the same way.

Then the bird entered a "terminal" phase where it earned food after a set delay. The team varied three things: how long the wait was, how many food pellets came at the end, and how long the terminal phase stayed on.

Longer waits usually hurt a key’s popularity. But when the wait came with extra pellets, birds often stayed loyal to that side.

If the terminal phase ended quickly, birds picked it even when the wait was long and the payoff small. Quick escape trumped both delay and amount.

Lloyd (2002) seems to clash: keeping a fixed 10-second gap between delays still killed preference as absolute delays grew. The difference is procedure. Schmitt (1984) paired longer waits with more food; Lloyd (2002) held food amount flat. Extra pellets, not just the gap, kept birds choosing.

Gowen et al. (2013) extend the story into one session. They showed choice can flip from "impulsive" to "self-control" as delays rise within the same hour. Schmitt (1984) showed the static snapshot; Emma et al. showed the movie.

Rider (1983) used rats and mixed delays instead of concurrent chains. Both papers agree: sprinkle in enough short delays and subjects forgive the long ones. Species and setup differ, but the tolerance pattern holds.

When you set up token boards, response chains, or delayed praise, remember three levers: wait time, payoff size, and how fast the student can exit the activity. If the task is tough, shorten the terminal requirement first; that alone can save the program. Then add extra reinforcers if you must stretch the wait.