What researchers studied

Meta-analysis extending an earlier review with six additional years of evidence, synthesizing 841 effect sizes from 141 studies of educational apps targeting reading and math skills for children ages 3 to 9. 141 studies evaluating educational apps for children ages 3 to 9, yielding 841 effect sizes across reading and mathematics outcomes

What they found

  • The pooled effect of educational apps was moderately positive at d = 0.40, with substantial heterogeneity across studies.
  • Effect sizes were larger on researcher-developed outcome measures and on outcomes measuring constrained skills, suggesting that how learning is measured materially affects how strong an app appears.
  • Randomized controlled trials and studies with larger samples produced smaller effect sizes on average than less stringent or smaller studies.
  • App dosage and app quality ratings did not predict effect size in the meta-analysis, and evidence about effectiveness at scale and over long-term follow-up remains limited.

What the study does not prove

  • This is a 2026 EdWorkingPaper and should not be described as a peer-reviewed journal article unless a later publication is independently verified.
  • The pooled estimate combines many different apps, study designs, skills, measures, and implementation settings, so d = 0.40 should not be treated as the expected effect of any specific app.
  • Substantial heterogeneity means average effects mask large variation across products and contexts.
  • The authors explicitly caution that evidence at scale and with long-term follow-up remains limited, so short-term gains should not be equated with durable academic improvement.

Evidence strength: Strong synthesis of a large evidence base; 141-study meta-analysis, with important heterogeneity and measurement caveats; EdWorkingPaper.

Why this matters for families

Educational apps can be useful, but families should be cautious about taking a headline effect size or product rating at face value. This 2026 meta-analysis found that the apparent size of learning gains depended partly on study rigor and how outcomes were measured.

Noor interpretation

How Noor translates the evidence into practice

The key Noor-relevant lesson is not that educational apps are ineffective. It is that evidence quality matters. Products can look more impressive when tested on narrow or researcher-created measures, while effects may shrink in larger and more rigorous studies. Technology should therefore be judged by independent outcomes, durable learning, and whether it improves understanding rather than by engagement metrics or promotional ratings alone.

When Noor Lyra Educators use or recommend learning technology, prioritize tools that complement instruction and have evidence from credible, independent measures. Do not equate time-on-app, star ratings, or polished interfaces with learning. Use apps selectively for practice, retrieval, fluency, or feedback, then verify transfer through student explanation and independent work.

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/d9yw-r226

Research notes

educational appslearning technologyreading appsmath appsmeta-analysisevidence qualityinstructional technology

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.