Revisiting the Debate Over the Davis Phonics Pamphlet: Part 1

According to this article on denialism in public health, one of the tactics most often used by denialists is the use of:

…individuals who purport to be experts in a particular area but whose views are entirely inconsistent with established knowledge. They have been used extensively by the tobacco industry since 1974, when a senior executive with R J Reynolds devised a system to score scientists working on tobacco in relation to the extent to which they were supportive of the industry’s position. The industry embraced this concept enthusiastically in the 1980s when a senior executive from Philip Morris developed a strategy to recruit such scientists (referring to them as ‘Whitecoats’) to help counteract the growing evidence on the harmful effects of second-hand smoke. This activity was largely undertaken through front organizations whose links with the tobacco industry were concealed, but under the direction of law firms acting on behalf of the tobacco industry. In some countries, such as Germany, the industry created complex and influential networks, allowing it to delay the implementation of tobacco control policies for many years.In 1998, the American Petroleum Institute developed a Global Climate Science Communications Plan, involving the recruitment of ‘scientists who share the industry’s views of climate science [who can] help convince journalists, politicians and the public that the risk of global warming is too uncertain to justify controls on greenhouse gases’. However, this is not limited to the private sector; the administration of President George W Bush was characterized by the promotion of those whose views were based on their religious beliefs or corporate affiliations, such as the advisor on reproductive health to the Food and Drug Administration who saw prayer and bible reading as the answer to premenstrual syndrome. A related phenomenon is the marginalization of real experts, in some cases through an alliance between industry and government, as when ExxonMobil successfully opposed the reappointment by the US government of the chair of the Intergovernmental Panel on Climate Change.

Phonics denialists often use a similar tactic. So much so that you can often predict which “authorities” they will end up quoting. Perhaps Stephen Krashen, an American professor who is reported to have claimed “[A]ny child exposed to comprehensible print will learn to read, barring severe neurological or emotional problems”. Perhaps one or all of of the trio Torgerson, Brooks and Hall who managed to do a review of the evidence on phonics which rejected all but 4 of the studies on reading comprehension and then concluded that what remained was insufficient to support the consensus that phonics teaching benefited comprehension. Perhaps Michael Rosen, the children’s author who described phonics as “barking at print”. Perhaps Henrietta Dombey, another educationalist, who, despite not being any kind of neurologist, gave us her interpretation of brain scans as part of an argument against phonics. The same denialist organisations will also pop up again and again, particularly NATE and UKLA. There’s no shortage of people with opinions that are against phonics, often with highly respected positions in the education establishment, but with an utter lack of good evidence for those opinions. While some denialists will catch you out (for instance, Dr Mary Bousted pretending her PhD was about phonics) mostly you encounter the same people and sources again and again.

The latest “go to” expert without evidence for the phonics denialists is Andrew Davis. Another educationalist (although as I understand it, this time with a maths background)  he wrote a pamphlet for the once prestigious Philosophy of Education Society of Great Britain. This received some publicity at the time of publication for the sensationalist claim that teaching a child with systematic synthetic phonics was “almost a form of abuse”.  A recent letter in the TES (admittedly one whose signatories include Davis himself, Brooks, Torgerson, and representatives of UKLA and NATE) cited it and started the discussion going again, so I thought I’d take the time to look at it. However, it turned out I had already discussed it here. In that post I noted Davis’s main argument:

In this short book I am going to argue that research into the teaching of reading involves some fantasies. These take the form of imagining that specific teaching methods could be identified, and that their efficacy would be open to empirical investigation. I show that if any schools were actually implementing such strategies, the adults responsible would have abdicated their role as teachers. In reality, implementations of SP in any one school will not and should not precisely resemble those in other schools and in any case, current research into SP ‘effectiveness’ is not informed by a detailed blow by blow description of what actually happens in the classrooms concerned. Hence, it is never really made clear what the research is actually investigating. If teachers are actually teaching, there will be and should be nothing common to all SP programmes. The effects of drugs or fertilisilisers can, of course, be investigated using orthodox scientific methodologies, but we lack the equivalent here in terms of teaching approaches.

