So often, when we look at meta-analysis, we examine these very broad ideas, and within these broad ideas, we can lose a lot of important nuances. When we look at phonics, for example, we often just explore the efficacy of phonics as a whole. It is important to do this because it tells us if phonics is overall a high yield strategy (which it clearly is). However, by focusing on this information, we can lose focus on the fact that phonics can be more efficaciously applied, in different individual circumstances and less efficaciously applied in others. I was recently reviewing a secondary meta-analysis of the NRP meta-analysis, by Linnea et al, on the topic of systematic phonics instruction, done in 2001. In this paper, she only looks at the effect sizes, calculated with a control group. I thought ranking and comparing the effect sizes from this data would be an excellent way to provide people with a better contextual understanding of when and how phonics instruction is most effective. Please see the graph below, as well as some contextual information about the implications of these effect sizes.


Just a reminder to the reader that generally speaking: an effect size above .80 is seen as high, between .79 and .50 is seen as moderate, between .50 and .20 is seen as low, and below .20 is seen as extremely low. On average we see effect sizes of .40 in education research. So any effect size above .40 can be seen as statistically more relevant. 

.Effect size comments:


  1. Kindergarten Decoding, Spelling and Comprehension: Overall, we see strong evidence from this meta-analysis that phonics instruction is most useful for kindergarten students, especially in terms of their decoding and spelling ability. This result would suggest that phonics instruction provides an extremely consistent benefit to kindergarten readers. Ultimately I believe that this effect size is conclusive proof that the DAP movement is wrong in its assertion that systematic reading instruction should not start until grade 1. 

  2. One of the most interesting discoveries I made, reviewing this literature, was the difference of effect sizes for readers from low economic backgrounds, compared to moderate economic backgrounds. It appears there is significant evidence to suggest that phonics instruction is far more important for impoverished students than for middle-class students. This might be because middle-class students often receive certain advantages in life and start school ahead as readers. Ultimately, I think this effect size shows that early phonics instruction is not just a matter of best practices, but of equity. 

  3. At-Risk Kindergarten Students: This data suggests that providing at-risk kindergarten students with phonics instructions is a moderate yield strategy. This data suggests, it is likely helpful to identify struggling readers early, so they can be provided with phonics instruction. 

  4. Delivered by a Tutor: This intervention clearly shows that one on one phonetic instruction is clearly superior to both small-group instruction, and classroom instruction. This is likely obvious. However, I believe that this effect size does lead credence to the argument that we should be providing at-risk readers with additional one on one phonics instruction. 

  5. Grade 1: The effect size of phonetic interventions for grade 1 students was high, albeit substantially lower than for kindergarten. Moreover, phonetic interventions for grades 2-6 were substantially less effective than for grade 1 students. This suggests that phonics becomes less and less effective, as students advance in their reading ability. Moreover, it also reinforces the idea that whole class phonetic instruction should start in Kindergarten and end in the junior grades. 

  6. Grades 2-6: Decoding, Spelling, and Comprehension: This effect size demonstrates that phonics instruction provides a moderate result for spelling/decoding outcomes and a low result for comprehension outcomes in grades 2-6. However, this analysis shows that there is clearly a diminishing return to whole class phonics instruction as students get older. Given that this effect size covers grades 2-6, it is safe to assume that the effect size is much higher on the lower end of this spectrum and that the reverse is true for the higher end of this spectrum. Dr. Timmothy Shanahan has confirmed as much in a previous interview on Pedagogy Non-Grata. 

  7. Direct Instruction: This effect size showed that teaching students about phonics, using direct instruction, had a moderate result. Phonics is not an inquiry skill, it is a knowledge-based skill. Considering that direct instruction provides superior results for knowledge outcomes, (as shown, by other previously reference meta-analysis evidence) I think it makes sense that direct instruction should be used for phonetic instruction, especially within the early grades. 

  8. Synthetic Phonics Compared to Unsystematic Phonics: This effect size is based on comparing synthetic phonics to other high yield strategies, so while it appears to be a moderate yield, it should be interpreted as evidence that synthetic phonics is a high yield strategy. 

  9. Small-Group Instruction: This effect size showed that phonics in a small group setting had greater results than in a classroom setting, but lower results than in a one-on-one setting.

  10. Reading Disabled: This effect size showed that providing phonetic instruction to students diagnosed with a reading disability was on average a moderate yield strategy. 

  11. Phonics Interventions Compared to Whole Language Interventions: This effect size looked at studies, in which the control group used a Whole Language approach. In many ways, this is the most important effect size on this list. Whole Language and Phonics are competing ideas, this means that the most academically honest way to evaluate the efficacy of either approach would be through a comparative study. This effect size clearly demonstrates (again) that Whole Language approaches are inferior to phonics ones. 

  12. Phonics Interventions Compared to no Specific Classroom Strategy: In this meta-analysis, phonics interventions were evaluated against control groups, in which teachers were allowed to teach as per normal. The effect size clearly indicates that phonics was a superior approach. 

  13. Moderate Economic Backgrounds: This effect size shows the phonetic instruction is less effective for middle-class students, likely because middle-class students often start school ahead, academically and are past the emergent reader stage, which benefits from phonetic instruction. 



Written by Nathaniel Hansford

Last Edited on 06/20/21



Linnea, et al. (2001). Systematic Phonics Instruction Helps Students Learn to Read: Evidence From the National Reading Panel’s Meta-Analysis. Ontario Institute for Studies in Education. Retrieved from <>.