Systematic Phonics

There is little evidence that systematic phonics is better than whole language and other common methods used in schools (and no, this is not evidence for whole language)

The title of this post will no doubt be surprising to many.  Proponents of phonics can point to multiple meta-analyses and 1000s of papers that strongly endorse systematic phonics.  For example, the National Reading Panel or NRP (2000) that has been cited over 20,000 times concludes:

Students taught systematic phonics outperformed students who were taught a variety of  nonsystematic or non-phonics programs, including basal programs, whole language approaches, and whole word programs. (NRP, 2000, p. 2-134)”.

Similarly, The Rose Report that led to the legal requirement to teach systematic phonics in English state schools concludes:

Having considered a wide range of evidence, the review has concluded that the case for systematic phonic work is overwhelming …” (Rose, 2006, p. 20).

Strong claims are routinely made in academic articles, such as:

It should be clear that I am advocating here a strong ‘phonics’ approach to teaching, and against a whole-word or whole-language approach… theoretical and laboratory-based arguments converge with school-based studies that prove the inferiority of the whole-word approach in bringing about fast improvements in reading acquisition. (Dehaene, 2011, p. 26).

Similarly strong claims are made by leading researchers in popular books (e.g., Seidenberg, 2017; Willingham, 2017).  For example, in “Reading at the Speed of Sight: How We Read, Why So Many Can’t, and What Can Be Done About It”, Mark Seidenberg (2017) writes:

The phonological pathway requires knowing how print relates to sound, the focus on “phonics” instruction… For reading scientists the evidence that the phonological pathway is used in reading and especially important in beginning reading is about as close to conclusive as research on complex human behavior can get. (p. 124)

Given all this, how can I claim that there is little evidence in support of systematic phonics?  Why would I even question claim?

Why question the evidence for systematic phonics?
Let me start with the second question first: what motivated me to question this claim?  I started looking more carefully at evidence in response to the difficulties I had in publishing a paper co-authored with my brother that detailed an alternative to systematic phonics instruction called Structured Word Inquiry or SWI (Bowers and Bowers, 2017; see previous blogpost for brief summary of the main arguments in this paper).  I lost count of the number of places the paper was rejected, and what made the process of publishing this paper particularly frustrating is that reviewers repeatedly mischaracterized many of our claims.  For example, reviewers consistently criticized us for rejecting fundamental role that phonology plays in reading despite the fact we clearly wrote the opposite, as in the following quote:

SWI is motivated by a fundamental insight from linguistics, namely, the English spelling system makes sense when the sublexical constraints of morphology, etymology, and phonology are considered in combination… Consistent with phonics, SWI agrees that it is important to teach sublexical grapheme–phoneme correspondences, but it emphasizes that English spellings are organized around the interrelation of morphology, etymology, and phonology and that it is not possible to accurately characterize grapheme–phoneme correspondences in isolation of these other sublexical constraints

Our impression was that many reviewers did not understand our claims because they did not even try.  We still find it hard to get many researchers to engage with the ideas now that they are published.  Why? Because so many researchers assume that the case for phonics is settled — the “reading war” is over, phonics won, and it is almost irresponsible of us to question the efficacy of systematic phonics given the continuing resistance to phonics in many schools.

It was this absolute commitment to systematic phonics that led me to start looking more carefully at the evidence for phonics.  And with a critical eye, it does not take long to discover that the case for systematic phonics is weak at best.  The case is built on a house of cards, with 1000s of researches citing the NRP (2000), Rose (2006), and multiple additional meta-analyses without understanding what the reports actually show.  Indeed, the authors of these reports and meta-analyses often misunderstand their own findings.  The disconnect between the claims and evidence are so striking that I ended up writing the long and detailed paper you can download here.

What is wrong with the evidence with systematic phonics?
Briefly, I’ll summarize two key reasons why the current evidence does not support phonics. First, the findings reported in the meta-analyses are often poorly summarized in the abstracts and conclusions, with small and mixed results described as providing evidence for systematic phonics.  These conclusions are then echoed and amplified in subsequent papers that cite these reports.  However, as I detail in my review, if you carefully read the result sections (rather than the abstracts) of the meta-analyses, and accept the findings at face value, then there is no reason to be a strong advocate of systematic phonics.  A more appropriate conclusion would be something like this:  Systematic phonics supports small short-term effects on a pseudoword and word naming, the evidence for long-term effects on pseudoword and word naming is mixed, and that no evidence to suggest that systematic phonics has long-term effects on reading fluency, vocabulary, or reading comprehension (the aspects of reading we should care about most).   This is a far cry from the standard claim that the evidence for systematic phonics is overwhelming.

Second, and more importantly, you should not accept the findings at face value. There is also a fundamental conceptual problem with the design with most of the meta-analyses that undermines even these modest conclusions.  In two important but largely ignored papers, Camilli et al. (2003, 2006) showed that the NRP (2000) did not even test the hypothesis that systematic phonics is better than whole language and other common methods used in schools.  This motivated their new analyses, with Camilli et al. (2003) showing that the effect of systematic phonics is roughly half the size as reported by the NRP (2000), and with an improved analysis, Camilli et al. (2006) failed to observe a significant benefit of systematic compared to non-systematic phonics (as practiced in whole language for example).  In my review I show that this same flaw applies to all subsequent meta-analyses, as well as point out multiple additional problems that further undermine the conclusions that are commonly drawn.

What do I mean that the NRP (2000) and subsequent meta-analyses have not even compared systematic phonics to whole language?
What is this flaw that Camilli et al. (2003, 2006) identified?  The NRP (2000) assessed the following hypothesis:

…findings provided solid support for the conclusion that systematic phonics instruction makes a more significant contribution to children’s growth in reading than do alternative programs providing unsystematic or no phonics instruction [bold added] (NRP, 2000, p. 2-132).

The bold highlights the key point that systematic phonics was compared to control condition combined to separate conditions, namely (1) intervention studies that included unsystematic phonics and (2) intervention studies that included no phonics.  Why is this important? Within the NRP analysis, most whole language interventions included non-systematic phonics, and accordingly, it is possible that the advantage of systematic phonics in the NRP analyses was due to the poor performance in the non-phonics condition, with children in the systematic and non-systematic phonics conditions doing similarly.  Indeed, this motivated the Camilli et al. analyses, and when they compared systematic phonics compared to non-systematic phonics, the effects were greatly reduced (Camilli, 2003) or eliminated (Camilli et al., 2006).

This is not nitpicking.  The “reading wars” was about whether systematic phonics is better than whole language, and the research community overwhelmingly take the NRP (2000) and many subsequent meta-analyses as providing evidence in support of systematic phonics.  Indeed, this has led to the legal requirement in the UK to teach systematics phonics rather than whole language (and related methods) in English schools since 2007.  But given that whole language includes non-systematic phonics, and given that there is little or no evidence that systematic phonics is better than non-systematic phonics, the conclusion does not follow.  That is, the meta-analyses (other than Camilli et al., 2003, 2006) have not even tested the hypothesis that systematic phonics is better than whole language.

I expect many readers will claim that whole language rarely includes non-systematic phonics.  But here is the description of whole language as practiced in the USA from the NRP (2000):

Whereas in the 1960s, it would have been easy to find a 1st grade reading program without any phonics instruction, in the 1980s and 1990s this would be rare… Whole language teachers typically provide some instruction in phonics, usually as part of invented spelling activities or through the use of graphophonemic prompts during reading (Routman, 1996). However, their approach is to teach it unsystematically and incidentally in context as the need arises. The whole language approach regards letter-sound correspondences, referred to as graphophonemics, as just one of three cueing systems (the others being semantic/meaning cues and syntactic/language cues) that are used to read and write text. Whole language teachers believe that phonics instruction should be integrated into meaningful reading, writing, listening, and speaking activities and taught incidentally when they perceive it is needed. As children attempt to use written language for communication, they will discover naturally that they need to know about letter-sound relationships and how letters function in reading and writing. When this need becomes evident, teachers are expected to respond by providing the instruction. (p. 1-202)

Similarly, in the UK, prior to the mandatory requirement to teach systematic phonics in the UK, Her Majesty’s Inspectorate (1990) reported on the teaching and learning of reading observed in 470 classes and over 2,000 children.  They wrote:

…phonic skills were taught almost universally and usually to beneficial effect’’ (p. 2) and that ‘‘Successful teachers of reading and the majority of schools used a mix of methods each reinforcing the other as the children’s reading developed (p. 15).

So to summarize, apart from the Camilli (2003, 2006) meta-analyses, no meta-analysis even tested the hypothesis that systematic phonics is better than whole language.  When the hypothesis was tested, the benefits of systematic phonics was no longer significant.

My review identifies many additional problems with the evidence taken to support systematic phonics (the evidence for systematic phonics is even weaker than Camilli et al. claim), but I’ll leave this for readers interested enough to download my paper.

What not to conclude
I am NOT claiming that grapheme-phoneme correspondences are unimportant.  Indeed, Camilli’s reanalyzes of the NRP (2000) report showed that systematic phonics is better than a no phonics control condition.  In addition, my review should NOT be taken as an excuse to continue to teach whole language.  Camilli et al. (2006) and I have shown that there is no evidence that one approach is better than the other.  It turns out that the reading wars was a draw.  This contradicts the common view that the systematic phonics is best practice, but it does not provide any support for whole language.

What to conclude
I am claiming that researchers should be less committed to systematic phonics, and more open alternative methods. Indeed, this is the main message I want to get across with my detailed review of phonics – it is time to consider alternatives to whole language and systematic phonics

This brings me back to Structured Word Inquiry or SWI.  A key feature of SWI is that is systematically teaches grapheme-phoneme correspondences (something that proponents of phonics should like) but within the context of morphology and etymology.  As a consequence, SWI does a better job at describing grapheme-phoneme correspondences.  At the same time, SWI teaches children systematic spelling-meaning correspondences with the goal of making sense of the English spelling system, and highlights the meaningful relations between words (something that critics of phonics might like).  Bowers and Bowers (2017) emphasized the theoretical motivation for SWI, but we also reviewed some empirical evidence for this approach, including one study that showed the SWI was more effective than phonics in a randomized controlled study with 5 to 7 year-old children (Devonshire et al., 2013), and more generally, systematic reviews and meta-analyses found morphological instruction to be more effective in younger and struggling readers (Bowers et al., 2010; Carlisle, 2010; Goodwin et al., 2010, 2013).  There is no evidence for the claim that SWI or morphological instruction should only be introduced in older children after systematic phonics (e.g., Rastle, 2018).

