Syntes: Camilli, Wolfe & Smith (2006) — systematic vs "some" phonics, d = 0.12, not significant, once language activities are modelled

Källa: https://doi.org/10.1086/509525 (The Elementary School Journal 107(1), 27–36)

Verified content

• [abstract] Re-examines the NRP phonics meta-analysis and Camilli et al. 2003, and reports "new analyses that substantially alter previous interpretations of the effect of systematic phonics instruction".
• [companion 2008, full text] Using the 2006 multilevel model (Table 1 of the 2006 article), systematic versus "some" phonics in the treatment group: d = 0.123, not significant (p = .183); tutoring 0.489; systematic versus some literacy (language) activities 0.210 in the treatment group and 0.403 when they occur in the comparison group (the authors' design-consistency argument).
• [companion 2008, full text] Stuebing, Barth, Cirino, Francis & Fletcher 2008 (Journal of Educational Psychology 100(1), 123–134, DOI 10.1037/0022-0663.100.1.123; abstract read) re-estimated the same contrast on Camilli's coding at d = 0.183 (p = .042) and concluded the re-analyses do not contradict the NRP's small-to-moderate effects; Camilli et al. reply that the comparison with a "critical competitor" (programmes that already teach some phonics) is the one that matters for policy.

Limits

• [assessment] The 2006 article was not read; numbers come from the authors' own 2008 account. Pooled reading outcomes; decoding and at-risk readers are not estimated separately. Whether the contrast is significant depends on the model (0.123, p = .183 vs 0.183, p = .042).

Role for C12

• [assessment] Held as open question 16 (2026-10-24), with Camilli et al. 2003 — not a narrowing of C12.

Added 2026-10-24 by the project (nightly attack run). [abstract] = published abstract; [companion 2008, full text] = read in Camilli, Kim & Vargas 2008; [assessment] = the project's own assessment.