How Effective Are One-to-One Tutoring Programs in Reading for At-Risk Students? A Meta-Analysis
Elbaum, B., Vaughn, S., Hughes, M. T., & Moody, S. W. · 2000
grade Cmeta-analysisindependentreplicated
Sample
29 studies / 42 samples / N=1,539
Population
US at-risk elementary readers, mostly grades 1-3.
Design
Older literature (1975-1998); mixed RCT/quasi.
Key findings
Anticipated two modern findings: 1:1 holds no advantage over small groups (undermining the costliest design choice in tutoring), and Reading Recovery's headline effects were inflated by analyzing only successful completers.
Genetic confound
The Reading Recovery 'discontinued' analysis conditions on post-treatment success — selecting on ability; the clearest ability-confound artifact in the tutoring literature.
Replication notes
The 1:1-vs-small-group null has replicated repeatedly (Nickow; Saga's 2:1/4:1). The completer-selection warning was validated by the later negative long-term RR findings.
DOI / URL
10.1037/0022-0663.92.4.605
Effects
| Outcome | Metric | Value | Measure | Timing | Vs | Horizon | Class |
|---|---|---|---|---|---|---|---|
| Reading, overall | ES | 0.41 | mixed | end of treatment | unclear | end-of-treatment | domain-skill |
| 1:1 vs small group | comparison | NO advantage for 1:1 over small-group supplemental instruction | mixed | end of treatment | active-alternative | end-of-treatment | domain-skill |
| Reading Recovery subgroup | artifact | large effects only for 'discontinued' (successful completer) students — completer-selection artifact | mixed | end of treatment | unclear | end-of-treatment | domain-skill |
Cited by
- Tutoring — the honest effect, the Bloom 2-sigma myth, and what survives scalestrong supportconf: highgc: low