An assessment of differential reinforcement procedures for learners with autism spectrum disorder

Johnson et al. (2017) tested three ways to set up reinforcement for kids with autism.

They compared giving a favorite snack (quality), giving it every time (schedule), or giving a bigger piece (magnitude) against plain non-differential praise.

Each child tried all four setups while learning one new skill so the team could see which one helped the child learn fastest.

For the first skill, the favorite-snack setup won for every child; kids finished trials faster and made fewer errors.

When the team moved to new skills, the snack advantage did not always repeat; sometimes another setup worked just as well.

Bottom line: quality-based reinforcement looked best at first, but the winner did not stay the same across skills.

Ward-Horner et al. (2017) ran a similar 2017 study and found that boosting quality or size flipped kids’ preference from intermittent to continuous schedules; their positive result lines up with the early wins Johnson saw.

Weston et al. (2020) later used token boards and still saw higher work rates under big, accumulated reinforcement, showing the same pattern holds when you swap snacks for tokens.

Chou et al. (2010) had complained that we lacked clear rules for picking a reinforcement format; Johnson’s mixed findings explain why—what works can change with each new task, so one rule will not fit every lesson.

You cannot assume the reinforcer that works today will win tomorrow. Run a five-minute probe—quality vs schedule vs magnitude—whenever you introduce a new program. If the child slows down, rerun the probe instead of blaming motivation. This quick habit saves session time and keeps learning efficient.