Research Library · Study page
Is the proof for phonics weaker than it looks? A 2020 paper, explained
One man reread 12 reviews of phonics. He found "little or no evidence" that planned phonics beats the usual ways to teach. Others disagree.
Study Facts
Bowers, 2020
Critique of 12 earlier reviews
- Question
- Is the proof for planned phonics as strong as people say?
What it found
"Little or no evidence" that planned phonics beats the usual methods. He calls it "a draw".
EvidenceExpert view
One researcher's reading of other reviews. Three researchers replied and disagree.
- What it is
- 12 earlier reviews
- Looked at
- What each review compared and found
- By
- 1 researcher, University of Bristol, UK
- Paid by
- Not stated
- Conflicts
- Not stated. He argues for another method
Jeffrey S. Bowers (2020). “Reconsidering the Evidence That Systematic Phonics Is More Effective Than Alternative Methods of Reading Instruction.” Educational Psychology Review, 32, 681–705. Read the free full text · Publisher's version
Free to read on the publisher's site. This is the copy we read.
Last reviewed November 17, 2026 · Used in: Phonics or whole language: which works better?
Go deeper
The study in eight short parts. Tap a part to open it.
1What is the study?
A paper by one researcher, Jeffrey Bowers, in the UK. It is not a new study of children. He reread 12 earlier reviews of phonics. He also looked at England's school results since phonics became the law there.
His question: is the proof that phonics in a planned order beats other ways of teaching reading as strong as people say?
2Why this matters to you
You may hear that the research on phonics is settled. This paper says it is not. Bowers writes that there "can be few areas in psychology" where researchers agree so firmly on a view so "at odds with available evidence".
Three researchers replied to him, and they disagree. Read their reply too.
3What was measured?
No children were tested. For each earlier review, he asked what it really compared, and what it found.
4Who was measured?
No one, directly. The paper measured no children of its own. It reads 12 earlier reviews, and the studies inside them. It also reads England's national test results.
5How was it measured?
- He reread the reviews, one by one. He did not pool the numbers into a new average.
- He looked at the comparison groups. Some got phonics with no set order. Others got no phonics at all.
- He read England's results. Phonics in a planned order has been the law in England's state primary schools since 2007. Since 2012, all children aged 5 to 6 take a phonics check. This part is his reading of national results, not an experiment.
6The results
These are his findings. Where he reports another review, it is as Bowers tells it. We did not read those reviews.
His main point. Whole language classes teach some phonics too, he notes, just not in a set order. The panel's comparison groups mixed classes with some phonics and classes with none, he argues. In most cases, he writes, the reviews did not test planned phonics against phonics with no set order. So "it is not possible to conclude" that planned phonics beats it, he argues.
The panel's test against whole language. The National Reading Panel's gain over whole language classes was smaller than its overall gain, but still beyond chance, he notes. He says none of those studies put kids into groups at random.
Older struggling readers. He calls one result "the most important limitation": the panel found that phonics did not help low-achieving readers above first grade beyond chance. More research "may show" it does help them, he adds. The panel had only eight comparisons for them.
A re-analysis, as he tells it. Other researchers (Camilli and others) went back over the panel's studies. As Bowers tells it, they found that planned phonics beat phonics with no set order. But the gain was about half the size the panel reported. A later version found a smaller gain that could be chance.
A review of fair tests only, as he tells it. A 2006 review (Torgerson and others) used only studies that put kids into groups at random. As he tells it, it found a gain for reading words correctly. But most of that gain came from one study, he argues. Gains for understanding text and for spelling could be chance.
England. The share of children meeting the standard on the phonics check rose from 58% in 2012 to 82% in 2018. In his reading, the only clear change is on the check itself, with "no discernable effects on reading more generally".
What he does not say. He does not say whole language is better. His results "should not be taken as an argument in support of" the other methods, he writes. He does not say letter sounds are unimportant. His findings "do not challenge the importance of learning" how letters link to sounds.
What he wants instead. He wants teachers and researchers to consider other ways to teach reading. One possibility, he writes, is to give more weight to how meaning shapes spelling, through the meaningful parts of words. He points to a method called structured word inquiry, which he has written about before.
7Why this matters to you, again
- The debate is real. One researcher says the case for planned phonics is overstated. Three researchers answered him and disagree.
- Neither side says skip letter sounds. He does not question that learning how letters link to sounds matters. His question is whether a planned order beats phonics with no set order.
- He calls the reading wars "a draw". He means he finds no clear winner. He does not say whole language won.
His own limits:
- His results are not an argument for whole language, he writes.
- Planned phonics may still work, he grants. The reviews may just fail to pick it up.
- Future research may show that phonics helps older struggling readers, he writes.
Also keep in mind that it is one researcher's reading of other reviews, not a new study. Three researchers replied and dispute his reading point by point.
8Who did the study, and who paid
Who did it. One researcher, Jeffrey S. Bowers, School of Experimental Psychology, University of Bristol, UK.
Who paid. Not stated. The paper has no funding statement.
Conflicts of interest. Not stated. He argues for another way to teach, structured word inquiry, and cites his own earlier papers on it.
Numbers that disagree. Some numbers in the paper's table do not match its text. So we print none of its gain sizes.