To avoid any conclusion, I want to emphasize that more studies are needed to assess the efficacy of SWI when introduced at the start of instruction or later.  I hope a better appreciation of the (lack of) evidence for systematic phonics will help motivate more research.  Indeed, given the lack of evidence that systematic phonics is better than whole language, there is every reason to assess new methods, including the hypothesis that children should be taught how their writing system works.  For an easy introduction to SWI, I suggest going to the publication page on this website and reading Bowers and Bowes (in press), and for more detail, Bowers and Bowers (2017).  And if you want even more information, I suggest you go to my brothers’ website here.

Some references related to this topic

Any comments welcome!

34 thoughts on “Systematic Phonics

  1. Just getting introduced to your work and can see the value in it, though have a couple questions, if you have time.

    1) What about us perfectly okay spellers/learners who never had instruction through morphemes,etc. and have picked up the language just fine? While I’m sure it would deepen our understanding of the language, it is not totally necessary – hence is such an approach most beneficial for that 30% of students struggling with the language?
    2) What about the research on orthographic mapping, particularly some of the things David Kilpatrick is bringing out. He has pointed out that indeed phonics is not enough – as you critique – but major advances have come with heavy phonemic training in the most successful studies to date.
    3) Some spelling programs DO incorporate morphological and etymological details ( I primarily use Louisa Moats’ material), albeit more so starting around 2nd to 3rd grade. I understand your claim is to start early, or that you can start early – but are there issues you see in kids getting exposed just a little later?
    4) Going through your materials feels very random – or it is very random and I have always felt morphological knowledge to be this way from the materials I have looked at. Are there any succinct resources you might point teachers to, who we all know are limited on time, in order to continue study in this area?

    Thanks for your time.

    1. Hi Frank thanks for your interest. A few quick replies in order.

      1) Most children will learn to read and write well with whatever methods are used. So SWI is not needed for all children (thanks goodness as few people know about it!). But it may help children who struggle, and it might make learning for many children easier. It certainly will give all children better insights into how their writing system works.

      2) I just deny that the evidence supports the conclusion that systematic phonics or heavy phonics works better than other methods.

      3) It is an outstanding question when morphology should be taught. The only evidence to date suggests that morphology is useful from the start, and there is no evidence against this conclusion. But more data are needed. Regardless of when SWI should be introduced, the most important thing is that SWI uses tools like morphological matrices, word sums, etymological dictionaries to teach children (and adults) the logic of their writing system. This is not done by Moats or others.

      4) There are no standardized resources to use for SWI for now. For some it is important that teachers can be flexible and respond to questions that arise with students. I think this a great approach for skilled teachers who have a strong grasp of SWI, but I think it is important to have more structured forms of instruction so more teachers can adopt SWI more confidently. That is something I hope to contribute to in the future.

  2. Hi Harriet, although there is no evidence for your claim that SWI is too difficult (like you say that is an empirical question), there is evidence that learning and memory is better for meaningful things, and children find answers to the question “why” motivating. I think teachers/researchers are too quick to assume SWI is too difficult for beginning readers, especially as it is done in many schools in grade 1, and the only empirical study that exists found it more effective than phonics in Year 1. But it does require a lot of word for teachers to learn to teach SWI.

    As far as Dehaene goes, I already have a paper that claims that neuroscience has nothing to offer teaching. With at least one quote making a bit of fun of Dehaene. In case you are interested in this, go to my publications section of my webpage and download:

    Bowers, J. S. (2016). The Practical and principled problems with educational neuroscience. Psychological Review, 123, 600-612.

    1. Thanks, Jeff. I found the article fascinating. I’d like to read the article responding to your piece before I comment, but I’m unable to access it. Are you able to post the link?

        1. These two articles on EN really were fascinating. There are two parts that stood out based on my experiences working with over 500 struggling readers.

          The first is the statement that “rather than the common view that letters are primarily designed to represent sounds, English is a morphophonemic system in which spellings have evolved to represent an interrelation of morphology, etymology and phonology.” Jeff, you conclude that “this has obvious implications for literacy instruction given that learning and memory is better when information is encoded in a meaningful and well-organized manner (e.g., Bower, Clark, Lesgold, & Winzenz, 1969).”

          The first thing that struck me is this reference to a study done in 1969. Since the premise that morphology facilitates learning and memory is central to your recommendations, I was surprised that the study cited is 50 years old. Then when I looked it up, I was further surprised by this conclusion:

          “The message of these studies is simple: If S can discover or learn a simple rule or principle which characterizes the items on a list and which relates them to one another, then he uses that rule as a retrieval plan in reconstructing the items from memory, with a consequent improvement in his performance.”

          If the aim is to learn a “simple rule or principle” and use this rule “as a retrieval plan”, then I see a direct application to phonics teaching in beginning readers. Instead of having my kindergartners memorize a whole word like “play” or teaching them to simply spell out all four letters, I taught them the three phonemes and the “rule” that the two letters “a” and “y” represent a single sound. This facilitated their “learning and memory” enough for some of my students to apply it to their writing by independently composing the word “mayk”. I think this is the level of complexity appropriate for beginning readers as they make meaning out of the spelling variations for a given sound when they blend and segment. If this is how we can apply the term “meaningful”, then it very much describes the “meaning” my beginning readers take away from phonics.

          Secondly, the comments responding to your article refer to a direct application of EN to the development of the computer program Graphogame, which I am familiar with. The authors state:

          “Bowers (2016) interprets this statement negatively: ‘There is no indication how the neuroscience provides any additional insight into how instruction should be designed.’ However, as that paper also showed, EN researchers can build on this neuroscience finding, combined with psychological and educational findings on the development and training of number representations, to target the concepts learners lack. Specifically, the article cites two examples of interventions based explicitly on learning targets informed by neuroscience—Graphogame (Räsänen, Salminen, Wilson, Aunio, & Dehaene, 2009) and Number Race (which has been the subject of another successful trial more recently; Sella, Tressoldi, Lucangeli, & Zorzi, 2016).”

          I know that reading Dehaene’s book Reading in the Brain prior to teaching kindergarten did indeed allow me to “target the concepts learners lack”, and I think Seidenberg sheds light on this link between research and practice when he states in Language at the Speed of Sight:

          “The unresolved issues about reading education matter because instructional practices make a difference, affecting children’s proficiency but also whether they engage in the activity, enjoy it, or assign value to it. The reading wars and their aftermath also brought to light broader issues about education itself: the philosophical and empirical bases of educational theories and practices; relationships between education and cognate disciplines such as psychology and neuroscience; the functions of elementary education and how well they are being fulfilled; where change is needed, what to do, and who should decide; who goes into the profession and how they are trained. Disagreements about these issues are why debates about reading continue, despite exhaustion with the topic and pressures to move on.”

          Thanks again for this discussion. I look forward to following the continued debate.


  3. I would like to thank Pete and Jeff for introducing me to Anne Castles’ research–and also thank the three of you for this discussion. Those of us in the trenches depend on those of you in the (ivory) towers to digest and disseminate the relevant research so that we can apply it in the classroom. I have been rereading Suggate (2016) and Blachman (2014) and have been trying to reconcile their conclusions. Suggate states that “the current analysis would suggest that preschool and kindergarten interventions would target phonemic awareness alone, leaving decoding skills to Grades 1 and 2”, which are exactly the grades of the struggling readers I work with on decoding skills. Of most interest is the statement that “a key and unique finding from this meta-analysis is the greater retention of intervention effect to follow-up for at-risk, low, and disabled readers in compared to normal readers. This finding is certainly encouraging for interventionists targeting struggling readers, suggesting that promising long-term effects are attainable.”

    Since I am one of those interventionists, I am certainly encouraged. As for Blachman (2014) which looked at the effects of interventions with 2nd and 3rd graders 8 years later, the conclusion states that “ideally, one would want to build on the initial large effects seen immediately posttreatment on word recognition, reading rate, spelling, and passage reading (with respective effect sizes of 1.69, .96, 1.13, and .78), by providing the kind of extended instruction that would facilitate an accelerated growth rate over time, especially in fluency (automaticity) and comprehension. To close the achievement gap between struggling readers and typical readers, more extensive efforts are clearly required.”

    Although I primarily work with struggling first and second grade students, I also spend time each week in 3rd-6th grade classrooms teaching comprehension strategies and often encounter in a general classroom setting students I have in previous years worked with in small group intervention. Based on my having emphasized decoding skills in these small 1st and 2nd grade groups and now comprehension strategies in grades 3rd-6th whole class, I would say that teaching phonics to struggling students (especially those caught in the achievement gap) is necessary but not sufficient to see longterm gains in reading.

    So I would agree with the conclusions of Fiona J. Duff and Margaret J. Snowling in their chapter “Reading Disorders”, the one after Ann’s in The Cambridge Handbook of Child Language: “This, and other findings, caution against viewing one-off interventions as being fully able to remediate dyslexic difficulties. More realistically, those with dyslexia, especially where it is severe, are going to need early and ongoing additional support.”

    Phonics: necessary but not sufficient.

    1. Thanks Harriett for your thoughts. The main point from the Suggate meta-analysis is that the long-term benefits of systematic phonics instruction are small, and less than alternative methods. This goes against the standard view that it is important to teach systematic phonics first (the view of Anne Castles and colleagues).

      But the main thing I want to respond to is your final point: “Phonics: necessary but not sufficient”. It is important to note that phonics is a form of instruction. Clearly systematic phonics is not necessary as most children learn to read without systematic phonics, and indeed, as I show in this blogpost, there is little evidence that systematic phonics is better than non-systematic phonics instruction. That said, I agree that learning GPCs is necessary. But this is not the same thing as to say phonics is necessary. I think it makes more sense to teach GPCs in the context of morphology and etymology for the reasons I detail above.
      All the best,


      1. I would love to see what this looks like, Jeff. I’ve watched several videos that Pete has recommended featuring pre-school and Kindergarten children, but I don’t see them learning GPC’s. Most recently, I watched K children dissecting the word “getting” by looking at “ing” and the double “t”. I have first graders–even a few second graders–who still can’t blend /g/ /e/ /t/ to read the word “get”, let alone “getting”. I just can’t wrap my head around what your recommendations look like for children at the very beginning stages of learning to read, and I have all the same questions and concerns that Anne has already expressed. After kids crack the code, I think what you’re advocating is wonderful and will make a real difference in the classroom. For me, morphology supplements rather than supplants phonics.


