A primer for using multilevel models to meta‐analyze single case design data with AB phases

Becraft et al. wrote a how-to guide. It shows BCBAs how to pool many AB studies into one number.

The authors use multilevel models. These models keep each client’s trend while still giving an average effect.

They walk readers through a real example. The data come from 21 brief differential-reinforcement studies.

The primer works. Readers who follow the steps can turn separate line graphs into one effect size.

The example run took minutes. The pooled effect was large and precise.

Cook et al. (2020) also help single-case work. They teach how to stop momentary time-sampling error. Use both papers: collect clean data first, then meta-analyze it.

Jessel et al. (2020) warn that short sessions can hide control. Becraft’s tool needs good data, so keep sessions long enough before you pool them.

Jessel et al. (2016) reviewed DRO the old way—story style. Becraft shows the new way—numbers style. The new guide does not replace the review; it upgrades it from story to stats.

You can now answer “Does this intervention work across all my cases?” in one number. Download the free code, plug in your AB graphs, and get an effect size you can show parents, teachers, or funders.