I then responded to this argument:

Now, the limits of scientific methods to isolate and evaluate what happens in the classroom is a real issue. I’m certainly sceptical about a lot of education research for that reason. However, the claim that we could never, even in theory, objectively evaluate a teaching method is as extreme a denial of science as anything you will hear from homeopaths or creationists (who are also often prone to claim that science cannot hope judge their claims). The claim that all the research in an entire field (not just the hundreds of studies on phonics, because this argument applies equally to all teaching methods) is a particularly extreme one. It entails that all those who have conducted empirical research in teaching methods were mistaken, and all those who found statistically significant results were deluded. Not only that, but if they were to test more extreme cases, say the efficacy of teaching by telepathy, their results would still be invalid and teachers would be fully entitled to teach telepathically. To dismiss empirical research on this scale is as extreme as dismissing all the evidence of climate change, in fact it is pretty much the same argument, that we cannot aggregate data, that climate change denialists use.

As a note, I should probably add that the defence could be made that the author, despite using an argument that could be applied to all research into teaching methods, actually meant to treat phonics as a special case; that it is only in the case of phonics or perhaps synthetic phonics specifically that researchers would have no idea what method was being used. But if it is inherently impossible to identify the teaching methods used as phonics, or synthetic phonics, why write a pamphlet condemning a policy encouraging synthetic phonics? If the teaching method is indistinguishable from other methods, it cannot possibly be enforced. The policy would be meaningless, and can safely be ignored. The argument assumes the very thing that would make the argument irrelevant. You cannot oppose the imposition of a teaching method by arguing that there is no method being imposed.

I also pursued this with Davis himself on the TES forum here (where he posts as “ded6ajd”) but couldn’t get a clear argument from him  He admitted his argument would cast doubt on more teaching methods than just systematic synthetic phonics, but he wouldn’t identify clearly where we could draw the line on which teaching methods were observable and which weren’t, or identify the wider consequences of his claim.

The problem is not simply that Davis has speculated but not found new evidence (although that does invalidate some of the uses this pamphlet is put to in arguments), it is that he has rejected existing evidence because of a speculation. The possibility that it is impossible to objectively evaluate a teaching method is not one which can reasonably be entertained without paying due attention to the examples in the real world of people objectively evaluating teaching methods. Apparently without considering the claims of a single study, Davis, in the words of the editor, concludes that “[w]hatever it is that empirical researchers take themselves to be doing when they investigate synthetic phonics … they are not investigating a specifiable method of teaching reading”. Davis’s argument is to claim that research methods which have already worked in practice would not work in theory. Just as speculation against the possibility of space flight might be considered flawed in light of the apparent existence of space shuttles, the speculation that there can be no actual evidence for phonics is flawed in light of the existence of evidence for phonics. All the speculation in the world that we cannot attempt to evaluate phonics is worthless in face of the fact people have evaluated phonics.

Of course, while the extreme nature of the claim should be highlighted, and it justifes the use of the word “denialist” to describe the argument, an extreme position could still be correct. Unfortunately, this mistake is not made in isolation. It is part of a pamphlet which argues against the imposition of teaching methods, particularly the teaching of systematic synthetic phonics. As I argued above, if teaching methods (or even just the methods of teaching reading) cannot be identified then they cannot be imposed, for there can be no possible consequence for not implementing them. Nor can the claim that teaching synthetic phonics is “almost a form of abuse” be squared with the picture painted of an invisible, undetectable method. I suspect that if Davis truly believed that teaching methods were as unidentifiable as he claimed, he wouldn’t concern himself with the detail of primary literacy policy, he would be considering the consequences of this insight for the whole system. How much of what is done in education assumes that teachers and observer can identify what method they are using? Phonics is a minor detail if teachers are genuinely that ignorant of their actions.

There are other claims and material in the pamphlet that I could have explored. There is plenty of the usual denialist misconceptions about the nature of reading and meaning, and attempts to rebrand the usual denialist methods as another brand of phonics, but it is this self-defeating argument which (other than a citation of Torgerson et al by the editor in the introduction) is the only real attempt to address the empirical evidence on phonics. This, perhaps, misled me to think that nobody would actually pay much attention to this pamphlet, but it has been cited again and again in online debate and there have been attempts to defend it from the obvious criticisms.

In part 2 I will address some blogposts defending this pamphlet.