The changing criterion design.

Lewis et al. (1976) wrote the first how-to guide for the changing-criterion design.

They showed step-by-step graphs so you could copy the method in your own study.

No new treatment was tested; the paper only explains the design.

The authors proved you can track slow, stair-step behavior change with one person.

Each week you raise the target a little, and the line should climb with it.

If the data steps up with your new goals, the treatment is working.

Ferron et al. (2023) later added coin-flip randomization to the same design. You now decide ahead of time which days the goal will shift, so your stats are cleaner.

Levin et al. (2019) give a free Excel tool that works for any single-case study, including changing-criterion. You type in phases and the sheet randomizes them for you.

Manolov et al. (2022) built a free web plot that lets you check if the stair-step effect repeats in new participants.

Use this design when you want to shape behavior a little at a time—like adding one more math problem each day. Plot the steps on your graph, and if the data follow the steps, you have a clear effect. Pair it with the new random tools from Ferron et al. (2023) to make your study tougher and easier to publish.

The changing criterion design is a single-case design, related to the multiple-baseline design, used to evaluate a treatment that gradually changes a behavior in a stepwise fashion. After an initial baseline, a performance criterion is set and reinforcement is delivered for meeting it.

Experimental control is demonstrated when the behavior repeatedly changes to closely match each new criterion. Each successive criterion phase acts as a baseline for the next, so behavior tracking the criterion up (or down) again and again rules out coincidental change.

On the graph, the criterion for each phase is drawn as a horizontal line, and the data path steps to a new level as the criterion changes. A convincing demonstration usually needs several criterion changes and phases long enough for responding to stabilize at each level.

Strength is increased by varying phase lengths and the size of criterion changes, and sometimes by briefly reversing the criterion to a prior level to confirm the behavior follows it. The design suits goals shaped gradually, such as exercise, fluency building, or reducing smoking.