What researchers studied

Systematic review and meta-analysis of randomized controlled educational interventions with adult follow-up outcomes. 29 randomized intervention groups reporting 179 posttest effects and 497 adult effects for 26,076 participants; adult outcomes were measured an average of 14 years later, with a range of 4 to 49 years

What they found

  • Average intervention-level effects declined from 0.14 standard deviations at posttest to 0.07 after six to twelve months and 0.06 after one to two years.
  • A small positive average adult effect of 0.04 standard deviations remained across the included interventions.
  • Immediate posttest skill impacts only weakly predicted adult outcomes, while medium-term impacts were more predictive.
  • The evidence did not support a widespread pattern in which effects disappeared and then broadly re-emerged in adulthood.

What the study does not prove

  • The review included only a small and selected set of randomized studies with adult follow-up data.
  • Programs and adult outcomes were highly diverse, so pooled effects should not be treated as one universal intervention result.
  • The review was not preregistered and did not conduct a formal study-level risk-of-bias assessment.
  • Publication-bias adjustments substantially reduced some pooled estimates, and some relationships were sensitive to individual studies.
  • The results do not show that every faded short-term gain produces a later benefit.

Evidence strength: Systematic review and meta-analysis of randomized controlled trials; working paper.

Why this matters for families

Ask not only whether the score moved today, but whether the learning still works weeks later and in the next course.

Noor interpretation

How Noor translates the evidence into practice

A large immediate score change is not the only meaningful signal, and a smaller visible effect later does not automatically mean support was wasted. Lasting value is better judged through retention, transfer, independence, and how the learner functions when the original support is no longer present.

Noor Lyra should measure what survives beyond the lesson: whether students can retrieve the skill later, apply it in a new format, explain their reasoning, and work with less prompting.

Read the original source

Noor links to the original or authoritative source so families can distinguish the evidence itself from our interpretation.

Open original source →

DOI: 10.26300/dtk5-ds56

Research notes

fadeoutrandomized trialslong-term outcomeslearning scienceeducational interventions

No single study determines a student's plan. Noor uses research as one input alongside the learner's goals, observed performance, academic context, and response to instruction.