        1. Harriett, did you see my comment pointing you to my latest newsletter with tons of examples of instruction in pre-school to Gr. 2. If you follow the links and images you will see tons of illustrations of grapheme-phoneme charts and investigations. Sometimes the investigations are focused on grapheme-phoneme correspondences. Sometimes a brief comment is made to plant a seed of a grapheme-phoneme correspondence. For example, in the video of the preschool class constructing a web on the word “rain” when a kid adds the word “rains” and uses it in a sentence. As the student discusses the word, she independently emphasizes that /z/ pronunciation at the end of “rains”. When the teacher adds that child’s word to the web, she spells out the base, and then she added the suffix . The teacher has the whole class attend to how it feels to pronounce that final phoneme in the word. The investigate whether it is a /z/ phoneme or a /s/ phoneme. They agree it is a /z/. What better way to encounter the fact that the grapheme can write both /s/ and /z/ than in a word the children have brought to the class activity to investigate? This is just a tiny example of orthographic phonology instruction. The Newsletter I pointed to has tons of examples.

          1. Yes, Pete, I did watch all about “rain”. My confusion is this: I’m still getting first grade kids to recognize that “ai” is usually pronounced /ae/, and therefore they need to blend /r/ /ai/ /n/ when they see those letters. That’s the video I want you to send me: How does your method teach the initial blending of phonemes? “Rain” has three phonemes and four letters–that’s where phonics comes in–teaching the sound/spelling variations. Look forward to watching this video!


          2. I’ve just rewatched the video recomended in your 2017 article that features Scott’s grade 1 class. His sound wall looks like my sound wall, and exploring spelling variations with highlighters is a big part of what I do (though I probably spend more time initially on decodable text featuring those variations–“Paul Saw Artctic Foxes” has the variations of /o/–which nowadays can be very interesting and not a forced march through contrived language).

            This is what I would consider teaching “explicit” and “systematic” phonics. So we agree on that part though we may have different labels for it. I still would love to see what SWI instruction looks like in teaching non-readers how to blend graphemes into actual words. To use Ehri’s terminology, the students in Scott’s class are moving from the “full-alphabetic” phase to the “consolidated-alphabetic” phase. What I would like to see is how you use SWI to get students to that point.


        2. Harriett, I can’t see how to reply to your follow up response, so I’m pasting your response here and then responding. You write:

          Yes, Pete, I did watch all about “rain”. My confusion is this: I’m still getting first grade kids to recognize that “ai” is usually pronounced /ae/, and therefore they need to blend /r/ /ai/ /n/ when they see those letters. That’s the video I want you to send me: How does your method teach the initial blending of phonemes? “Rain” has three phonemes and four letters–that’s where phonics comes in–teaching the sound/spelling variations. Look forward to watching this video!

          First of all I see my angle brackets that I used in my previous response for writing orthographic information like a grapheme just make the text disappear! So I’ll use single quotation marks…

          Understand that you have a specific purpose that didn’t need to be Carolee’s focus in her lesson. The pronunciation of the ‘s’ grapheme in the suffix ‘-s’ was what the child highlighted, so that’s what Carolee emphasized. She could have decided to focus on the vowel, but I want to be clear that there is nothing deficient in a different choice.

          One thing you’ll see in the Newsletter I linked to is that a standard practice in SWI is spelling out graphemes and markers in the base. When Carolee first showed me the video I mentioned, she told me she was embarrassed that she noticed that she hand not done that in the act of making the web. It’s part of the learning process! But my point is that if I were to spell out the word “rain” and my purpose was to identify the grapheme-phoneme correspondences in this word, I would start by spelling-out the graphemes as units.

          “r – ai – n”

          I might ask kids if they can spell the word like me. If they spell it out like this, I know they are not catching my cue.

          “r – a – i – n”
          That is spelling one-letter-at-a-time, not the graphemes in the base. If I try again and give one tap for each grapheme, kids can try until they see that I’m saying ‘ai’ as a unit, but NOT the ‘r’ grapheme or the ‘n’ grapheme.

          Now I can ask WHY do they think I announced the ‘ai’ together in “rain”.

          At that point we have a motivation to attend to the phonemes in the base. I might ask, “What do you feel at the beginning of saying ‘rain’?” to get them to pronounce /r/. “How do we write the ‘r’ in ‘rain’?” And now I can point to the grapheme ‘r’ that I said all by itself. Now I might ask, what do you feel at the end of ‘rain’?” to get the /n/ phoneme. Now I can say something like, “And we used the ‘n’ grapheme to write that phoneme.” And now we can identify the vowel phoneme. I might just pronounce it, or I might have kids “say “rain” without the /r/ and then that without the /n/ to get to the vowel phoneme. And now I point to the board, and the only letters left are . And now we can discuss that the reason I spelled the word rain out as “r – ai – n” with the ‘ai’ together is that those two letters are a digraph that spells the phoneme.

          See more on this “spelling-out-loud” and “writing-out-loud process that is central to SWI at this link:

          I do that kind of thing all the time. But not usually because I’ve decided randomly that today I’m teaching the ‘ai’ digraph. I do it because I’m studying a word family that has a base that happens to have this digraph. Think how many times kids end up spelling out ‘ai’ in a web of words with a base with that grapheme!

          This is just one example. I do not have films of many of the things I do. I capture what I can when I can. There is another film I’m trying to get from a school right now — but that means permission from parents and a whole process.

          Here is another video from a public school Grade 1 class in which the teacher investigates graphemes in the base in the context of word sums. You will see how motivated they are in part because they are investigating these grapheme-phoneme correspondences themselves (with guidance) in the context of words in a story they’ve been reading. It is not random words chosen to target abstract grapheme-phoneme correspondences. It is looking at those abstractions in the context of meaningful text.

          Please do read through that Newsletter and follow the links to other examples. I hope you pay close attention to the section on the difference between “phonics” instruction and instruction of “orthographic phonology”.

        3. Hey Harriett,

          I’m glad you liked this video from Scott’s class. (

          You write “This is what I would consider teaching “explicit” and “systematic” phonics”

          I understand why you think that this is the case and it is just that we are just using different terms. But there are crucial differences.

          What you see in phonics instruction and instruction of orthographic phonology in structured word inquiry in Scott’s class is that both address explicit instruction about grapheme-phoneme correspondences. This is exactly the point I make in the section of the Newsletter I pointed you to. But I also point out numerous crucial differences, many of which are evidenced in Scott’s class.

          Every grapheme addressed in Scott’s class is identified with in the morphological structure of a word sum. Instruction of orthographic phonology understands that graphemes never cross morphemic boundaries. If a child came across a word like “does” and thought s/he saw an “oe” digraph, Scott would help that child see the structure “do + es” and show that the ‘o’ grapheme is writing the /ʌ/ phoneme, but in the base ‘do’ and other relatives like ‘doing’ that same ‘o’ grapheme writes /uː/. Under no circumstances would Scott ever tell a child that a word that doesn’t follow the “phonics rules” is irregular. This is true even if Scott cannot explain that spelling at that time. Being unable to explain a spelling at a given moment is not evidence of an irregular word to memorize, it is evidence of a spelling that is rich for study, but we don’t need to understand it right now.

          I have never seen “explicit systematic phonics” instruction that studied graphemes within the context of morphological word sums, have you? You may not see it in this particular video, but many of the graphemes encountered in this lesson and likely some of the words will be used as the basis for studying a morphological family later. On Scott’s wall you will see matrices and word webs that are used to show that interrelated structure of words — AND show how graphemes in a base must be able to represent all of the pronunciations of that base. Again, that is something standard in teaching about orthographic phonology but absent in phonics instruction.

          You mention levelled books. Those are not in Scotts class. There is no sequence to the graphemes being taught. The text they are studying is one they’ve already read and enjoyed and now they are going back to a familiar story and language to inspect the orthography. That text has digraphs and trigraphs and there is no plan for which of those are taught. The kids discover them and test them out. They don’t start with single-letter graphemes and move on later to digraphs and finally trigraphs as I often see in phonics programs. They start with words they are interested in and find the graphemes in those words. The wall of graphemes for various phonemes you see are not from a book that said teach these grapheme-phoneme correspondences. They are simply the ones these Gr. 1 students have discovered. Note the bit at the very end where the kids are so excited that they found new graphemes for the wall. They are invited to be constructors of the bank of knowledge in the class.

          Did you see the process of “spelling-out-loud”? At the 4 minute mark you can see me and the student spelling-out the words ‘flew’ and ‘out’ in graphemes “f-l-ew” and “ou-t” and then discuss the phonemes that those digraphs represent. We then move on to the ‘igh’ trigraph of “high” for /aɪ/ with no problem.

          This is a process that we do all the time — building the mental representations for those graphemes. (See in the newsletter about the connection to cognitive load theory.)

          I do not know of this practice of spelling-out-loud and writing-out-loud in phonics. And in SWI it is in conjunction with spelling out affixes and the base from the start.

          You are wanting to see this instruction with children at earlier stages of learning to read. The thing is there are many students in Scott’s class who are in that stage. They are still engaging with this process — just with more support. In the “rain” video you saw in pre-school the same practices were going one just with more support.

          Here is a video (I think you’ve seen before) of me reading with my son in the first week or two of him reading independently at all.

          You will see that I’m doing exactly the same sort of practice. But my support in this early part of Skyler’s independent reading is the same thing I did when he couldn’t read at all and I would read stories with him at bed time. I would read along and every once in a while, I would pause on a word and we would check out how it was spelled out. So before this video, I know I would have encountered a word like “laugh” in a story. I would have paused and said something like, “Oh, that’s interesting that’s the word “laugh”. I spell that “l-a-ugh” and got him to do the same. This would have been teaching him the letter names at the same time. I might get him to feel the end of this word to get him to announce the /f/ phoneme. and then I could point out that the ‘ugh’ is writing that /f/. And then I’d just move on. I one-minute intervention here and there as we enjoyed reading stories together.

          With that background, this process when he starts to read independently is so powerful. You see it when he gets stuck on the word “caught”. In that word, he does not understand what the ‘ugh’ is doing. I bet few reading specialists do. It is something called an etymological marker. But it doesn’t matter right now that he doesn’t know what the ‘ugh’ actually does in this word. You can see from the video that his first inability to make sense of the letter sequence is clarified when I cover everything but the ‘ugh’ so that he can see that structure. He has a slot in his mind for a ‘ugh’ and that feels familiar. The only new thing is what it does, and we don’t need to address that now.

          Notice how he turns away at his first spelling when he says ‘th’ even though there is no ‘th’ sequence in ‘caught’. Why does he do that? Because he has a well-established mental representation for the super common digraph ‘th’ that we have spelled-out many times. He wants it to be something familiar, but it isn’t so he looks away to announce what he wants it to be! Once I cover up the letters in the word so that he can only see each orthographic structure one at a time, he says “Ohhhh! ugh” Because I spelled-out with him before he could read he was familiar that this is a common orthographic structure. He can’t read the word “caught” after we do this yet, but that is not my goal. My goal is to make sure that his experiences with English orthography always reinforce the fact that it is an ordered system that makes perfect sense and can be investigated.

