A comparison of positive and negative reinforcement for compliance to treat problem behavior maintained by escape.

The team worked with five children who hit, screamed, or ran off when asked to do schoolwork.

They compared two ways to reward compliance. One group got a bite of cookie after each task. The other group got a short break from work.

The design flipped the two rewards back-to-back in the same session so each child served as their own control.

Cookie reinforcement cut problem behavior to near zero for every child. Break reinforcement helped only two kids and did little for the other three.

No child needed escape extinction. The edible alone did the job.

Kahng et al. (1999) ran the same cookie-vs-break contest sixteen years earlier and saw the same win for edibles. Au-Yeung et al. (2015) is a clean replication.

Carter (2010) widened the prize bin to include toys and still found positive reinforcement superior, again with no extinction required.

Najdowski et al. (2003) and Siu et al. (2011) look like they disagree. In feeding studies, escape extinction was mandatory because escape reinforcement alone failed. The difference is the response class: swallowing food is harder to shape than completing a worksheet, so extinction stays on the menu for feeding cases.

If a learner’s problem behavior is all about avoiding tasks, try handing a highly preferred edible right after compliance before you block escape. You may get the same drop in problem behavior without the extinction burst. Start with a single, easy instruction, deliver the bite immediately, and track if the break option is even needed.