Ghana’s targeted instruction scale-up needs more than a training count

A programme can expand while a practice remains difficult to establish in classrooms. That is the tension in the World Bank’s account of targeted instruction in Ghana. As of September 2026, improved training had reached all grade 2–6 teachers in 10,000 schools, benefiting more than 2.9 million children, according to the World Bank. Yet its implementation survey found that only 32% of teachers grouped students by learning level daily, despite most teachers reporting that they followed the programme.

The important finding is not that training failed. It is that a measure of reach and a measure of routine practice answer different questions. For monitoring-and-evaluation teams, this is where implementation data becomes useful: it can reveal a gap while there is still an opportunity to change how a programme is supported. Reporting how far a programme has spread cannot, by itself, establish what happens during instruction.

Ask what teachers need to do the work

The survey offered a possible direction for that change. According to the World Bank, 84% of surveyed teachers identified more school-level support as the most important aid to implementation. The World Bank also reported that its team worked with the Ghana Education Service and National Teaching Council on a structured headteacher checklist covering follow-up on practices including student grouping. The response was more specific than another reminder to follow the programme: it gave school leaders a way to attend to the practice the survey had exposed.

That sequence matters for research design. A useful implementation survey should not end at a compliance score. It should help a team form a testable account of what may be preventing a practice from taking hold, then identify an adjustment that can be examined. Teacher reports are valuable here as evidence about perceived support needs; they should not be treated as direct observations of classroom behaviour.

Keep the measures distinct

The adjustment produced evidence on practice, not yet on learning. The What Works Hub for Global Education reported that an A/B test found the checklist increased the frequency of grouping students within Differentiated Learning lessons by 15 percentage points. That is a meaningful implementation result because grouping was the practice under scrutiny. It does not establish that pupils learned more. The World Bank said future research rounds would assess effects on learning outcomes.

Other monitoring evidence makes the measurement problem sharper. In a separate analysis, the What Works Hub said Ghana’s DL+ system had recorded more than 8,300 classroom observations over two school years. Yet grouping learners by level appeared in fewer than 1% of observation comments describing teacher strengths, (the Hub found). That figure concerns what observers wrote in a particular comment field; it is not an estimate of how often teachers grouped learners. For data teams, the distinction is crucial. A sparse mention might prompt a review of practice, observation prompts or coding, but the text alone cannot decide which needs attention.

Build a chain of evidence, not a single verdict

Ghana also has a separate route for examining outcomes. A UNICEF case study described a randomized evaluation of DL+ involving 320 primary schools—160 DL+ and 160 comparison schools—tracked over three years. That evaluation should not be conflated with the checklist test. They address related but different questions: whether the wider approach affects learning, and whether a particular form of school-level follow-up changes a targeted classroom practice.

The lesson for careful research teams is to preserve those distinctions from data collection through reporting. Track reach so decision-makers know where the programme is operating. Ask teachers about barriers, and observe practice with instruments that make the intended behaviour visible. Test a plausible response, then assess learning separately rather than treating a change in implementation as proof of impact. The World Bank’s Ghana findings are a case for monitoring that helps teams make a decision—and for being explicit about what the evidence cannot yet say.