          That is simply not a lesson that can possibly be learned from explicit systematic phonics that comes with lists of irregular sight words to memorize.

          I have not heard any thoughts from you about the Newsletter. I really hope you take the time to read that carefully Harriett. You’ve long asked me for examples of SWI instruction with young children. I spent a lot of time trying to put together a piece on that topic with multiple examples and the research context.

          Now I’m particularly curious to know after you study that document if I manage to convey that phonics and instruction of orthographic phonology are definitely not really just two different names for the same thing. But I sincerely hope that your observation makes you feel more comfortable about SWI. It is absolutely true that both SWI and phonics treat explicit instruction of grapheme-phoneme correspondences as a non-optional key part of early literacy instruction. A key difference is that the orthographic phonology taught in SWI teaches how those grapheme-phoneme correspondences are constrained by morphological and etymological factors. Both teach the “what” — the available grapheme-phoneme correspondences, only instruction that addresses the context in which those correspondences happen can teach the “why”. Orthographic phonology in SWI can explain the spelling of the super common words like ‘does’ ‘rough’ ‘one’ ‘because” and all the other words struggling children are taught they need to memorize because they don’t make sense.

          And I don’t know anything that can motivate learners to keep studying and investigating than to teach in such a way that they gain an understanding. We dyslexics demonstrate every day that we are terrible at memorizing spellings we don’t understand.

          1. Thanks for the follow-up, Pete, which will take me a while to pursue. In the meantime, here are some thoughts.

            1. I agree with Anne Castle that the complexity of what you describe is not appropriate for teaching beginning readers to decode text. Yes, I have done all those wonderful language explorations you describe (using, for example, a class set of Cloudy With a Chance of Meatballs with my kindergarteners), but I do these explorations in addition to phonics, not in lieu of. I recommend Stanislaus Dehaene’s 30 minute lecture on the brain to emphasize the monumental changes that take place as kids learn to read.

            2. It is precisely because most of the students I work with do not have the literacy experiences at home that you describe that we need to be strategic in the choices we make in the classroom, and here again I believe what you describe is too complex–at the beginning–for these students who are stuck in the achievement gap. I highly recommend Anne Fernald’s TED talk about her work with infants 18-24 months.

            3. Anne Castle said it best: These are empirical questions, so we’ll have to put them to the test.

            Thanks again!


          2. Hey Harriett,
            Thanks for this discussion. I hope it is useful for you and others who may be following.

            Yes, the other resources I’ve sent should take quite a bit of time to process. There is no hurry — but I do hope you study them carefully.

            I’m going to offer responses to your three points and then let’s wait until after you have time to process before we pick up the conversation again. It could be here or in emails, or via video conference. (I hope anyone who is interested in this discussion will consider the SWI conference going on in Chicago March 1 & 2 to study this work like they do other work at all the other reading conferences. Here is the link with info:

            So, to your final three points…

            1. You write: ” I agree with Anne Castle that the complexity of what you describe is not appropriate for teaching beginning readers to decode text.”

            I note how you say this as a conclusion. Anne presents it as a hypothesis. You may think this, but I hope you will see lots of counter evidence in the Newsletter I point to.

            You write “I have done all those wonderful language explorations you describe (using, for example, a class set of Cloudy With a Chance of Meatballs with my kindergarteners), but I do these explorations in addition to phonics, not in lieu of. ”

            I’m glad you draw from books like this. But are you including spelling-out loud graphemes and markers in the base and announcing affixes in word sums? Are you building matrices and word sums and investigating graphemes that link words with common meanings and spellings but different pronunciations like studying the link between “action” and its word sum (act + ion → action) to investigate the grapheme-phoneme correspondence of the ‘t’?

            This is not practice that can be learned quickly. It is learned over time working with these tools.

            You write: “I do these explorations in addition to phonics, not in lieu of.”

            This statement is why I’m so keen on you reading the section addressing the difference of “phonics” and teaching “orthographic phonology”. When you saw the video of Skot’s class, you thought it was an illustration of systematic, explicit phonics. I explained that the overlap was the fact that both phonics and Skot’s class did use very explicit instruction of grapheme-phoneme correspondences from the beginning (a necessary aspect of SWI) but all the other aspect of orthographic phonology that he was teaching that is no in phonics. When you say “not in lieu of phonics” it suggests that SWI does some instruction in the place of grapheme-phoneme correspondences. That is clearly not the case. One could say that SWI teaches about “orthographic phonology” in lieu of “phonics instruction”. That means that it explicitly teaches the available grapheme-phoneme correspondences — and how we can understand why one grapheme is chosen over another based on morphological and etymological relatives.

            I have seen Dehaene’s talk. I can say that his work has no bearing on the basic principle of SWI that literacy instruction should reflect a valid understanding of the orthography we teach children to read and write. He places a huge emphasis on the role of phonology in learning to read. I would argue that he should love SWI because it places a greater emphasis on teaching children how orthographic phonology operates than does phonics. Phonics teaches about letter-sound correspondences in isolation of other features of the language system. Why would the human brain make better sense of literacy in a context in which children are taught the “t sound” when that sound is not in the word “action” and is in the word “jumped”? If we teach the “z sound” when the most common way to write that phoneme is the “s”, I suspect that misinstruction about how orthographic phonology works in English will not make it easier for the human mind to learn how English orthography represents it’s phonology. I’m not saying every phonics program uses terms like “z sound” and “t sound” but boy is that standard practice.

            2. You write “I believe what you describe is too complex–at the beginning–for these students who are stuck in the achievement gap.” I know you believe this. But you are coming to that conclusion without seeing it in action. I hope that Newsletter helps you get a wider view. I will watch the TED talk you mention.

            3. You write Anne Castles says it best, ” These are empirical questions, so we’ll have to put them to the test.”

            I agree that these are empirical questions and we should put them to the test. But what seems to be lost is that we have two hypotheses that need to be put to the test:

            Hypothesis 1) Phonology-first hypothesis: teach about associations between spelling and pronunciation first before including information about morphology and phonology.

            Hypothesis 2) Teach the orthography system from the start: Instruction should address the interrelation of morphology, etymology and phonology from the start.

            The sense I get is that your view is that Hypothesis 1 is already established by the research, and that we require new research evidence before we encourage people to attempt Hypothesis 2.

            For Hypothesis 1 to be a research-based conclusion which we treat as the default, we should be able to point to evidence comparing these to hypotheses and show that Hypothesis 1 is more effective. That means, for example, we should be able to point to evidence of instructional research in which including morphology from the beginning resulted in lower gains in learning compared to teaching isolated phonics from the start.

            I have yet to have anyone point me to any research showing that including morphology from the start resulted in any loss compared to including morphology from the start.

            Until that evidence is presented, the statement that these are empirical questions that need to be put to the test is EQUALLY true to both hypothesis. Otherwise, on what research basis can we recommend teachers AVOID morphology from the start.

            Ironically, all the evidence I can see points in the opposite direction.

            Results from all the meta-analyses of morphological instruction point to particular benefits of including morphological instruction for less able and younger students. This is the opposite of what the phonology-first hypothesis would predict.

            Goodwin & Ann (2013) pointed out that in both their meta-analyses (2010,2013) the greatest outcome effect was phonological awareness outcomes. Morphological awareness outcomes had the second highest effect sizes. A key argument for the phonology-first hypothesis is that time spent on morphology would reduce the learning of phonological aspects in learning to read and write. That hypothesis would predict a reduction in effect sizes when time was spent on morphology at the beginning. Again the finding is in the opposite direction of that hypothesis.

            Goodwin and Ann (2013) argue that this finding for phonological outcomes may signal the synergistic relationship between morphology and phonology in our morphophonemic language. These findings like the previous ones all point in favour of Hypothesis 2 and against Hypothesis 1.

            And then there is the one direct instructional study by Devonshire, Morris & Fluck (2013). (See here:

            They compared a condition with SWI type instruction teaching the interrelation of morphology, etymology and phonology (with word sums and matrices) to 5-7 year olds and compared it to a phonics condition. They found significant effects in standardized measures of reading and spelling compared to the phonics condition.

            Again, I have yet to find any evidence supporting the hypothesis to AVOID instruction of morphology (or etymology). That would be the necessary evidence for the “phonology-first” hypothesis to be a research-based instructional recommendation.

            But more than that, we do have empirical evidence that can speak to these to hypotheses and all of it points in the direction of Hypothesis 2 — teach how the writing system works from the beginning.

            If we are going to restrict our instructional recommendations to the best evidence of empirical research, we should include morphology from the beginning of instruction and for less able students. If you or anyone can point to empirical evidence that counters that conclusion, I would be delighted to look at it.

            But any argument that there is more research evidence for the phonology-first hypothesis than the hypothesis of teaching how orthography works from the beginning should be duty-bound to account for the counter evidence I have provided above. I’m totally happy to engage with that research discussion.

            Actually, I would hypothesize that the evidence from the meta-analyses are way under-representative of what we would see with SWI instruction. Why? Because central to SWI is that we explicitly teach the interrelation of these features of language. Of the 22 studies in the meta-analysis I was involved in only 5 even addressed that interrelation at all.

            Even with that lack of explicit instruction about the interrelation, Goodwin & Ann find the greatest effect of including morphology was phonological awareness outcomes. Shouldn’t we expect explicit instruction should increase those kinds of gains?

            I have not yet heard any challenge to the research case I’ve outlined above and that Jeff and I have been publishing (also that I’ve published with John Kirby and John Kirby and Helen Deacon) over the years.

            We are totally open to be challenged on that research case, but that part always seems to be swept away with the claim “it might be useful to include morphology early on, but we need more empirical evidence.”

            Actually, as far as I can see there is no empirical evidence supporting the phonology-fist hypothesis. and I’m still waiting to be presented with counter evidence to that claim. The empirical evidence is fairly early for Hypothesis 2 — we totally agree. We need more. But we are not going to get more if the message from researchers is to avoid morphology until later.

            One reason I think people have a hard time accepting the research case we are making is because they mistake the idea of teaching the interrelation of morphology, etymology and phonology from the start means not teaching phonological aspects of orthography from the start. But that is simply not true. As you have seen in the videos and examples I’ve shown, explicit instruction of grapheme-phoneme correspondences is a non-optional aspect of SWI instruction from the early grades. It is just that it is taught within the wider context of morphological and etymological relatives with the help of matrices and word sums and other practices like “spelling-out-loud” word structure.

