Evidence Based Education
Does Phonics Help Dyslexic Students?
Earlier this week, I saw an article claiming that phonics instruction did not help dyslexic students or older struggling readers. The authors cited the 2001NRP meta-analysis as proof for their claim. This was an interesting choice, as the NRP meta-analysis is usually cited as the best evidence for phonics and against whole language instruction. However, the NRP meta-analysis did show lower outcomes for older struggling students, and dyslexic students (as can be seen in the below infographic.) Something that I discussed with Dr. Timothy Shanahan in 2020 when I interviewed for the Pedagogy Non-Grata podcast.
Despite the fact that the NRP showed low outcomes for these two demographics, I am truthfully, more skeptical of these findings than any others from the paper. Firstly, the claim seems to lack a sort of intuitive logic. Phonics is a foundational skill, we know that teaching young readers phonics, greatly improves the rate at which they learn to read. It, therefore, makes rational sense that students at older ages, with similar skill deficits, would benefit from the same type of instruction, delivered more explicitly. Similarly, it also directly contradicts some of the most compelling anecdotal evidence I have heard. When I speak to individuals interested in the science of reading, the most common story I hear is that of a mother whose child had dyslexia and failed to learn how to read within the regular school system, until they were given intensive phonics instruction. Indeed, I have heard this story over and over again, from individuals who reached out to me to discuss the science of reading, after they found this blog. It seems to me that these parents have really been the driving force behind the SOR movement.
That said, anecdotal evidence is still anecdotal and should not be used, on its own, to discount scientific research. However, I did my own meta-analysis of phonics studies that was done 20 years after the NRP phonics meta-analysis and included almost twice the number of studies. Most of my results were statistically similar to the NRP paper, indeed, we even found the same overall average effect size. Interestingly, the only results of mine that significantly differed from the NRP paper, were on this topic. I found higher results for phonics studies on dyslexic students, and for older struggling readers, than I did for core instruction, as can be seen in the below infographic.
Moreover, not only did I find the highest overall effect sizes for phonics studies on dyslexic and older struggling readers, I also found the individual programs that showed the highest results for these demographics were also all very explicit programs that focused on teaching the connection between graphemes and phonemes. Indeed, the top 4 programs on average showed more than double the outcomes of the programs ranked 5th or lower.
While my meta-analysis was conducted 20 years more recently and included far more studies, it was also not published in a peer-reviewed journal (although I did have it reviewed by an independent research firm). Moreover, I lack the credibility and qualifications of those individuals that conducted the NRP meta-analysis. So I fully accept that some might prefer to take the findings of the NRP meta-analysis, over my own findings of this issue.
That said, there have been additional meta-analyses of this topic. In 2014 Galuschka, et al conducted a meta-analysis of randomnized control trial studies on treatments for dyslexic students. While this meta-analysis included far less total studies (22 studies and 49 comparisons) than my own or the NRP paper, the inclusion criteria was far more rigorous and included more total papers, for this specific topic. The results of this meta-analysis can be seen below.
Phonics instruction showed the second highest effect size in this meta-analysis and was the only intervention with enough studies to be statistically significant. Suggesting that at the time of this meta-analysis, phonics instruction was the most evidence-based practice for teaching dyslexic students. Now that said, the Galuschka and NRP meta-analysis did find the same effect size for teaching phonics to dyslexic students and this might suggest that my results were inflated. However, it’s also possible that my more modern inclusion criteria affected the results. My meta-analysis included 6 studies on dyslexic students published after the NRP paper and 4 studies published after the Galuschka paper. I also included 4 papers, on struggling readers written after the publication date of the two other meta-analyses.
Colby Hall, et al. published a meta-analysis on this topic, in 2022 on interventions for dyslexic students. Their paper looked at 53 experimental/quasi-experimental papers and coded for both random vs fixed effects. Their paper is the most in-depth analysis of the topic and in my opinion, is the most important paper on the science of reading, since the NRP report. While the authors did not look at the issue of phonics instruction, they did find the most valuable forms of instruction to be phonemic awareness and spelling. Both of these instructional methods also happen to relate to phonics, as phonics instruction is meant to teach the connection between letters and sounds.
When I read all of the above research, I interpret it to mean that there is a strong scientific argument for the use of phonics instruction with dyslexic students. Now, let’s be honest though, someone is going to read this and loudly exclaim “see this isn’t settled science, we need to take a holistic approach that includes all of the research”. However, some of the best evidence for phonics is actually what we don’t have evidence for. For example, I am not aware of any meta-analyses on the use of whole language or balanced literacy instruction for teaching dyslexic students. Moreover, I have never even found a single balanced literacy study that focused on an entirely dyslexic sample of students. My own meta-analysis looked at the use of balanced literacy instruction for at-risk/struggling readers and found results that were almost 3 times lower than phonics instruction for at-risk/struggling readers.
