Effects of alternative reinforcement sources: A reevaluation.

Attwood et al. (1988) worked with pigeons that pecked a key for food. They compared two ways to cut the pecking. One way was DRO: food arrived only if the bird stopped pecking for a few seconds. The other way was VT: food just dropped on a timer, no matter what the bird did.

Both schedules gave the same amount of food. The team watched which schedule slowed the birds more.

DRO won. Birds pecked less under DRO than under VT. Longer gaps between peck and food made the drop even bigger.

The result says timing matters. When food can arrive right after a pause, birds learn to pause more.

Rey et al. (2020) ran almost the same test and got the same answer: DRO beats VT. They added a third wing—plain extinction—and still saw the biggest drop with DRO. Their paper calls itself a direct replication of the 1988 work.

Rey et al. (2020) also asked why DRO works. They saw birds do other things—like standing still—right before food came. They say DRO might work partly by accident: it rewards any behavior that is not the target peck. The 1988 paper only talked about delay, not about these accidental rewards. Same data, two lenses.

Dove (1976) did an earlier pigeon study that also reinforced “not pecking.” That paper found the same edge over noncontingent food. Attwood et al. (1988) built on that idea but used cleaner VT controls.

If you run DRO with a client, remember it is more than a timer. The pause itself gets reinforced, and any other move the client makes during the pause can get accidentally strengthened. To boost the effect, keep the delay short at first, then stretch it. Pair DRO with cues so the learner can feel when the pause starts. Watch for new, quiet behaviors that might pop up—they are your allies.