In education and psychology, much value is placed on interventions that help children and adolescents develop, both in the short and, preferably, in the long term. But, recent studies point to a worrying phenomenon: the effects of these interventions often fade over time. This phenomenon, known as “fadeout,” raises important questions about the effectiveness of educational interventions and their ability to produce lasting changes in children’s development. A new meta-analysis examined this phenomenon. In particular, there are concerns about how this “fadeout” affects both cognitive skills (such as language, math and intelligence) and social-emotional skills (such as self-regulation and social interaction).

Hart and colleagues’ meta-analysis (2024) seeks to understand whether the effects of interventions on social-emotional skills persist longer than those on cognitive skills. This is important because social-emotional skills are often seen as crucial to long-term success in personal relationships and work.

In their meta-analysis, they examined 86 randomized controlled trials (RCTs) that collectively included 56,662 participants. These studies were selected from eight previously published meta-analyses and focused on interventions that included children from early childhood to adolescence. The researchers checked how an intervention’s impact at completion (post-test) could predict its long-term impact (follow-up).

One of the study’s most striking findings was that the impact of interventions on both cognitive and social-emotional skills tended to decrease over time. This was especially true for interventions that initially had large effects. Interestingly, the study found that effects on social-emotional skills did not last longer than those on cognitive skills. In some cases, cognitive skills seemed slightly more likely to be retained over the long term.

An unexpected finding was that interventions with more minor initial impacts often led to more lasting results. This suggests that interventions that produce modest but meaningful changes may be more effective in the long run than those with a large but short-lived impact.

The findings of this study have important implications for both the theory and practice of educational interventions. First, the results suggest that policymakers may need to temper their expectations about the long-term effects of interventions. While some interventions can undoubtedly have positive long-term results, these effects are unlikely to be due solely to the direct impact of the intervention on cognitive or social-emotional skills.

In addition, the results highlight the importance of ongoing support after the initial intervention. Since the effects of interventions can diminish over time, children and adolescents must continued opportunities to develop and reinforce the skills they have learned.

Abstract of the study:

Researchers and policymakers aspire for educational interventions to change children’s long-run developmental trajectories. However, intervention impacts on cognitive and achievement measures commonly fade over time. Less is known, although much is theorized, about socialemotional skill persistence. The current meta-analysis investigated whether intervention impacts on social-emotional skills demonstrated greater persistence than impacts on cognitive skills. We drew studies from eight pre-existing meta-analyses, generating a sample of 86 educational RCTs targeting children from infancy through adolescence, together involving 56,662 participants and 450 outcomes measured at post-test and at least one follow-up. Relying on a meta-regression approach for modeling persistence rates, we tested the extent to which post-test impact magnitudes predicted follow-up impact magnitudes. We found that post-test impacts were equally predictive of follow-up impacts for cognitive and social-emotional skills at 6- to 12- months follow-up, indicating similar conditional persistence rates across skill types. At 1- to 2- years follow-up, rates were lower and, if anything, cognitive skills showed greater conditional persistence than social-emotional skills. A small positive follow-up effect was observed, on average, beyond what was directly predicted by the post-test impact, indicating that interventions may have long-term effects that are not fully mediated by post-test effects. This pattern of results implied that smaller post-test impacts produced more persistent effects than larger post-test impacts, and social-emotional skill impacts were smaller, on average, than cognitive skill impacts. Considered as a whole, intervention impacts on both social-emotional and cognitive skills demonstrated fadeout, especially for interventions that produced larger initial effects. Implications for theory and future directions are discussed.