The Campaign for Real Education is a pressure group which aims to raise standards in state education in the UK. It is not politically affiliated, although its proposed changes to education policy – grammar schools, a return to a ‘traditional’ teaching philosophy and increased parental choice – are more typical of right wing or conservative agendas.
McGovern appeared on BBC Radio 4‘s Today programme this morning, to discuss an academic paper published in the Economic Journal. The study analysed primary school performance data to show that, contrary to what some might expect, having a high proportion of pupils from non-English-speaking backgrounds in a school class does not reduce its performance.
I’m not going to discuss the paper, nor the CRE’s policies. I would just like to quote some of McGovern’s responses to the paper, and leave open the question of his credibility as an educational advocate.
“It’s a fact that if you’re teaching a class and half the children can’t speak English, it’s going to take up teacher time, it’s going to have a negative impact on those children.”
It’s not a fact. The reason McGovern was on the programme in the first place was to discuss a paper which proves exactly the opposite.
“This research does actually say that the raw data shows that it does have a negative effect, but they try to offset that by saying that it’s all to do with social background.”
The data on class performance has been adjusted for a known bias. I haven’t read the paper, but from McGovern’s comments and my understanding of statistics, I would guess it’s something like the following: classes of a lower social background perform worse on average by a known degree; classes with a high proportion of non-English-speaking pupils tend to be from lower social backgrounds as well; so the data has been adjusted to eliminate this bias. This is a sound piece of analysis: otherwise, if you compare the non-English-speaking classes against the national average and identify a lower performance, you could be wrongly attributing the negative effect of the lower social background to the presence of non-English-speakers. You have to compare like for like: classes of equal social background, and therefore expected equal performance, with greater and fewer non-English-speakers. This is presumably what the study does, and concludes there is no negative effect that can be directly linked to non-English-speakers.
The validity and accuracy of this bias adjustment may be debatable in its details, and this is what the peer-review process of academic papers is designed to provide: a quality check on such calculations. McGovern doesn’t discuss this though. Instead he seems to dismiss all statistical analyses and bias adjustments as a fraudulent hand-waving exercise. His attitude suggests a profound misunderstanding of statistics, and of science in general.
“I don’t think we should take this too seriously, it’s about as reliable as those experts who are saying that grade inflation in examinations represents a genuine improvement.”
Putting McGovern’s shoe-horned mention of his own personal bugbear, grade inflation, to one side, what he effectively did here was to say, “I don’t like the results of this study, so let’s ignore it.”
The CRE’s home page deplores the recent history of education policies driven by governments “without any evidence to support them”, suggesting a desire to see evidence-based policy. However, there’s no mention of evidence-based policies anywhere else in the CRE’s various manifestos and subject pages. It seems to be driven as much by dogma as the various government policies it opposes. It even boasts of being seen as the voice of “common sense” (a quantity which is notoriously unreliable in public policy, and generally the antithesis of hard, reliable data), and criticises statistical data analysis in point 13 of its manifesto.
Back to Today…
After Evan Davis defended the study against McGovern’s unreliability accusation, by pointing out that it wasn’t published by politicians with a point to spin, but by disinterested academics from the London School of Economics, McGovern made the following extraordinary statement:
“They’re economists, and we know what record the economists have had…”
There are two ways I can think of to understand this swipe at economists. The more generous interpretation is that it was simply an act of desperation, by a man confronted by solid, knock-down evidence against his position, who grasped at whatever bit of rhetoric he could think of to undermine it.
The more worrying alternative is that McGovern is truly dismissive of the entire discipline of economics. This would imply that he is unaware that there’s a big difference between macroeconomists, some of whom make predictions about national economies, and some of whom failed to predict the global financial crisis (which is presumably the “record” he mentions), and other economists whose work is nothing to do with economies: those who specialise in analysing and interpreting the social and political data on which public policies like education are (ideally) decided.
If this is the case, it further undermines McGovern’s understanding of science, statistics and mathematics in particular, and academic subjects in general. Where, for instance, does an understanding of the basic tools of economics, which help us understand the world we live in, and make good policy decisions, sit in the CRE’s plan for a revised academic curriculum?
Davis also pointed out the absurdity of a supposed champion of academic standards dismissing the validity of an entire academic discipline and the results of an academic study. At this point McGovern changed tack and fell back on his own anecdotal evidence that non-English-speaking pupils create extra work for teachers and negatively impact their classes, and wouldn’t budge, even when Davis quoted a particular piece of data from the study, that the presence of Polish children in a class actually improves other pupils’ performance in maths.
When the discussion returned to McGovern, he first (erroneously?) quoted Churchill, saying that “experts should be on tap, not on top” – another statement of opposition to evidence-based policy – and then gave his most controversial opinion:
“I think what we should do is tell parents the number of children in a class who do not speak English, and I think parents will be choosing other schools.”
I don’t think much more commentary is needed. School choice as a means of voluntary racial segregation, and based on a willful disregard of evidence, and misunderstanding of science. Is this a man who should be a leading voice in education policy?