            And that brings me to the reason I keep pointing to that Newsletter. My guess is that a reason the phonology-first hypothesis persists despite lack of empirical evidence — and in fact, a good deal of counter evidence — is that few educators or researchers have any image of what instruction could possibly look like from the early years. You have repeatedly asked for examples. I don’t know if you’ll find exactly the video you are looking for in that Newsletter, but you will see tons of examples of what SWI can look like from the start.

            I’m hoping that any researchers reading this blog will be motivated to not only check out that Newsletter, but contact me about going to actually visit the schools around the world doing this work. Why not take advantage of these labs having a go at the kind of instruction the research says we should be figuring out. The meta-analyses of morphological instruction all agree that we should teach morphology from the start — but that what w lack is research that can tell us more effective ways of teaching it. There are examples around the world that you can go and see to get ideas for intervention studies to design. Why not take advantage of teachers who have been working for years at developing practices that the research says we need to explore? Email me at if you are looking for such a connection.

            Why not attend the SWI Conference in Chicago March 1 & 2 to see presentations on this research and practice including teachers presenting what they are doing right now in their classrooms across the grades (See information here:

            Hariett, I have long been impressed at your willingness to enter these discussions with me. I know I make strong cases that you may or may not agree with. I hope you get a chance to read this before or after you study the newsletter, but please take your time to do that before we pick up the conversation again.

            In particular, I hope that you or anyone responding to this post with reference to the research specifically address my claims about the evidence for Hypothesis 1 and Hypothesis 2.

            I’ll end by repeating the key point in this post that somehow never gets responded to:

            Any argument that there is more research evidence for the phonology-first hypothesis than the hypothesis of teaching how orthography works from the beginning should be duty-bound to account for the counter evidence I have provided above.

            Hope that helps!



  4. First of all, I want to emphasize again how much Jeff and I appreciate Anne Castles’ openness to engage in this discussion in a comment string building on a research debate that challenges some of her published work. Anne has taken time to read various comments and then to respond thoughtfully and openly. There is no requirement to agree to have a productive discussion. As a result of Anne’s willingness to engage, this comment string offers readers so much more to consider — including views that counter the case Jeff and I are putting forward. This is exactly the kind of scholarly debate Jeff and I are hoping to spark with our recent articles and this blog.

    Anne recently responded to a comment of mine that I wanted to add one more response. I feared if I just responded directly, it would be a comment that might get lost in the thread. That’s one reason I’m starting a new comment string here to highlight what her comment made me think about. The other reason to start a new string is that the point that Anne makes is a reflection of a larger point I want to make that is not really limited to her personal take. Here is the section of her comment that grabbed my attention:

    “As I’ve mentioned before, though, the question of when and how to introduce this knowledge [interrelationship of morphology, etymology and phonology] to children learning to read is a separate one, and an empirical one. As you would know, the range of abilities of children learning to read is enormous: I see many children who are barely at the point of being able to recognise all their letters by the end of the first year of schooling. As I have said to Jeff, this leads me to hypothesise that keeping things really simple, and at the individual letter-sound unit of instruction, is the best way to get the ball rolling to get children started in reading. Then much more can follow. But this is just my hypothesis – I would welcome studies that tested this against your method.”

    There is much in this statement that I agree with. As Jeff and I have published again and again, there is very little research on the effect of teaching the interrelationship of morphology, etymology and phonology, that characterizes structured word inquiry instruction – and even less that looks at this instruction in young classrooms. Devonshire, Morris and Fluck (2013) do provide one compelling study with 5-7 year olds that showed significant gains on standardized measures of reading and spelling in an SWI based condition compared to a phonics condition. So while direct evidence is limited, what we have points in the direction of this kind of early instruction.

    Jeff and I are in complete agreement with Anne’s point that the question of when to do this kind of instruction is an empirical one. Our recent publications, and my recent chapters with John Kirby (Kirby & Bowers, 2017,2018) and the earlier meta-analysis (Bowers, Kirby & Deacon, 2010) made this point explicitly about when morphological instruction (not the same as structured word inquiry) should be introduced. We cited evidence from our own meta-analyses and those by Goodwin and Ann (2010, 2013) that found the greatest benefits on including morphology in literacy instruction was for younger and less-able students. Kirby and I have repeatedly explicitly argued that the evidence is that we should teach morphology from the beginning of schooling, but we do not have instructional evidence telling us HOW best to include morphology in early literacy, or in any grade for that mater. SWI is not the same as morphological instruction, but it is a necessary and central aspect of SWI, so we are putting this out there as something that we hope other researchers will explore given the current state of data on literacy instruction. And as Anne states, she welcomes such studies.


    I want to take another look at the thinking Anne presents…

    “I see many children who are barely at the point of being able to recognise all their letters by the end of the first year of schooling…. this leads me to hypothesise that keeping things really simple, and at the individual letter-sound unit of instruction, is the best way to get the ball rolling to get children started in reading.”

    Again this is presented as a hypothesis, which is totally fair. I would also say that the frame Anne presents reflects the thinking of the vast majority of researchers and educators.

    Instead of complicating matters with regard to direct instructional research on SWI, let’s move over to instructional evidence regarding just morphological instruction (a necessary, but not sufficient aspect of SWI). Over at another blog post, Jeff cites some statements on that question including this one by Taylor and Rastle (2018):

    “We believe that a focus on these morphological regularities is likely to be more appropriate in the later years of primary schooling.”

    Again this statement is made as a hypothesis, not as a conclusion. But here there is a larger problem. There is evidence about the effect of including morphological instruction in early and later grades, and for less able students. As we have shown, all the meta-analyses that could provide evidence on this question point in the opposite direction of the hypothesis as stated above. Counter to decades of assumptions, the students who gained the MOST from the inclusion of morphological instruction were the younger and less able students. In our meta-analysis (Bowers, Kirby & Deacon, 2010), pre-school – Grade 2 students showed better gains from the inclusion of morphology in their instruction than was found for students in Gr 3 and above.

    Also consider this finding from the research on morphological instruction. In their meta-analyses of morphological instruction, Goodwin and Ann (2013) noted that in that study and their earlier meta-analysis (2010), the outcome with the greatest effect was not morphological awareness, but phonological awareness. Interpreting those results, they wrote, ” Goodwin and Ahn (2013) reasoned, “Similar to Bowers et al. (2010), results suggest that early morphological instruction may be particularly helpful perhaps because of the synergistic relationship between phonology and morphology and the larger repertoire of root [base] and affix meanings available for use. If a reciprocal relationship exists between morphological knowledge and literacy…it makes sense to jump start this knowledge from an early age” (Goodwin & Ahn, 2013, p. 23).

    A point I’m trying to be very clear about is this — it is one thing to say that there is not enough direct instructional evidence to conclude SWI instruction (interrelationship of morphology, etymology and phonology) should begin from the start. (What we have is positive, and there is no evidence to avoid it).

    But it is a very different situation to present a hypothesis that morphology should be introduced after a certain amount of grapheme-phoneme instruction. To present such a hypothesis, researchers now need to account for the existing evidence that challenges that hypothesis. I have yet to see this hypothesis made post those three meta-analyses about morphological instruction while also explaining why that evidence does not negate the hypothesis to avoid morphological instruction at the beginning. I’m open to considering the argument, I just don’t see the argument being made.

    It may seem like it is obvious that children need to be taught the letter names and common letter-sound correspondences. That is certainly the experience most teachers and researchers have had for the bulk of their careers. It’s a totally understandable assumption. But that is why we do research with instructional studies, and then do meta-analyses of such instruction to see if we can find evidence that supports or contradicts our hypotheses. What I find troubling is the fact that many researchers are presenting a hypothesis that may seem logical, but without addressing the empirical evidence that contradicts it. As I say their could be good arguments for why these meta-analyses do not challenge their hypothesis. But given the evidence we have, the onus should now be on researchers to account for that evidence.

    But there is another issue behind all of this.

    It makes sense to me that researchers and teachers could wonder to themselves when they see that evidence from meta-analyses, “but what on earth would instruction about morphology look like with kids who can’t yet read and are still learning the letter names?” There is very little institutional knowledge from classrooms for what this could look like. It seems to me that what is happening is that researchers are not addressing the research that counters long-held assumptions in part because they can’t picture what such instruction could look like. But that lack of experience is not what should be driving research hypotheses and conclusions — especially when the published evidence counters their hypotheses.

    It seems to me that a key next step for the research community is to pause and take note of the fact that the best current instructional research counters assumptions that we have been working with for decades. Many researchers have long suggested that if we teach morphology, it should not occur until later years, or at least until basic ‘letter-sound correspondences” are established. The fact that the results of multiple meta-analyses are not in line with that prediction should be a cause for a pause and a celebration! We should be delighted to find that a basic assumption about literacy instruction does not seem to hold up to the evidence.

    Why delighted?

    Because that means that a whole new rich line of inquiry has opened up. For decades we’ve been acting as though the need is to refine our instruction of “letter-sound correspondences”. And we’ve been doing that, but not getting the results we would hope for. Discovering that there might be another way to address this problem that has not yet been carefully addressed is great news. You can only refine current instruction so much. If it turns out that our assumption that morphology should come later has been hindering our ability to see another possibility, we should be delighted to move past that assumption. Consider the final paragraph from an article by Al Otaiba and Fuchs (2006) on Response to Intervention that identified the challenge of understanding why so many continue to struggle despite what has been considered “best-practice” (that makes no mention of morphology). They wrote:

    “Our nonresponders clearly required a secondary level of intervention, using a different method or combination of methods delivered with greater intensity than was available to our study participants. Given the heterogeneity of the nonresponders, we believe that this secondary level of instruction must not only be more intensive, but must be tailored to children’s individual strengths and weaknesses. Perhaps the most humbling implication of our study is that although we and others have learned much about what is necessary to reduce the number of nonresponders and to understand some of the characteristics that make it difficult for them to learn to read in mainstream class- rooms, we do not yet know nearly enough about which specific approaches are key to ensure that no child is left behind” (p. 428).

    Note that the authors are recognizing that we are not getting the results from the current instruction and they are looking for any way to find approaches that can build on children’s strengths and reduce the effects of any weaknesses. But nowhere in the article is there any suggestion that we look at the possibility of including instruction about the way morphology and phonology interrelate in English spelling. This is 2006, but it is not like we did not know about the morphophonemic nature of English. And the point is, we now do have a great deal of evidence and theory to motivate an intensive search for instruction that is not restricted to phonological factors from the beginning of schooling.

