Systematic Review of Differential Reinforcement of Alternative Behavior Without Extinction for Individuals With Autism.

Stephens et al. (2018) looked at every paper they could find on DRA without extinction for kids with autism. They kept only studies that gave rewards for good behavior while still letting problem behavior contact normal consequences.

Ten studies made the cut. The team asked: can we drop extinction and still cut problem behavior by tweaking how we deliver reinforcement?

Nine of the ten studies showed clear drops in problem behavior. The trick was to change one of four levers: make the good-behavior reward bigger, faster, better quality, or come around more often.

No study had to add extinction once these levers were tuned. Kids kept the new skills and problem behavior stayed low.

Kunnavatana et al. (2018) is one of the ten studies in the review. They tested each child to see which lever—size, speed, or quality—mattered most, then tilted only that lever. The review shows their single-case success is part of a wider pattern.

Rey et al. (2020) looks like a clash at first. Their lab work says DRO works because it accidentally rewards any other response. Stephens et al. (2018) says we should purposefully reward one specific alternative. The gap is in the procedure: accidental rewards help, but planned, rich rewards for one clear behavior help more.

Weinsztok et al. (2023) adds a timing tip. Their 2023 review found edible or toy rewards speed up new skill learning. Stephens et al. (2018) says the same tools can also replace extinction when we want to stop problem behavior.

You can run a gentler plan tomorrow. Pick the replacement behavior you want to grow, then test if the child cares more about bigger candy, faster praise, or a favorite toy. Push that one lever first. You may never need to place problem behavior on extinction, which saves you from emotional bursts and keeps parents on board.