I am only aware of a single Whole Language study on dyslexic students. The experimental study was conducted by Mayfield, et al in 2000, on 60 grade 1 dyslexic students. The study found an effect size of .07 on decoding outcomes, which is statistically negligible. Similarly, to the best of my knowledge, we don’t have a single meta-analysis on the use of three-cueing, let alone for dyslexic students. So it seems that whole language, balanced literacy, or interventions that use three cueing, cannot be viewed as evidence-based alternatives to phonics for dyslexic students.
What about comprehension instruction? I often see the argument made for comprehension instruction being more important than phonics instruction, because after all, we want students to be able to understand the text right, not just read the words? Except the Galuschka 2014 meta-analysis looked at this factor and found statistically negligible results for providing dyslexic students comprehension instruction. This makes sense when you look at the Won Lee, et al 2022 correlational meta-analysis, which showed comprehension skills, unlike phonological and morphological awareness skills, did not correlate with general reading ability. This is likely reflective of passage comprehension being a by-product of other skills. In other words, decoding, vocabulary, and fluency knowledge lead to comprehension knowledge, comprehension knowledge does not lead to decoding, vocabulary and fluency knowledge.
Speaking of morphology, could it be an alternative for teaching dyslexic students how to read? Bowers et al 2010 meta-analysis showed some very high outcomes for the use of morphology instruction with at-risk readers. However, Colby Hall looked at this in greater detail in 2022 and found no statistically significant difference between studies that included or did not include morphology/vocabulary, for dyslexic students. And while previous meta-analyses on vocabulary instruction have shown really high impacts for dyslexic and struggling readers. It is important to note that these studies did not control for standardized vs non-standardized assessments, this is a big caveat because vocabulary studies that use non-standardized assessments tend to show massively inflated results.
The only other form of reading instruction that I am aware of, which has shown high-yield results for dyslexic students is repeated reading. In 2004 Therrien, et al conducted a meta-analysis on 18 experimental repeated reading studies, for grades k-12. They found a mean effect size of .60 for learning-disabled readers (on transfer assessments). Of course, this is phenomenal. But one cannot just do repeated readings for an entire intervention session, as such a task would get tremendously boring, tremendously fast.
According to the Oxford English dictionary, there are 171,476 words in the English language. There is just no way students diagnosed with learning disabilities could be expected to just memorize the spelling of the English language. Phonics, phonemic awareness, spelling, and repeated reading instruction are the most evidence-based instructional strategies for teaching dyslexic students. This is highly logical when we realize that phonological processing deficits are theorized to be the most common cause of dyslexia. To the best of my knowledge, there exists no meaningful evidence that whole-language or balanced literacy interventions are superior to phonics-focused interventions for dyslexic students. While other forms of instruction like comprehension, morphology, and vocabulary, likely still have value and deserve their place within instruction for dyslexic students, they appear less impactful than phonics, phonemic awareness, and fluency instruction.
Written by Nathaniel Hansford
Last edited, 30/10/2022
References:
Bowers, Peter & Kirby, John & Deacon, Hélène. (2010). The Effects of Morphological Instruction on Literacy Skills: A Systematic Review of the Literature. Review of Educational Research - REV EDUC RES. 80. 144-179. 10.3102/0034654309359353.
Galuschka K, Ise E, Krick K, Schulte-Körne G (2014) Effectiveness of Treatment Approaches for Children and Adolescents with Reading Disabilities: A Meta-Analysis of Randomized Controlled Trials. PLoS ONE 9(2): e89900. https://doi.org/10.1371/journal.pone.0089900
Hall, C., Dahl-Leonard, K., Cho, E., Solari, E.J., Capin, P., Conner, C.L., Henry, A.R., Cook, L., Hayes, L., Vargas, I., Richmond, C.L. and Kehoe, K.F. (2022), Forty Years of Reading Intervention Research for Elementary Students with or at Risk for Dyslexia: A Systematic Review and Meta-Analysis. Read Res Q. https://doi.org/10.1002/rrq.477
Therrien, William. (2004). Fluency and Comprehension Gains as a Result of Repeated Reading: A Meta-Analysis. Remedial and Special Education - REM SPEC EDUC. 25. 252-261. 10.1177/07419325040250040801.
Lee, J. won, Wolters, A., & Grace Kim, Y.-S. (2022). The Relations of Morphological Awareness with Language and Literacy Skills Vary Depending on Orthographic Depth and Nature of Morphological Awareness. Review of Educational Research, 0(0). https://doi.org/10.3102/00346543221123816
-NRP. (2001). Teaching Children to Read: An Evidence Based Assessment of the Scientific Literature on Reading Instruction. United States Government. Retrieved from <https://www.nichd.nih.gov/sites/default/files/publications/pubs/nrp/Documents/report.pdf>.
N, Hansford. (2022). Confirming the NRP Findings in 2022. Teaching by Science. Retrieved from <https://www.teachingbyscience.com/_files/ugd/237d54_2fc4b30fff474501944754c178c79b43.pdf>.
Mayfield, L. G. (2000). The effects of structured one-on-one tutoring in sight word recognition of first-grade students at-risk for reading failure. Louisiana Tech University.