    When researchers hypothesize that children should be taught about grapheme-phoneme correspondences before morphology – but make no mention of the evidence countering that hypotheses, we simply reinforce the same path of inquiry that we see in this 2006 article.

    Another barrier…

    It seems to me that a major challenge is that one result of decades of the phonology first assumption has left teachers and researchers with little or know ability to imagine what including morphology or etymology instruction could possibly look like in the early years. It is very hard to get behind a hypothesis for instruction for which one has no experience. This is why I try to point to videos of such instruction all the time.

    At this link you can find a series of classroom videos. ( The first is a Gr. 1 class in which I address the spelling of homophones and function and lexical words. This is not specifically morphology (although that comes into play in subtle ways). But it is instruction that addresses an aspect of etymology (e.g. homophone principle). If you watch it, you will see that I am teaching explicit grapheme-phoneme correspondences at the same time that I’m teaching the names of letters and graphemes, but doing so in the context of meaningful connections. Children do not need to know all their letter names and grapheme-phoneme correspondences to engage with this lesson.

    At this link ( you will see a pre-school teacher engaging her students in an investigation of relatives of the morphological family built on the base “rain”. Again, many children in this class are still learning their letter names, some can read. But they all get to engage in a meaningful conversation about words they know, and the build their letter knowledge and grapheme-phoneme correspondence knowledge in the context of spelling-out these words and studying them together. Note how the teacher highlights the phonology of the “s” when they get to the “-s” suffix in “rains”. It’s a small seed she plants, but it’s a great introduction to how their orthography system works in given its morphophonemic nature.

    At this link ( is a video of me introducing kindergarten students to morphological word sums. Again you’ll see that there is no lack of explicit grapheme-phoneme correspondence instruction. Instead, that instruction of abstract associations is made more interesting (meaningful and concrete) because it grows out of the study of words from their book. Notice when the child announces “s means more” to arcuate whey the author added an “-s” suffix to the word “eggs” even though that “s” was pronounced /z/. Children do not need to wait until they’ve learned all their grapheme-phoneme correspondences before this particularly important one is encountered in a meaningful context provided by word sums for “chicks” and “eggs”.

    Finally, my hope is that researchers who read Jeff’s blog and this (too long) comment are realizing that there are actually tons of examples of early morphological instruction and early SWI instruction going on around the world all the time. Explore the very accessible examples at Rebecca Loveless’ website ( where she shows many examples from the work going on at the Nueva School near San Francisco and other schools in the area. Go to Lyn Anderson’s blog ( and see countless examples of work here — which has been inspiring for the work at Nueva and teachers around the world.

    But better yet. I encourage any researcher who feels skeptical that instruction could address morphology and or etymology at the same time as grapheme-phoneme correspondences (or those that are excited about it) to contact me directly ( so that I can put you in touch with teachers / schools where this kind of instruction is going on. I worked with the Nueva School for years as a visiting consultant before I went there as an SWI coach for the 2015-2016 school year. The work going on in their lower school is so far beyond what most think is possible. Given the questions about work with younger students, the work in pre-school and kindergarten there is particularly significant in terms of proof of concept that researchers should go study. But there are schools all over — Chicago, Edmonton, Melbourne, Bangkok — and beyond where innovative work is going on.

    My challenge to researchers is to contact me so that I can connect you with a school you could go visit to see such instruction in action. Wouldn’t it be great to talk to teachers who are doing this work and students who are learning in this way to get an image of what it can be like? The whole time I was at Nueva, Marcia Henry was the only reading researcher who came to visit what was going on there. Our old hypotheses about early literacy instruction were not supported by the evidence of meta-analyses. That should put a fire under researchers to dive into a rich area that is not yet explored enough. The teachers I work with are totally excited to share what they have been learning. Email me at if you would like me to help you find some teachers and students who can give you a picture of what this kind of instruction can look like. That kind of research of innovative practice is needed to guide researchers to design interventions that can take a look at a question that has now long been highlighted as a need in the research. In our meta-analysis of morphological instruction (Bowers, Kirby, & Deacon, 2010) we pointed out that we had evidence that morphology should be included in instruction from the beginning of formal instruction and with students who are struggling, but that we still did not have evidence from instructional studies about how best to teach morphology. The next step for researchers, I think, is to go visit classrooms where this kind of instruction is going on so that they can have a picture of what might be the key ingredients of this instruction that can be tested.

  5. Hi Anne, I appreciate your comment below “My gut feeling is that this is too complex for very early readers, the single letter-sound level should be the unit of focus initially – but I’m the first to admit that we haven’t done the study to establish that. But the problem is that you and pretty much everyone else in the field claim in their publications that the evidence for systematic phonics is compelling. For instance, here is a passage from your recent Castels et al. (2017) paper describing the various meta-analyses:

    “The evidence for the effectiveness of phonics instruction is extensive and has been surveyed comprehensively elsewhere. The most influential analysis arose as a result of the National Reading Panel convened by the U.S. Congress in the 1990s. Part of the work of the panel was to conduct a quantitative meta-analysis evaluating the impact of systematic phonics instruction compared with nonsystematic or no-phonics 38 experiments involving 66 treatment-control comparisons, this meta-analysis yielded a moderate impact of phonics instruction (i.e., effect size) of 0.41,1 which was much larger when phonics instruction began early (d = 0.55) than when it began after the first grade (d = 0.27). Phonics instruction improved decoding, spelling, and text comprehension. This result is broadly consistent with a subsequent meta-analysis of 14 randomized controlled trials investigating the impact of phonics instruction on reading accuracy (Torgerson, Brooks, & Hall, 2006), although the overall effect size was reduced (d = 0.27). More recently, two meta-analyses have concluded that phonics instruction is an effective intervention for poor readers (Galuschka, Ise, Krick, & Schulte-Korne, 2014; McArthur et al., 2012). The only meta-analysis that has examined the longer-term outcomes of phonics instruction produced a variable pattern of results, but there was clear evidence of benefits on spelling (Suggate, 2016).”

    This provides the reader with the impression that the evidence is strongly in support of systematic phonics, contrary to fact.

    Briefly, let’s consider the meta-analyses you cited in this paper. As opposed to the NRP (2000) overall effect of .41, Camilli et al. (2003) found that the overall effect of systematic phonics was .24 when systematic phonics is compared to non-systematic phonics (the correct comparison). Let’s ignore the Camilli et a. (2006) study that re-estimates the effect size to be a non-significant .14 and just focus on the Camilli et al. (2003) paper that you seem to accept. The .24 is an overall effect, and presumably will be much smaller if you considered the most important variables that were assessed in the NRP, including reading text orally and comprehending text (these effects were .25 and .27 in the original report). And the effect would be cut roughly in half if we considered the long-term effects (given that the effect went from .41 to .27 following 4-14 months delay in the original NRP). And the NRP included many non-randomized studies, and there was evidence of publication bias influencing the effects from Torgerson, Brooks, and Hall (2006). So this .24 effect is not a good estimate of the long-term benefits of systematic phonics instruction on various important measures of reading. It is also almost 20 (not 30 as I said above) years old now, and multiple more recent meta-analyses have been carried out with more up-to-date research.

    Next you cite the Torgerson, Brooks, & Hall (2006) noting that the overall effect size was reduced (d = 0.27). But again, this is likely an overestimate. The following is a passage from my review paper:

    “There are, however, reasons to question the significant word reading accuracy results. This result was largely due to one outlier study (Umbach et al., 1989) that obtained a massive effect on word reading accuracy (d = 2.69). In this study, the control group was taught by two regular teachers with help from two university supervised practicum students, whereas the experimental group was taught by four masters degree students who were participating in a practicum at a nearby university. Accordingly, there is a clear confound in the design of the study. As reported by Torgerson et al. themselves, when this study was excluded the effect was much reduced (d estimates between .20 and .21) and the effect was only marginal on one analysis (p = .03) and non-significant (p = .09) on another. And once again there was evidence that bias may have inflated the estimate of effect sizes in this study.”

    Note, even before removing this outlier study Torgerson et al. reported that there were no significant effects on comprehension or spelling.

    You mention the Galuschka, Ise, Krick, and Schulte-Korne (2014) meta-analysis, but as detailed in my post in Pamela Snow’s blog:, this paper provides no support for the claim that systematic phonics is better than alternatives. The various methods the authors tested reported similar effect sizes, and there was no interaction between methods, meaning there is no evidence that systematic phonics is better than alternative methods. The only reason why systematic phonics was significant and other methods were not is that there were many more systematic phonics conditions.

    You cite McArthur et al., (2012), but again the study provides no reason to support of systematic phonics. Most importantly, they compared systematic phonics to studies that had no phonics at all (that is, they excluded studies that used non-systematic phonics, as typical in whole language, so can’t make any claims about whole language). And even then, the authors failed to obtain significant effects word reading fluency (d = -0.51; expected direction), reading comprehension (d =0.14), spelling (d = 0.36), and nonword reading fluency (d = 0.38, the unexpected direction). There are additional problems that listed in my review paper.

    Finally you cite Suggate (2016) citing benefits on spelling. But in fact, systematic phonics did the worse compared to various methods. Here is from the abstract of Suggate:

    Overall, comprehension and phonemic awareness interventions showed good maintenance of effect that transferred to nontargeted skills, whereas phonics and fluency interventions, and those for preschool and kindergarten children, tended not to.

    The main point of my review paper is that there is no evidence for systematic phonics. This needs to be communicated to the research literature as the strong commitment to systematic phonics is discouraging people looking for alternative methods, and it is making it hard to publish alternative views. Part of the reason I’m posting on blogs is that I cannot get my review paper published, just as it took about 5 rejections across different journals before the Bowers and Bowers 2017, I am now on my second rejection on my review paper, with all comments factually incorrect. I’m tempted to post the reviews and my response… we will so how things go. The bias in this field is really problematic.

    But let me end with an agreement — I think GPCs should be taught explicitly. But for the reasons detailed by my brother Peter Bowers in his post, this is very different than systematic phonics. That said, there is not much empirical evidence for this approach either, and more research in this direction is important given that the “reading wars” has ended up a tie.

    1. Hi Jeff,

      These discussions have all occurred post the publication of our 2017 paper, so I think it is a bit unreasonable to suggest that paper should have reflected them! And as I said at the outset, I stand by the conclusion that overall the weight of the evidence supports systematic phonics (by which I mean the structured and explicit teaching of GPCs – I wonder if some of the disagreement here stems from subtle differences in what we mean by systematic phonics). I was not aware of the Camilla et al papers and, as I have also said, I agree that the point about the phonics vs systematic/no-phonics comparison in the NRP is an important one. Had I been aware of this work at the time of writing the 2017 paper, I would have included it.

