Evidence-Based Teaching Strategies, Ranked by Real Impact
Education is full of confident advice. Very little of it comes with numbers. Here's what large-scale research actually says — including two results that surprise almost every teacher.
Every staff room has strong opinions about what makes teaching work: smaller classes, more homework, stricter discipline, smart boards. But for two decades, researchers have been measuring these things across thousands of classrooms — most famously the Education Endowment Foundation (EEF), whose Teaching & Learning Toolkit summarises evidence from millions of students into one simple unit: months of additional progress per year.
The ranking most schools have never seen
Education Endowment Foundation, Teaching & Learning Toolkit — averages across thousands of studies; individual results vary by implementation.
1. Metacognition: the +7-month strategy that costs almost nothing
The single highest-impact, lowest-cost approach on the list is teaching students to plan before a task ("what strategy will I use?"), monitor during it ("is this working?") and evaluate after ("what would I change next time?"). In practice this is as simple as ending a lesson with two questions: What did you find hardest today? What did you do about it? Students who reflect on their own thinking stop repeating the same mistakes — and the effect compounds all year.
2. Feedback: +6 months — but only a specific kind
Research is clear that feedback is among the most powerful tools a teacher has — worth roughly six additional months of progress. But the evidence favours feedback that is specific, early, and actionable: "your conclusion doesn't answer the question you set in your introduction" beats "good effort, 6/10" every time. One study found clear communication of goals had a 32% higher effect on performance than simply holding high expectations. Marks without commentary are the least useful form of feedback there is.
3. Structured group work and peer tutoring: +5 months each
Group work earns its reputation only when it's structured — defined roles, a task that genuinely requires talk, and individual accountability at the end. Unstructured "discuss in groups" mostly produces noise. Peer tutoring is the quiet star here: the tutor often gains as much as the tutee, because explaining a concept is a form of retrieval practice (more on that in our student guide).
What barely works — and what that means for your school
- Extending school time: +2 months, high cost. More hours of the same teaching produces little. Better teaching in existing hours produces much more.
- Technology by itself: +4 months, moderate cost. Devices amplify pedagogy; they are not pedagogy. A smart board running a lecture is still a lecture.
- Homework: it depends on age and design. Purposeful, feedback-linked homework helps secondary students; hours of repetitive work helps almost no one.