FSM Pupils Excluded 5x More: The Behaviour Evidence Gap

Everyone knows the struggling kid needs one-to-one — except the evidence isn't sure. We read the EEF Toolkit's behaviour interventions strand (83 studies, updated May 2025): +3 months' additional progress on average, with the honesty labels attached — an average, at moderate evidence security. What carried the biggest impacts: approaches with an academic component or built on role-play and rehearsal, run frequently over a sustained period — and large-group or whole-class approaches slightly outperformed one-to-one. Plus the starkest line in the strand: pupils on free school meals are suspended at more than four times the rate of their peers and permanently excluded at around five times, most commonly for persistent disruptive behaviour. Behaviour support is disadvantage policy.

Printable one-pager for this episode, free: https://theeducationcommute.co.uk/#newsletter-staff

Two free newsletters, one for staff and one for parents and carers, same evidence written for the person reading it.

Sources:

• Behaviour interventions — EEF Teaching & Learning Toolkit (updated May 2025; 83 studies) — https://educationendowmentfoundation.org.uk/education-evidence/teaching-learning-toolkit/behaviour-interventions
• - FSM suspension/exclusion figures: DfE data as recorded on the EEF strand page (attributed that way in the episode)
• - Companion episode: Ep4 (First 20 Days) covers the EEF guidance report — the how-to; this episode covers the Toolkit strand — the how-much.

Disclosure: both voices on this show are synthesised (NotebookLM, Google) — the research, reading and editorial judgement are done by a serving practitioner.

Actions for consideration:

• Pull your suspension/exclusion data split by pupil premium this week — know the FSM gap before a governor or LA officer finds it for you.
• Favour interventions with an academic component or role-play/rehearsal, run frequently over weeks — not one-off assemblies or theatre visits.
• Reconsider automatic one-to-one placement for struggling pupils — groups of six or more slightly outperformed it; be ready to explain that choice to parents.
• Check whether a behaviour intervention is displacing more learning time than the disruption it's meant to fix.
• Choose evaluated programmes with real training over an unevaluated binder skimmed at inset.