      You will also note that we did not come down in favour of synthetic over analytic phonics in the 2017 paper, because we did not conclude there was sufficient weight of evidence to support that. So we explicitly did not conclude that there was only one way to explicitly teach GPCs and that this was by teaching each one separately, outside the context of words. And this is what I mean when I say that, although my gut feeling (based on lots of observations of kids) is that the unit of focus in very early reading instruction should be the single letter-sound unit, I am aware there is not the evidence to support that. There may indeed be more than one way to successfully teach GPCs explicitly. One of those ways might be SWI, but that also awaits evidence. So, I stand by the overall conclusions of our 2017 paper.

      I’m going to leave it there, because I think there’s only so far we can get in blog comments!


  6. Anne,

    I really appreciate your engagement with this conversation. I have to say that I am convinced that Jeff’s description of the analysis of the NRP report by Camilli et. al makes it clear that the studies addressed in that 2001 report does not provide evidence that allows us to conclude that there is any clear evidence that “systematic phonics” is more effective than “unsystematic phonics”. They simply did not have a design that could draw that conclusion. But I want to leave that aside for the moment to address this from your earlier message. You write:

    More generally, a response you made on Twitter made me realise the distinction you draw between phonics and GPCs. So, to get things back to basics, let’s just focus on GPCs. Do you agree with all of these points:

    1. It is important to teach GPCs right from the commencement of reading instruction.

    2. It is important to do so explicitly.

    3. It is important that children get to practice the GPCs they have been taught, close to the time that they’ve learned them.

    If you agree with all these points, I think we have a very firm platform of agreement about the evidence to build on.

    Just so everyone reading is clear, by GPC, I take it you mean “grapheme-phoneme correspondences”.

    I can unequivocally say that any instruction that all three of these points should be occurring in any instructional context that is described as “structured word inquiry”.

    Grapheme-phoneme correspondences are a foundational aspect of how English orthography works, so any instruction making sense of the writing system MUST include explicit instruction of this aspect of orthography.

    One point I am trying to really emphasize around this question is the distinction between teaching “orthographic phonology” (essential in SWI) and “phonics instruction”. Phonics instruction and instruction of orthographic phonology in SWI BOTH emphasize the importance of teaching the available grapheme-phoneme correspondences in English orthography. The crucial distinction is that phonics is an educational term about teaching grapheme-phoneme correspondences, without any reference to the morphemic and etymological constraints on grapheme choice for a given word. Without addressing those constraints, there is no way for a student to understand why these spellings do not work in English: *dogz (for ‘dogs’), *acshun (for ‘action’) or *becuz (for ‘because’). (Note: I’m not using angle brackets for spellings as I don’t know if that works in this blog.)

    the ‘z’ grapheme can write /z/, but if we teach /z/ as the “z sound”, why on earth would a student use the spelling ‘dogs’ where there is no “z sound”? Well, one thing both phonics and orthographic phonology instruction can tell children is that the ‘s’ grapheme can write /s/ and /z/. But that is not enough information to understand which one to use in this word. The only way to understand is to teach about how we can mark “more than one” with an ‘-s’ suffix. When we have a word and we add either a /s/ or /z/ phoneme to mean “more than one” there is no question left. We always use the ‘s’ grapheme, because that is the ‘-s’ suffix for more than one. There is no ‘-z’ suffix anywhere.

    This is such a simple little example but it brings such clarity to the fact that morphemes (bases or affixes) do not have pronunciations until they are in a word. This lesson doesn’t just help the learner spell ‘dogs”, it helps them understand their orthography system better. Phonics and orthographic phonology both teach the “what” of grapheme-phoneme correspondences explicitly. Only orthographic phonology can teach the “why”. Children currently struggling with spelling are desperate to have a “why”.

    So what about *’acshun’? Well once we know the meaning of the word ‘action’ we can draw a connection between this word and its base spelled ‘act’. We can study the structure of related words with word sums like this:

    act + or → actor
    act + ing → acting
    act + ive → active
    act + ed → acted
    re + act → react
    act + ion → action

    By collecting a set of words connected by a common base element (an orthographic morphological family) students get to study words that are related in both meaning and spelling-structure. This is not one-at-a time learning. And when we study morphological families explicitly like this, we encounter the fact that the base spelled ‘act’ has more than one pronunciation. In ‘actor’ the base is pronounced /ækt/ but in ‘action it is pronounced /ækʃ/. It is by encountering the change in pronunciation of this base that the pronunciation of the ‘t’ grapheme becomes interesting. We can now explain that while the default pronunciation of the ‘t’ grapheme is /t/, it can also write the /ʃ/ (and /tʃ/ phoneme). If phonics teaches common grapheme-phoneme correspondences explicitly, every child who gets intensive phonics should know this. My experience is that most teachers and students — even working with intensive phonics programs — are surprised to learn of this grapheme-phoneme correspondence that is so common.

    Because we are studying orthographic phonology in SWI, we can now explain that we need the ‘t’ in ‘action’ because (think ‘be + cause’) to understand grapheme-phoneme correspondences work in English we can’t study a word in isolation. It is not that we need to know the grapheme-phoneme correspondences in a specific word, we need to find the graphemes that can write all of the phonemes it needs to represent in the FAMILY. The ‘t’ can write the /t/ we need for ‘act’, ‘react’, ‘actor’ but it can also write the /ʃ/ we need for words that use this base like ‘action’ and ‘reaction’. If we keep going down this track we can learn that for the ‘t’ to write /ʃ/ there has to be an ‘i’ or ‘u’ immediately following it.

    It’s not easy to explain this in text in a blog post. But it is super easy with word sums and matrices and grapheme-phoneme correspondences charts that are ever-present in SWI classes.

    I don’t have a video of teaching ‘action’ but here is one introducing kindergarteners to the word sum, the orthographic phonology of the ‘s’ grapheme for /s/ and /z/ in the context of the ‘-s’ suffix at this link:

    Here is a lesson with a young class explaining the spelling of the word ‘does’ with a matrix and word sums.

    In terms of explicitness of grapheme-phoneme correspondence instruction in SWI, I would highlight this page on my website on an instructional practice I call “writing-out-loud” and “spelling-out-loud”.

    The basic point is that when we write-out word sums, you might only see the morphemes in script — but the convention work very hard to establish is to “spell-out-loud” the graphemes and orthographic markers in the base. Affixes are announced quickly as a group, and we pause at morphemic boundaries. This ties a motor-movement of announcing and writing common digraphs and trigraphs into every day work. The word “the” is not spelled ‘t-h-e’ and this hide access to the ‘th’ digraph. It is announced as ‘th-e’. The word “knight” would be spelled-out ‘kn-igh-t’. A digraph, trigraph and single-letter grapheme.

    There is much more to understand about this, but that link has more on this, and videos of examples of how this practice plays out.

    But to answer your question, all of your basic statements about instruction of grapheme-phoneme correspondences are addressed in SWI. Crucially, the orthographic phonology taught in SWI does much more by teaching how those grapheme-phoneme correspondences operate in morphological and etymological word families.

    One last point I can’t resist. Look at these two spellings:


    How can a reader know which has an ‘ea’ digraph if they don’t have instant access to the word? The necessary clue comes from morphology that can see with word sums:

    re + act → react
    reach → reach

    The word ‘reach’ is a base, so it does have an ‘ea’. The ‘e’ and ‘a’ of ‘react’ are NOT in the same morpheme, so by definition this letter sequence cannot be a digraph. I had a grade 8 student who was reading at a Gr. 4 level look at the word ‘react’ one day and asked if he knew what that word was. He sounded it out from left to write and got to the pronunciation of the word he knew in his oral vocabulary “reeked”. I asked, “You mean your socks reeked from playing soccer yesterday?” and he answered confidently, “yes.” He perfectly applied what he learned from phonics, got it wrong and thought he was right. How frustrating that must be! We had worked on the phonology of the ‘-ed’ suffix that included the fact that it could be pronounced /t/. I said, “but there is a problem. You mean that to be your socks reeked in the past. That would mean you need the ‘-ed’ suffix, but we don’t have one here. Have a go again, and see if you notice any bases or prefixes or suffixes. Without hesitation he said “a… react”.

    When his strategy was sounding out the letter-sound correspondences from left to write, there was not enough there for him to read the word. When my provocation was to look at structural cues, his mind brought him to notice the base ‘act’ (I’m sure that’s why the /æ/ of “act” came out first) and now he could no longer get confused by thinking of an ‘ea’ digraph here. This was our second day working together.

    Note the key point here. A definitional aspect of graphemes is that they NEVER cross morphemic boundaries. If we teach grapheme-phoneme correspondences with out this kind of knowledge of the interrelationship of morphology, etymology and phonology, we simply can not make sense of spelling. But I hope you see that in all of this there is no lack of emphasis on the role of phonology in instruction of how English orthography works. In fact I would argue that SWI provides far greater explicit instruction of how grapheme-phoneme correspondences work than we see in phonics.

    1. Hi Peter,

      Thanks for your comment, and these interesting examples. I absolutely understand the points you are making about the interrelationship between morphology, etymology and phonology – and I think this is something all reading researchers should be aware of.

      As I’ve mentioned before, though, the question of when and how to introduce this knowledge to children learning to read is a separate one, and an empirical one. As you would know, the range of abilities of children learning to read is enormous: I see many children who are barely at the point of being able to recognise all their letters by the end of the first year of schooling. As I have said to Jeff, this leads me to hypothesise that keeping things really simple, and at the individual letter-sound unit of instruction, is the best way to get the ball rolling to get children started in reading. Then much more can follow. But this is just my hypothesis – I would welcome studies that tested this against your method.

      Best wishes,


  7. This is a fascinating analysis and one that I am following with interest as a literacy teacher educator. Having worked with schools that tried “systematic, explicit phonics” programs such as Reading Master that “didn’t work”, this is so helpful. I think we are still looking for a “balance”. I feel the same with comprehension research. We don’t quite have all the nuance there either, even if we have the early “word recognition” story almost figured out. I don’t conduct research, but this conversation is immensely helpful.

  8. Hi Anne, thanks for engaging! But you will not be surprised that I disagree. You are right that almost all forms of instruction in the classroom include some degree of phonics (including whole language as noted by the NRP), but in the dataset that the NRP analysed, about 30% of the interventions did not. As shown by Camilli et al. (2006), when you compare systematic phonics to non-systematic phonics (and exclude these “no phonics” studies), and you factor out other obvious factors (class size, etc.) then the effect of systematic phonics is not reliable. The NRP has been cited over 20K times, whereas the Camilli et al. (2006) study that shows that there is no significant advantage of systematic phonics over non-systematic phonics as been cited 58 times. This is a problem.

