Don’t let phonics denialists move the goal posts after PIRLS 2016

December 9, 2017

A big difference between scepticism and denialism is that sceptics can identify what evidence would persuade them and then change their position when they have it. Denialists will move the goalposts, acting as if the evidence has no consequences for their arguments. When dealing with denialists you have to constantly remind them of their own arguments otherwise they will simply move on.

The recent PIRLS results, that assessed reading in “4th grade” in 61 countries, and allowed for comparisons between countries and with previous scores was a perfect example of this. This was the first PIRLS cohort to have been through the phonics check. They indicated who was right and who was wrong in their predictions about the effects of the phonics check. The results showed that since the previous round of PIRLS, reading scores in England had improved (to their highest ever) with spectacular gains for the weakest readers.

Supporters of the phonics check had claimed it would improve reading, and that it would be particularly beneficial for the weakest readers. The supporters of the check had made a correct prediction and had reason to feel vindicated. Phonics denialists, who had not predicted the improvement, and often claim that phonics won’t help the weakest readers, immediately started to find reasons these results were irrelevant to the debate. A summary of the denialist arguments can be found here and roughly speaking are:

  1. England is not the only country with an increase in reading scores.
  2. It’s not the only increase ever to have occurred in England.
  3. For the students in the sample, the phonics test score, was not a good predictor of PIRLS scores.
  4. You cannot attribute an improvement to one policy.

Point 3 is just ridiculous. There was a statistically significant correlation of 0.52 on tests 4 years apart where the worst performers would have had extensive remedial help after the first test. The data actually proves the opposite point to the claim being made.

However, the other points are correct. They also show complete amnesia about what phonics denialists were arguing before these results came out.

Did the phonics denialists claim that the phonics check would be followed by improvements in reading but that would prove nothing? Of course not. They suggested reading would be harmed. Opposition politicians, trade unionists, activists and education establishment figures wrote a letter to the Guardian in 2012 claiming:

Many of those working in primary education fear that these tests could undermine rather than benefit children’s progress and development.

Yet somehow reading has improved.

Did phonics denialists claim that an improvement in reading would be irrelevant to assessing the effects of the phonics check before there was evidence of an improvement in reading? Of course not, when Australia was looking at implementing the phonics check a few weeks ago, this was in one of the most popular articles arguing against it:

As the test has been has already been in use for six years in England we are fortunate to be able to learn from their experience. A major evaluation of the test conducted by the Department for Education in England found that the test is not delivering improvements in literacy capabilities, and in fact, is delivering some unwanted side effects…

…What does the research say?

Claim: The phonics test has improved reading results in England since its introduction.

Evidence: Year 1 children in England are certainly getting better at passing the phonics test. Over the past six years, pass rates have increased by 23%. This means around 90% of Year 1 children in England can now successfully read nonsense words like “yune” and “thrand”.

However research has found that the ability to read nonsense words is an unreliable predictor of later reading success.

And so far, the phonics test in England has not improved reading comprehension scores.

So in September, phonics denialists were arguing that a lack of evidence for an improvement in reading showed the phonics test might not be working. Now in December, they argue that an improvement in reading is irrelevant to assessing the effects of the phonics check.

Did phonics denialists argue that PIRLS results were irrelevant to analysis of the policy on phonics before these results came out? Of course not. Janet Downs of the Local Schools Network, a loose online coalition of hard leftists and education progressives, dismissed the 2016 PIRLS results as irrelevant in articles like this one this week. Back in 2012,  she was arguing in an article entitled “10 year-olds from England and Northern Ireland shine like PIRLS in global reading test” that PIRLS results showed that reading was fine without the phonics check and the long tail of underachievement was perfectly normal:

Schools minister, Elizabeth Truss, said the rise in reading performance was “encouraging” but England was being held back because of a “too long tail of under-performance”. While it’s true that 5% of English 10 year-olds didn’t reach the lowest level, 3% of Singapore’s 10 year-olds didn’t reach it either. And in Australia and New Zealand 7% and 8% respectively didn’t reach the lowest level.

Having thrown cold water on these results, Truss used them to promote the controversial phonics test. But PIRLS doesn’t test decoding – it tests comprehension.

So it seems that analysis of PIRLS results was fine when claiming that everything was okay before the phonics check, but cannot be used to judge improvement after the test.

On three counts here, the goal posts have been moved. Of course, there is a limit to what the PIRLS results show. What they show is not that the phonics check definitely caused an improvement in reading, but that the predictions of the supporters of the phonics check were right, and the opponents of the phonics check do not actually respond to evidence. The latter is something we already knew, but now we have even more evidence that opposition to the phonics check is based on ignoring evidence.

Oh, and one final point, one of the claims made by the opponents of the check, and by phonics denialists generally, is that assessing decoding does not tell us anything useful about reading ability in the future. The 2012 letter I quoted above claimed:

…we do not believe that this will help parents know how well their children are learning to read…

They will not show whether a child can understand the words they are reading, nor provide teachers with any information about children’s reading ability they did not already know. The use of made-up words – like snemp, osk, jound – risks confusing children for whom English is a second language and those with special educational needs, and frustrate those who can already read

This argument has already been discredited by looking at Key Stage 1 reading results. However, we now have additional evidence. Here is a graph of average PIRLS score against phonics check scores:

The link between decoding scores and reading performance here is remarkable. So this is another case where, if the phonics denialists were honest, they’d be admitting right now they were wrong about this.



  1. Reblogged this on The Literacy Echo Chamber.

  2. Thank goodness for your forensic analysis of the debate regarding phonics and phonics results in England. Thank you yet again.

    Added here at the International Foundation for Effective Reading Instruction:


    It has been so time-consuming and tiring to argue for the uptake of phonics teaching in the first place (for decades now) such that is gratifying to have you on board to comment on events so forensically.

  3. hi
    the national report for England (https://www.gov.uk/government/publications/pirls-2016-reading-literacy-performance-in-england) from which the impressive predictor of phonics test Year 1 (e.g. better predictor than self-reported amount of books at home) in this post is taken also notes that the picture of phonics and PIRLS results internationally is a mixed bag both in terms of positive change/no change/negative change and PIRLS country rankings.

    i wonder if we can speculate that the weaker Year 2 phonics test correlations with PIRLS shows diminishing effects of phonics as student reading proficiency rises?


    • Is that English language nations only?

      • good question, the two on the list in the UK gov report that seem not English – Austria, Oman are also ones that improved apparently (section 9.1.8)

  4. […] The data shows that the Mathew effect continues to strengthen – those that have much continue to gain much; those who have little continue to gain little. The gap is widening and this rings alarm bells for SSP and PSC supporters. We need to take action now to ensure all students, not just the ones at the top, are given appropriate instruction and support to learn to read. Make no mistake, if students have basic code skills, their reading comprehension will improve too. Many will tell you otherwise; don’t believe them. […]

  5. […] Don’t let phonics denialists move the goal posts after PIRLS 2016 […]

  6. Reblogged this on The Echo Chamber.

  7. […] Don’t let phonics denialists move the goal posts after PIRLS 2016 […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: