This meta-analysis provides the first quantitative synthesis of how student engagement changes over the course of adolescence, drawing together evidence that previously existed only as individual studies with varied and sometimes conflicting findings. The theoretical anchor is stage-environment fit theory, developed by Eccles and colleagues since the 1980s, which proposes that engagement declines because school environments increasingly fail to meet adolescents’ developing psychological needs for autonomy, competence and meaningful relationships — particularly at points of school transition.
The search identified 125 peer-reviewed longitudinal studies published between 2006 and 2022, spanning samples from North America, Europe, Asia, Australia and beyond. Studies had to include repeated measurements of student engagement across the adolescent age range (10–18 years) at intervals of at least days to weeks. From these 125 studies the team extracted 223 separate repeated engagement measurements, yielding 544 individual effect sizes computed using Glass’s Δ. The dataset was tested for publication bias through multiple methods and found to be unbiased with high evidential value.
The headline finding is clear: engagement declines across adolescence. The overall effect size is small (Δ = −0.09, p < 0.001) across an average interval of 1.4 years, meaning that on average student engagement falls by roughly one-tenth of a standard deviation per measurement interval. The authors are careful to note that this small average likely conceals considerable heterogeneity — most students have gently declining or stable trajectories, while smaller subgroups with steep declines pull the mean downward. The result is nonetheless statistically robust and consistent across the full international dataset.
Five moderator analyses sharpen the picture considerably. First, the type of engagement dimension — behavioural, emotional, cognitive or mixed — made no significant difference to the rate of decline. All four fell at a similar pace, which is a notable finding given prior single-sample studies suggesting emotional engagement is most vulnerable. Second, school grade at first measurement mattered: decline was steeper in early adolescence and became less pronounced at higher grades, meaning the most critical window for disengagement is the early secondary years. Third, school transition significantly accelerated decline. Students moving from primary to lower secondary schooling showed declines nearly twice as steep (b = −0.15) as those without a transition, and the same was true for the transition from lower to upper secondary (b = −0.14). This directly supports stage-environment fit theory and flags school transition as a high-risk moment. Fourth, geographic area was a moderator, with European and North American samples showing steeper declines than Asian and Antipodean samples — though the authors note this may partly reflect unequal study representation across regions, and that sociocultural factors likely play a role. Fifth, gender had no significant effect on engagement trajectories, which challenges assumptions that boys and girls disengage at different rates.
The paper also confirms that the pattern of decline is evident whether engagement is self-reported or teacher-rated, and that it holds across studies with validated and non-validated measures — lending the findings additional robustness.
The implications the authors draw are direct: declines in engagement are normative across adolescence, not an artefact of particular studies or populations, and they are deepened by school transition. Educators and policymakers should pay particular attention to the early secondary years as a period requiring targeted support, and school transitions should be treated as structural moments of heightened disengagement risk rather than neutral administrative events.