Saving the best for last? A cross-species analysis of choices between reinforcer sequences.

Grindle et al. (2012) let pigeons and college students pick between two strings of food or money. Each string had three rewards, but the delays were shuffled. In one condition the tokens could be traded in right away; in another they had to wait until the end.

The birds pecked keys; the humans clicked boxes on a screen. Researchers recorded which sequence each species chose most often.

Both pigeons and humans picked the sequence that got the first reward to them fastest when exchange was immediate. They did not care which reward came last.

When exchange was delayed until the session ended, the pattern fell apart. Choice became messy and no longer lined up with simple delay-discounting math.

Gowen et al. (2013) later saw the same reversal inside a single session: pigeons started by taking the small-quick reward, then switched to the large-late one as waits grew longer. The two studies line up—delay structure, not reward size, steers the choice.

Tvan der Miesen et al. (2024) looks like it disagrees. They found pigeons failed a two-event sequence task and seemed too dumb to learn order. The difference is the task. TR’s birds could win by just pecking the last color they saw. F’s birds had to pick whole token strings; no shortcut worked. Same species, different rule—no real conflict.

Delano (2007) and Green et al. (2004) set the stage by showing hyperbolic decay handles pigeon choices when only one delay is in play. F extends that work by proving the rule still holds when several delays are chained together.

If you give tokens a client can cash in right away, they will pick the schedule that delivers the first reinforcer fastest—even if the later payoffs are smaller. To promote patience, build in a forced wait before exchange. This tiny tweak, shown in both birds and humans, can make self-control programs more effective without extra tokens or larger prizes.