    But put aside the NRP, there have been almost a dozen meta-analyses since, and none of them provide evidence that systematic phonics is better than non-systematic phonics (because people keep making the mistake of combining interventions that include non-systematic phonics with no phonics). And even when you ignore this flaw, there is little or no evidence for systematic phonics (as I detail in my review paper). For just one example, Pamala Snow on her blog (see: cites Galuschka et al. (2014) meta-analysis as providing evidence for systematic phonics. But as I show in my response it is just not true – the meta-analysis finds that various methods produce the same effect size and there is no interaction between methods (meaning that there is no evidence that one method is better than another). And then there is Suggate (2016) who shows that in a meta-analysis that systematic phonics does not have a long-term effect on reading performance (if anything, he finds systematic phonics does the worst of various methods). I also show that the PIRLS results, the phonics check results, etc. also provide little to no evidence in support systematic phonics despite researchers citing this work repeatedly in support of phonics. So I just disagree with you conclusion. I would be interested in which meta-analysis or findings more generally provides the best evidence in your opinion and I’ll respond.

    I do agree with you when you write that there is a “ logical case that it makes sense to teach kids this systematicity within the writing system explicitly”. But as I detail in various papers, there is systematicity in the grapheme-phoneme correspondences (GPCs) and systematicity in how morphemes are spelt. Indeed, you can better understand GPCs when you understand that GPCs are embedded within morphemes (see Bowers, J.S., and Bowers, P.N. (2018). Progress in reading instruction requires a better understanding of the English spelling system Current Directions in Psychological Science; can download from my webpage). According to systematic phonics children should only focus on one set of regularities and ignore others at the start. It is a logical possibility, but so is it possible (even plausible) that all regularities should be taught from the start. As noted above, the empirical evidence does not support systematic phonics, and accordingly, it seems like it should be a matter of priority to test the hypothesis that children should learn all regularities rather than only a subset at the start. In the following paper we detail strong theoretical motivation for this approach. Bowers, J.S., & Bowers, P.N. (2017). Beyond Phonics The Case for Teaching Children the Logic of the English Spelling System. Educational Psychologist, 52, 124–141. 2017

    Great to have this exchange, hope others comment too!


    1. Hi Jeff,

      I’d still cite the NRP study as coming out overall in favour of systematic phonics. I did read the Camilli et al (2006) study, but found it very hard to figure out what they’d actually done, and why they’d made the decisions they did about which factors to control. No doubt there are many, many different ways these data could be carved up – I really have no way of knowing whether things they have done have introduced other confounds. So, imperfect though it may be, I place more weight on the results as reported against the original set of criteria.

      More generally, a response you made on Twitter made me realise the distinction you draw between phonics and GPCs. So, to get things back to basics, let’s just focus on GPCs. Do you agree with all of these points:

      1. It is important to teach GPCs right from the commencement of reading instruction.

      2. It is important to do so explicitly.

      3. It is important that children get to practice the GPCs they have been taught, close to the time that they’ve learned them.

      If you agree with all these points, I think we have a very firm platform of agreement about the evidence to build on.


      1. Dear Anne, Camilli showed that there is a straightforward flaw in the logic of the NRP. The authors claimed that the meta-analysis supports systematic phonics over non-systematic phonics and whole language, as in the following:

        “Students taught systematic phonics outperformed students who were taught a variety of nonsystematic or non-phonics programs, including basal programs, whole language approaches, and whole word programs” [bold added]. (NPR, 2000, p. 2-134).

        But the study did not even test that claim because they combined non-systematic and non-phonics (whole word) conditions together. On top of this, as I detail in my review, the meta-analyses included studies that should have been excluded based on the exclusion criteria, used the wrong baseline in at least one study in a way that contributed to the large effects, most studies where not RCTs, and as Torgerson et al. (2006) noted, there is evidence of publication bias in the RCTs that were included. At minimum, researchers should cite the Camilli et al. study and should not use the NRP to support systematic phonics over non-systematic phonics (or whole language) given the design of the meta-analysis.

        But this meta-analysis is now almost 30 years old. As I show, the evidence is weaker still in subsequent meta-analyses (people focus on the NRP as it shows the largest effects), and all subsequent meta-analyses in support of systematic phonics have the same logical flaw pointed out by Camilli. It is also worth noting that the Galuschka et al. study that Pam cited the other day in her blog ( has such a basic flaw in statistics that it should never be cited for support of systematic phonics (see my comment in that blog), Suggate (2016) shows that the impact of phonics does not last, and indeed, descriptively does the worst in a comparison of multiple methods. I just see no empirical evidence that justifies the strong commitment to systematic phonics.

        You ask me if I agree if GCPs should be taught from the start. As noted above there is no empirical evidence for this (given the lack of evidence for systematic phonics), but yes, given the logical organization of English, I think this makes good sense (perhaps systematic phonics does not outperform whole language because it is a bad idea to ignore meaning with written words at the start?). But systematically teaching GPCs does not mean systematic phonics. I’m pasting a paragraph of Bowers and Bowers (2018) just out in Current Directions in Psychological Science that highlights some of the differences between systematic phonics and Structured Word Inquiry (SWI). But perhaps the main thing (not mentioned in the paragraph) is that systematic phonics and SWI could not be more different in their approaches. SWI focuses on the meaningful organization of words from the start, and teaches children to make sense of the system as a whole. This can be a life-long interest. By contrast, phonics focuses on non-meaningful correspondences, and if all goes well, is put aside in grade 2 or 3. Children do not attempt to understand the spelling of “site” words like and , and indeed, some researchers want to change the books children read early on with decodable texts.

        From Bowers and Bowers (2018)

        To avoid any confusion, it is important to emphasize that the explicit instruction of orthographic phonology — how grapheme-phoneme correspondences work—is a core feature of SWI. However, unlike phonics, SWI considers grapheme-phonemes within the context of morphology and etymology. For example, consider the crucial role of morphology for understanding the grapheme-phoneme correspondences in the word [react]. Absent morphology, it is not possible to determine the graphemic structure of the [ea] letter sequence: is the [ea] a digraph corresponding to a single phoneme (pronounced /iː/) or two graphemes associated with distinct phonemes? Morphology clarifies the phonology. The word [react] has the morphological structure [re + act] and this rules out [ea] as a digraph because graphemes never cross morphemic boundaries. Successful application of learning from SWI negates the mispronunciation /riːkt/ (homophonous with “reeked”) while successful application of phonics learning makes “reeked” and “react” equally plausible readings. More generally, the morphological context provides an explanation as to why specific graphemes- phoneme correspondences occur in words (e.g., why the word [action] includes the [t] rather than the [sh] grapheme to represent the /ʃ/ of [action]).

        1. uhmmm I see some critical words got lost in the paragraph I pasted. Here it is again:

          To avoid any confusion, it is important to emphasize that the explicit instruction of orthographic phonology — how grapheme-phoneme correspondences work—is a core feature of SWI. However, unlike phonics, SWI considers grapheme-phonemes within the context of morphology and etymology. For example, consider the crucial role of morphology for understanding the grapheme-phoneme correspondences in the word [react]. Absent morphology, it is not possible to determine the graphemic structure of the [ea] letter sequence: is the [ea] a digraph corresponding to a single phoneme (pronounced /iː/) or two graphemes associated with distinct phonemes? Morphology clarifies the phonology. The word [react] has the morphological structure [re + act] and this rules out [ea] as a digraph because graphemes never cross morphemic boundaries. Successful application of learning from SWI negates the mispronunciation /riːkt/ (homophonous with “reeked”) while successful application of phonics learning makes “reeked” and “react” equally plausible readings. More generally, the morphological context provides an explanation as to why specific graphemes- phoneme correspondences occur in words (e.g., why the word [action] includes the [t] rather than the [sh] grapheme to represent the /ʃ/ of [action])

          1. Jeff,

            When I’m talking about Camilla et al, I’m referring to the 2006 paper. I understand the problem addressed in the 2003 paper, and that is fine (agree it should be cited), but that analysis still showed a significant effect of systematic phonics.

            As you know, I think the question of whether GPCs are taught in the context of morphology and etymology in initial instruction is an empirical one – we simply don’t know the answer. My gut feeling is that this is too complex for very early readers, and the single letter-sound level should be the unit of focus initially – but I’m the first to admit that we haven’t done the study to establish that. So I don’t think there is a huge amount of point in debating it. But I still think it would be good if we were in agreement about whether and when GPCs should be taught (and whether the learning of them should be practised).

  9. Hi Jeff,

    I’ve read your detailed analysis and I think you make some important points – and I agree with you that there has been a tendency to take some of the phonics results at face value and without sufficient attention to the devil in the detail.

    On reading it all through, though, I still come away with conclusion that the overall pattern of results favours systematic phonics over non-systematic phonics. You are right to point out that the control in the major meta-analyses such as NRP is “non-systematic or no phonics”, but as you note yourself, there are going to be very few studies that have no phonics at all of any kind (I find it hard to imagine how that would even be possible – what does the teacher say when a kid asks?) And, yes, the effect size is reduced when you remove those no-phonics studies, but it is still there on most measures (at least before further moderator analyses are done). So my overall take on the pattern of data would be that (1) phonics is good (2) there is a graded effect such that the more systematic and explicit the phonics is, the better the results tend to be (though the differences are not as pronounced as some of the literature has suggested).

    This, combined with the logical case that it makes sense to teach kids this systematicity within the writing system explicitly (as you are also arguing for in the case of morphology), makes me conclude that we should just agree to teach phonics systematically and then get on with other things! That doesn’t preclude there being a range of ways to be “systematic” in phonics instruction, and I think there is definitely scope for more research here.

    Anecdotally, from things parents and teachers have told me, one real positive of explicit, sequenced exposure to GPCs in the very early stages of reading (and associated decodable readers) is that it gives children a huge feeling of success – they can do it themselves, and they see themselves as “readers” right from the outset. I don’t know if there is any work on phonics vs whole language instruction and academic self-concept, but it would be interesting.

    Thanks again for your very detailed study of all the meta-analyses – I think it is important that all reading researchers engage with this complexity, and avoid the temptation we all have to see things through a particular lens.


  10. This is such important insight and analysis. I look forward to reading the whole of the article. It’s time to move into new understandings that are based on the reality of the language rather than false pedagogical claims. (I wonder what would motivate someone to make false pedagogical claims…)

    BTW, even though I like NPR far more than the NRP, you probably want to stay with the latter here.

Leave a Reply

Your email address will not be published. Required fields are marked *