The Causal Role of Phoneme Awareness and Letter-Sound Knowledge in Learning to Read: Combining Intervention Studies With Mediation Analyses
Hulme, C., Bowyer-Crane, C., Carroll, J. M., Duff, F. J., & Snowling, M. J. · 2012
grade Brctdeveloper-ledunreplicated
Sample
152 children, mean age 5;0, 19 schools
Population
5-year-olds at UK school entry with poor verbal ability (at-risk).
Design
Mediation reanalysis of the Bowyer-Crane 2008 RCT. Randomized Phonology+Reading (letter-sound + oral PA + guided book reading) vs Oral-Language control; 20 weeks; 5-month follow-up. Caveat: the PA-vs-letters mediator split is measured, not independently randomized, so that decomposition is non-experimental.
Key findings
A randomized PA+letters+reading bundle produced a durable word-reading/spelling gain (d=0.49 at 5-mo follow-up) fully mediated by end-of-treatment gains in PA and letter-sound knowledge, with no residual direct effect. Supports PA and letter-sounds as JOINT causal inputs - but cannot show PA-alone works.
Genetic confound
RCT controls selection; strongest causal design in the PA cluster.
Replication notes
The bundle (PA+letters+reading) cannot isolate PA-alone; both skills appear jointly necessary for the alphabetic principle.
DOI / URL
10.1177/0956797611435921
Effects
| Outcome | Metric | Value | Measure | Timing | Vs | Horizon | Class |
|---|---|---|---|---|---|---|---|
| Word-level literacy composite (P+R vs control) | d | 0.49 (0.27-0.71) | mixed | 5mo follow-up | active-alternative | under-1yr | domain-skill |
| Mediated path via phoneme awareness | SD | 0.31 | mixed | 5mo follow-up | active-alternative | under-1yr | domain-skill |
| Mediated path via letter-sound knowledge | SD | 0.09 | mixed | 5mo follow-up | active-alternative | under-1yr | domain-skill |
Cited by
- Phonemic awareness training (and whether it needs letters)moderate supportconf: mediumgc: low