How Much Active Learning? A Rapid Systematic Review of Contextual Moderators and an Evidence-Weighted Dosage Model for Political Science Teaching

Meta-analyses agree that active learning raises student achievement in higher education by roughly half a standard deviation. They say much less about how much active learning a particular course should contain. An earlier proposal, the AL∗ formula, answered this question with a linear index of six course characteristics whose weights were set by narrative judgement. This article replaces those weights with a transparent derivation from the published evidence. A rapid systematic review of 58 studies asks, for each of the six inputs (class size, course level, cognitive objectives, programme mix, year mix, attendance policy), whether the factor changes the effect of active learning, in which direction, and with how much certainty. Three findings stand out. First, the dose-response relation appears to flatten: mid- and high-intensity active learning produce similar achievement gains, both far above low intensity, while narrowing of achievement gaps has been documented only at high intensity. Second, effects are larger in small classes. Third, two inputs, year mix and attendance policy, have almost no evidence on dose and act instead on the choice of method. The article therefore proposes a two-layer model. A dose layer, AL∗ = 0.43 − 0.13N + 0.07L + 0.13O + 0.07Hc, keeps recommendations between 30% and 70% of class time, with coefficients computed from the direction and certainty of the evidence. A format layer uses class size, attendance and student composition to select methods. Six propositions are stated for empirical testing.