Assessment & Research

Additional randomization test procedures for the changing criterion design

Tanious (2022) · Behavioral Interventions 2022
★ The Verdict

Two new randomization options—blocked alternating and randomized alternating—let you keep the stepwise look of a changing-criterion design while still randomizing.

✓ Read this if BCBAs who write or review single-case research using changing-criterion designs.
✗ Skip if Clinicians who only do discrete-trial or group-design work.

01Research in Context

01

What this study did

Tanious (2022) wrote a how-to paper. It gives two new ways to randomize steps in a changing-criterion design.

The new choices are called blocked alternating and randomized alternating. You still get the clean step-up graph, but the steps are chosen by chance.

02

What they found

The paper does not show client data. It simply hands you the new random rules and explains when to pick each one.

03

How this fits with other research

Elliffe et al. (2019) also gave randomization tools, but for trend tests. Tanious adds the same idea to the changing-criterion look.

Matson et al. (2013) found that a fixed order can speed up multielement FAs. That seems to clash with Tanious, who pushes random order. The difference is purpose: L wanted fast visual clarity; Tanious wants strong stats.

Killeen (1978) taught us to wait for stable data before moving on. Tanious updates that decision: use a coin flip to set the next criterion instead of eyeballing stability.

04

Why it matters

If you run changing-criterion studies, these two new random plans give you bullet-proof stats while keeping the easy-to-read stair-step picture. Pick blocked alternating when you want short runs; pick fully randomized when you have more time. Either way, you can now publish with randomization tests instead of just visual guesswork.

Free CEUs

Want CEUs on This Topic?

The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.

Join Free →
→ Action — try this Monday

Write your next criterion level on a slip of paper, draw it from a hat, and plot the point—your graph stays clean and your stats get stronger.

02At a glance

Intervention
not applicable
Design
methodology paper
Finding
not reported

03Original abstract

AbstractThe single‐point and range‐bound changing criterion design have been proposed as single‐case experimental designs. The only form of randomization currently available for these designs is phase change moment randomization. The present article describes two novel randomization procedures for changing criterion designs, the blocked alternating criterion randomization and randomized alternating criterion randomization, and illustrates their use. Appropriate randomization procedures are introduced for situations in which it is desirable to maintain the stepwise nature of the changing criterion design and situations in which the change in the dependent variable can be assumed not to follow the stepwise pattern. The article further discusses contexts in which each of the randomization procedures for the changing criterion design may be useful, and explains how the procedures maintain the researcher's flexibility.

Behavioral Interventions, 2022 · doi:10.1002/bin.1875