Syntes: Camilli, Vargas & Yurecko (2003) — against phonics taught unsystematically, the systematic-phonics edge is about half the NRP figure

Källa: https://doi.org/10.14507/epaa.v11n15.2003 (Education Policy Analysis Archives 11(15))

Verified content

• [full text] Re-coded the National Reading Panel (2000) phonics studies (40 studies after corrections) for the degree of phonics in both the treatment and the comparison group (none or unknown, some, systematic), plus language-based activities, tutoring and other moderators, and re-ran weighted regressions.
• [full text] Systematic phonics versus programmes with less systematic ("some") phonics: d = .241 with one weighting (WGT1) and d = .188 with the other (WGT3); the NRP's comparable figure was d = .41. Individual tutoring d = .399 (WGT1) and .290 (WGT3); language-based activities in the treatment group d = .257 / .228.
• [full text] Outcomes were pooled across reading measures; the authors did not estimate effects per outcome category (too few studies per category) and found no significant residual differences between categories (F(11, 212) = .805, p = .635). Reader-ability category (normal vs at risk / learning disabled) was among the candidate moderators but is not reported as entering the model.
• [full text] Figure 2: in early grades systematic phonics outperformed typical phonics or no phonics; differences were small after about grade 3.
• [full text] Tunmer & Hoover (1993; at-risk first graders, Reading Recovery with versus without an added systematic phonics component): recomputed against standard Reading Recovery, Word ID d = −.12, spelling −.25, nonwords −.12, oral reading .12.

Limits

• [full text] Same pool of mostly small, older studies as the NRP; randomised and quasi-experimental designs mixed (no difference found between them).
• [assessment] Pooled outcomes, so the decoding-specific contrast for at-risk beginners is not estimated. The Reading Recovery comparison is one small study, and Reading Recovery releases children once they reach a set level, so equal end scores may partly reflect the programme's design; the original article was not read.

Role for C12

• [assessment] Held as open question 16 (2026-10-24) together with Camilli, Wolfe & Smith 2006 — not a narrowing of C12.

Added 2026-10-24 by the project (nightly attack run). [full text] = read in the full text; [assessment] = the project's own assessment.