We Measure What Changes: A proportionate evidence standard for learning and skills programmes delivered by small providers in low-resource settings

Purpose: Funders, governments and employers increasingly ask training providers to show that their programmes work, yet the evidence base in education technology is thin and often supplied by vendors, and most workforce training evaluation stops at participant satisfaction. This paper proposes a proportionate evidence standard that a small provider in a low-resource setting can meet. Design: A theory-building narrative review integrates the four levels of training evaluation with the evidence tiers used in United States education law and tests the resulting framework against experimental and quasi-experimental evidence from Kenya, Sub-Saharan Africa and other low- and middle-income countries. Findings: Experimental evidence consistently favours instruction matched to the learner's level and sustained, in-classroom support for teachers; effects of technology depend on the pedagogy it carries; teacher knowledge is a binding constraint in the region; and effect sizes vary with context and typically shrink at scale. The reviewed literatures on training transfer and publication bias show that reaction data are weak evidence of impact and that unreported null results distort the evidence base. Implications: Small providers can meet credible standards without large budgets by publishing a logic model, pre-specifying outcomes, using comparison groups where feasible, engaging an independent evaluator, disaggregating results, reporting costs, and publishing null findings. A four-rung evidence ladder, a twelve-item minimum reporting standard and six testable propositions are offered. Originality: The paper adapts frameworks developed for large organisations and national systems into a graduated standard for small providers, and applies it reflexively.