Every holiday without fail, some controversy or other seems to grip edu-twitter. I remember something about eugenics last time around. Well, today the saga continued with a whole (mainly progressive) section of edu-twitter in uproar because Schools Minister, Nick Gibb, shared a blog post by Andrew Old, which critiqued the Education Endowment Fund’s meta-analysis of grouping by ability.
This was the tweet that got some people so enraged.
And here’s one of the scandalised tweets in response.
I’m not sure why this person felt a need to black out the names, as Nick Gibb’s tweet is in the public domain and easily accessible by all. In any case, the claim being made is obviously spurious. The whole point of the article by Andrew Old is that the EEF, a trusted research outfit, has seriously misrepresented the research. If one trusted research outfit is getting it wrong, then there is no guarantee that another would be infallible; thus the gatekeeping implied in the tweet is not only unnecessary but inappropriate. If not for an individual teacher blogger pointing out the serious flaws being presented as truth, the faulty statistics would just continue being widely publicised without being questioned. It is this engagement with research that is being celebrated by Nick Gibb in his tweet, rather than the ideological views of the blogger.
Let’s for a minute pause to reflect on the implications of the information dug out by Andrew Old in his blog. I’m not a mathematician, but I was able to understand the following:
- The meta-analysis of six research projects gives a +0.12 positive effect size for ability grouping.
- However, this figure masks inequalities in the effect depending on the ability group. For low attainers, the effect size is a negative -0.09, but this is based on only 2 meta-analyses which specifically measured the effect on low attainers.
- Given this disparity, the EEF decided to publicise the second figure rather than the overall one.
Now, I know many of us in the education sector are very focused on improving the life chances of the disadvantaged in society – and rightly so – but I can’t help feeling a bit nettled by this type of prioritising. If the overall effect size is a positive 0.12, then it would be even higher without the data on low attainers. Implicit in this is that ability grouping has a significantly positive effect for middle to high attainers, but that they don’t matter enough for this to be important. So all the spiel we hear about inclusion, and every child matters, is only true if that child is a low attainer. That’s not the kind of message I want to hear for my own child who is not a low attainer. He matters too.
As it is, I’m on the fence when it comes to ability grouping. I can see both the advantages and disadvantages, and it’s clear to me that whichever direction is taken, it has to be done well. I’ve seen lower set children consigned to Nowheresville. Equally I’ve seen mixed ability lessons that have been significantly dumbed down in order to be accessible to the lower attainers. So my problem is not so much with the findings of the EEF but with the deliberate misrepresentation of the data to speak only for the disadvantaged and not the many thousand other children who also matter.
Cherry picking education research to focus uniquely on the disadvantaged can also, counter-intuitively, lead to less than satisfactory outcomes for those very groups we want to give a leg up to. By publishing the headline -0.09 negative effect of ability grouping, the EEF is basically promoting the idea that setting or streaming are bad things – the term ‘symbolic violence’ comes to mind. But this may be doing a disservice to the lower attaining groups. By looking at the bigger picture and seeing that ability grouping can have positive effects on others, the question that should be asked is “What factors in ability grouping are causing this negative effect to happen for the lower attaining groups?”
Could it be due to behaviour, lower expectations, curriculum or the quality of the teachers assigned to the lower attaining groups? Could it be something entirely different? If we could identify the factors at work here, would it be possible to turn that statistic around so that ability grouping works equally well for all groups? These are the sort of questions the EEF should be asking, and attempting to answer, rather than succumbing to the bias of well intentioned social justice warriors decrying the whole thing as ‘symbolic violence’.