Divided stimulus control depends on differential and nondifferential reinforcement: Testing a quantitative model

The team worked with pigeons in a lab. Birds pecked a screen that showed two cues at once: color and line angle.

Sometimes only color told them which peck paid off. Sometimes only angle mattered. Sometimes both did.

The researchers changed the payoffs. Correct pecks earned most grain. Wrong pecks earned none, a little, or a lot. They tracked which cue the birds then followed.

When wrong pecks paid more, birds shifted control to the other cue. Extra grain for errors literally pulled attention away from the ‘right’ dimension.

A math model fit the shifts, but only when the payoff change was signaled. Without a signal, the model slipped.

Grosch et al. (1981) showed pigeons like signaled delays. Gomes‐Ng et al. (2023) adds that signals also lock in the power of reinforcer ratios.

Eugenia Gras et al. (2003) proved compliance rises only when requests climb a hierarchy. The new study mirrors this: reinforcement topography, not just presence, decides what the animal learns.

Lincoln et al. (1988) found constant time delay beats least prompts for kids with autism. Both papers use single-case designs and differential payoffs, showing the method works across species and skills.

Your client’s errors are also behavior. If they get any payoff—attention, escape, even tiny tokens—you may accidentally shift stimulus control to the wrong cue. Check what follows errors in your task. Strip reinforcement or make it clearly different from the reinforcer for correct responses. A quick signal ("good try, but no") can keep your discrimination plan on track.