Signalled and unsignalled percentage reinforcement of performance under a chained schedule.

Pigeons worked through a two-link chain. First they pecked 15 times. Then they waited 25 seconds.

Food came only some of the time. A red light sometimes told them "no food this round." The team counted how many chains the birds finished.

The red light boosted work. Birds finished more chains when the skip was signalled than when it stayed silent.

The lift was not the same for every bird or every food-odds setting. Results were mixed, but the signal helped overall.

Zentall et al. (1975) saw the same spark. Pigeons kept pressing a bar when peeps and flashes followed each press, even while free food sat in the dish. Both studies show that cues born from scarce food can feed the behavior by themselves.

Peiris et al. (2022) sounds like the opposite story. Dogs given a click but only occasional treats quit working and made more errors. The clash is only surface: the click had once meant every treat, so skipping broke a promise. Branch (1977) never promised food on every chain; the red light simply said "not this time," and that news itself became a tiny reward.

Keesey et al. (1968) set the base. They proved that a cue grows stronger the more times it hugs real food. Branch (1977) moves the idea forward: even the hug that says "no food" can turn into fuel if it shows up enough.

Your clients may keep working for tokens, points, or praise even when the big prize comes rarely. Build clear signals for skip trials instead of hiding them. The signal itself can keep the chain alive while you stretch reinforcement. Just guard the trust: if the learner expects a sure backup, deliver it. Use signalled omission only when the schedule was always lean.