John Bald is a former Ofsted inspector and has written two books on the history of writing and spelling. He is a Vice-Chairman of the Conservative Education Society.

When I argued in my last piece that setting up the Education Endowment Foundation had been an error, and that it should receive no more public money, I did not expect it to respond with a letter of congratulation, as indeed it did not. However, we have, since 1946, had a National Foundation for Educational Research, with an income of just under £18 million a year, and capable of carrying out or commissioning whatever research the government requires. Sir Kevan tells us in his letter that the research into ability grouping is to be evaluated, not by the researchers whose work the EEF has already published, but by the NFER. One has to ask, why the middleman?

Both the NFER and the EEF are registered charities, but is an organisation that receives almost all of its money from the government in fact a charity, or a quango? Since Labour developed the idea of the “third sector”, many charities have become government vehicles and are, in effect, quangos, a kind of pseudo-business, receiving money and providing services, but with no actual customers. The term “quango” has no legal status, but any organisation connected with education can be a charity, so the label is convenient. Last year, the 157 biggest charities had a combined income of £11.4bn, and it would be interesting to know how much of this came from government.

The EEF’s board includes Sir Michael Wilshaw, and Professor Becky Francis, Director of the London Institute of Education, whose intemperate language in reference to setting I criticised last week. The EEF may be justified in making grants to institutions led by its directors, though it seems odd if the director in question is leading the project. Sir Michael’s experience is in headship and inspection rather than research, but I can’t see how anyone could square a conclusion that ability grouping is “symbolically violent” with the opportunities Sir Michael has provided for children from Hackney to get top grades at A level and enter Cambridge University. Can Professor Francis, having used this term at an early stage in the project, continue to investigate it impartially? I suggest that she can’t.

The EEF’s approach to research has several major weaknesses. The first is its role as a one-stop shop, providing summaries of research for schools to use, an approach which gives far too much power to the evaluator, anonymous or not. In the case of setting, none of the raw data come from the present century, and so do not evaluate setting in the context of the academies programme, which dates from the establishment of Mossbourne in 2005. Mossbourne’s examination results in its first year outstripped all previous performance from comprehensive schools, and research has not kept pace with this development. In languages, where I have attributed the collapse of A level to the compromises made to accommodate mixed ability classes, often imposed on teachers against their professional judgement, there has been no research on grouping whatsoever. In this context, the reliability of research is not weak – there simply isn’t any.

The next weakness is that the body decides on what is to be investigated, and how. So, on behaviour, the effectiveness of the government’s policy of immediate detention, which has transformed behaviour in many academies, is not investigated, while the previous regime’s softer agenda of engagement and inclusion is the subject of multiple studies, including some imported from the US, with the vague gloss that they are “somewhat transferable”. Again, the US source organisation provides sticker-type evaluations of its own work, in terms of “strong” and “moderate” evidence.

As a reviewer of educational research for 15 years with the Times Educational Supplement, I would always go back to the raw data of a study, see how it was collected, what it actually showed, and what had been done to it. Many studies could not stand this level of scrutiny, with flaws including very small samples, poor control groups, lack of follow-up to show that any improvement was not a flash in the pan, and even poorly thought-out teaching methods. The current practice of combining the results of large numbers of studies- “meta-analysis”- as the EEF and Professor John Hattie do, can be seriously misleading if the studies on which the analysis is based are flawed, as errors are as likely to compound each other as to cancel each other out. Professor Hattie does not even trouble to read the studies on which he bases his analysis, and it is a pity that he has been listened to by so many politicians. I’ve discussed these and other errors, including elevating randomisation as a factor trumping all others, in a series of technical papers on my site.

Finally, Sir Kevan cites a useful study carried out by the EEF on teaching assistants, which showed that they could do valuable work with children in groups. This was an important element in Ruth Miskin’s success at Kobi Nazrul school in the nineties, and I have seen it used well in several other schools. Unfortunately, the study went beyond its evidence to challenge other uses of teaching assistants which might have been equally valuable if considered on their merits.

Sir Kevan will meet no argument from me on the importance of research in education. It won the battle for phonics, on a cross-party basis, and can provide genuine insights that go beyond the correlations that explain everything to some sociologists. Research, though, can be, and is, used as a political lever, by directing attention to the issues researchers think important, and avoiding areas where evidence would almost certainly go against their views. This is what is happening with the EEF and why, if it is to continue as a charity, it should do so without public money. There are plenty of competent avenues available to the government to carry out the research it needs, and to present it without fear, favour or spin.