Scaling Up a Peer-Mediated Program Statewide: Lessons Learned Through Peer to Peer.

Matthews et al. (2022) tell the 20-year story of Peer to Peer in Michigan schools. They explain how one small program grew to over 700 schools.

The paper is a narrative review, not an experiment. It lists what helped the program spread and what still blocks the way.

The team found three big levers: get principals on board first, train teachers to train other teachers, and fold peer support into everyday class life.

No test scores or social-skills data are given. The paper focuses on lessons from running the program, not on child outcomes.

Older single-case studies like Raslear et al. (1992) and Lowe et al. (1995) showed peer methods work. Amy et al. take the next step and show how to move those same ideas across an entire state.

Laermans et al. (2025) extends the logic downward. They had preschool teachers run Stay Play Talk without outside coaches and still saw big gains. This fits the train-the-trainer model Amy et al. praise.

Stephens et al. (2018) ran a small RCT on peer-mediated PEERS and saw only modest effects. That looks like a clash, but the difference is scope: L et al. tested one curriculum in a few classes; Amy et al. spread a flexible framework statewide. Small, tight trials and large, loose roll-outs can both be right.

If you are a BCBA helping schools adopt peer programs, copy Michigan’s moves: secure admin buy-in first, use train-the-trainer, and embed peers in daily routines instead of running separate clubs. Start with one willing school, document quick wins, then let principals sell the idea to neighboring schools. This roadmap saves you from piloting in isolation and boosts the odds the program survives staff turnover.