Fadeout and Persistence of Intervention Impacts on Social-Emotional and Cognitive Skills in Children and Adolescents
Hart, E. R., Bailey, D. H., Luo, S., Sengupta, P., & Watts, T. W. · 2024
grade Bmeta-analysisindependentunreplicated
Sample
86 RCTs, 56,662 participants, 450 outcomes
Population
Child and adolescent intervention RCTs across cognitive and social-emotional domains.
Design
Models posttest-to-follow-up persistence separately for cognitive versus social-emotional skills at 6-12 month and 1-2 year horizons. RCT-only.
Key findings
Fadeout is pervasive but not uniform, and two widely held beliefs are wrong: social-emotional skills are NOT more durable than cognitive ones, and BIGGER short-run effects fade MORE. The latter is a direct warning against buying interventions on the strength of impressive end-of-treatment numbers.
Genetic confound
RCT base.
Replication notes
The most current and best-powered statement of the fadeout base rate; supersedes Bailey 2017 on magnitude questions.
DOI / URL
10.1037/bul0000450
Effects
| Outcome | Metric | Value | Measure | Timing | Vs | Horizon | Class |
|---|---|---|---|---|---|---|---|
| Cognitive vs social-emotional persistence at 6-12 months | predictiveness | equally predictive | mixed | grade-3 | unclear | under-1yr | domain-skill |
| Cognitive vs social-emotional persistence at 1-2 years | persistence | cognitive skills slightly MORE persistent — contradicting the common assumption that non-cognitive skills are stickier | mixed | grade-3 | unclear | 1-2yr | domain-skill |
| Relationship between initial impact size and persistence | relation | larger posttest impacts fade MORE (inverse relation) | mixed | grade-3 | unclear | unclear | domain-skill |
Cited by
- Fadeout — why early gains disappear, and what actually persistsstrong supportconf: highgc: low