Implications for practice: Resurgence and differential reinforcement of alternative responding.

Austin et al. (2015) wrote a think-piece, not an experiment.

They asked: what if, during DRA, we teach several good behaviors in a row instead of just one?

The goal is to steer any later resurgence toward appropriate acts, not problem behavior.

The paper offers a plan, not data.

Serial DRA—reinforcing two or three alternative responses in rotation—may act like a safety net.

If resurgence hits, the behavior that pops back is more likely to be one of the taught good ones.

Perez et al. (2015) tested the idea the same year. Three adults with developmental disabilities got serial DRA. Resurgence dropped compared with classic single DRA. The lab result lines up with the theory.

Nist et al. (2023) ran a near-copy study with rats and found the opposite: five alternatives in serial DRA did not beat one alternative. The clash looks real, but species and procedure differ. Rats pressed levers for food; humans earned tokens for functional responses.

Fleck et al. (2023) extended the thought in a new direction. They used concurrent DRA—two choices at once—while keeping problem behavior on reinforcement. Resurgence vanished. Together the set shows “more alternatives” can help, but the how and when still shift.

You now have a low-cost tweak to try: teach two or three communication or play responses in rotation before thinning reinforcement. It may not guarantee zero resurgence, yet the human data give you a reason to test it, especially when relapse would be dangerous. Start small, track resurgence, and let the data guide you.