Duration comparison: relative stimulus differences stimulus age, and stimulus predictiveness.

Davis et al. (1994) tested how pigeons tell time. Birds pecked one key if a green light stayed on longer, another key if a red light stayed on longer.

The team made some durations easy to tell apart and others almost the same. They also changed how often the green light was the correct choice.

When durations got longer, the birds made more mistakes. Tiny duration differences were hardest to judge.

If green was usually the right answer, birds said 'green lasted longer' even when it did not. Predictiveness created bias.

Schmidt et al. (1969) showed that marking reinforcement duration with a hopper light sharpens discrimination. Davis et al. (1994) move the spotlight from reward length to stimulus length, showing timing errors grow with longer samples.

Parsons et al. (1981) found longer S+ exposure strengthens observing. The 1994 data flip the coin: longer samples weaken duration accuracy, pointing to different memory loads.

Reynolds (1966) saw contrast and peak shift vanish after extended training. Davis et al. (1994) add that extended duration itself hurts timing, hinting that some stimulus-control limits never go away no matter how much you train.

Your learners may 'know' the answer yet still err when intervals stretch or when one choice has a history of being right. Break long waits into shorter cues, and rotate which stimulus signals reinforcement so predictiveness bias cannot take root.