Phase duration and resurgence

Smith et al. (2022) tested how the length of each phase changes resurgence. Adults without disabilities pressed buttons for points. First, the team paid them for one button for a short or long time. Next, they paid for a new button instead. Finally, they stopped all payments and watched the old button come back.

The study used a single-case design. Each adult repeated the three phases several times. The researchers measured how much the old button returned when payments ended.

Longer first-phase pay made resurgence bigger. Longer second-phase pay made resurgence smaller. The results fit a math model of choice between responses.

In plain words: the more you reinforce a behavior at the start, the harder it bounces back later. The more you reinforce the new behavior before extinction, the softer the bounce.

Greer et al. (2023) and Greer et al. (2020) looked at the same question in children with problem behavior. They also found that longer extinction treatment did not reduce resurgence. The lab result now matches the clinic: duration alone does not protect you.

Matson et al. (2009) showed the same history effect earlier. Adults playing a baby-care game resurgenced more when the first behavior had a longer pay history. Smith et al. (2022) turned that pattern into a clear formula.

Shahan et al. (2025) adds a twist. They showed that bigger alternative rewards shrink the extinction burst. Together, the papers say: make the new payoff either longer or richer to protect against relapse.

When you plan extinction, think in phases. If the client has years of reinforcement for the old behavior, expect a big resurgence. Build a long, strong alternative-reinforcement phase before you stop all rewards. Do not trust extinction length alone; pair it with a rich differential-reinforcement plan. Track data after you end treatment so you can catch the bounce and respond quickly.