Posted on 6 January 2017


When we write up the research for Best Evidence in Brief we always try to be as objective as possible. It’s not up to us to have opinions, dear reader, that’s your job. But inevitably we have our favourites, and so, at Christmas, we indulge ourselves by tweeting our twelve favourite research articles of the year. They’re not necessarily the best, they’re not even the most significant, they’re just the ones that we liked the most. And here’s why…

  1. Promising results for Numicon

The National College’s Closing the Gap: Test and Learn project was exciting, carrying out a whole mass of experiments in one go. There were 50 small innovation projects (including small RCTs), and a larger-scale RCT testing several approaches suggested by schools. Unfortunately, it looks like it’s going to be a one-off, and that’s a shame, because I bet that a second iteration (like most second drafts) would have been much better. Although we didn’t make much mention of them in our BEiB write-up, the innovation trials were a terrific idea and I hope they can inform the way that we run the innovation evaluations in our Research Schools project. One concern I have, though, is how the results of these innovations are published and presented for future generations. They provide snippets of interesting information that can be built on further, but if they are tucked away in the annexe of a PDF in the corner of the DfE website, just under a sign saying “Beware of the Leopard”, how are people going to find them? I’d like a way of publishing small-scale research such as this that makes it easily accessible, while giving a fair assessment of its wider applicability.

  1. Ignoring the evidence on ability grouping

The evidence on ability grouping is that it doesn’t work, so why is it still so prevalent? This is a fascinating bit of research that goes beyond “just because” and looks at why it might be happening. We don’t often include studies that use a Foucaultian ‘analysis of discourse’, and I’m still not entirely sure what it means, but looking back after a year, it’s interesting to see how this analysis connects, in a not terribly encouraging fashion, with our new, post-truth world.

  1. Struggling scientists

It’s nice when we can include a study that would be easy to replicate if you wanted to. This was a straightforward intervention aimed at increasing student engagement with science. There were three clear intervention groups – stories of scientists struggling emotionally, stories of scientists struggling intellectually, and a control – and each intervention would be relatively easy to try yourself. The results of the intervention weren’t stunning, but they were positive, and it’s more actionable than the larger studies we often include.

  1. Picky eating

One for the parents; or for those who have to work with parents. This study provides reassurance that some picky eating is nothing to be worried about, while providing a warning flag that other behaviour may be more of a concern, and is worth looking into further. Arguably it’s a bit outside our remit, but I hope that it’s useful for teachers (at least in the early years) to know about these kinds of issues.

  1. Support is better than sacking

The previous issue of BEiB had included a study of how sacking teachers had improved outcomes for students, so it was good to find this study so soon after. According to this, if instead of sacking them you provide coaching from high-performing colleagues, you get pretty much the same positive result. I suspect the atmosphere in a school that takes this latter approach would be a lot better.

  1. Childcare

I like this study because it challenged my preconceptions. I think it’s best if young children stay at home and are looked after by their parents. I don’t like to see toddlers being shunted off to childcare at silly o’clock in the morning because their parents have to work two jobs to keep body and soul together. I hate that. It’s one of the things that make modern life rubbish. But guess what, according to this study I’m wrong, and it’s the best thing for them – the kids, that is. (I still don’t like it, though.)

  1. Sleeping through

Another study for parents who worry – this one looks to see if there’s a trend with sleep problems. It suggests that for most children sleep problems are on the decline by the age of five. However, for those where problems are escalating at that age, it is another red flag that there may be more going on.

  1. Children and tests

This isn’t a particularly robust study, but I thought it provided some interesting information about what children think about taking tests. At the time it was published, there was a lot of hysteria about the SATs, and indeed about this research. Some articles, if not exactly misrepresenting the survey, at least selectively reported it. As part of our mission to present the facts, ma’am, just the facts, I thought it would provide a useful service.

  1. Learning styles

I know that you, dear reader of BEiB, are spectacularly well-informed about evidence, but newcomers to the parish are sometimes a little bit sketchy. So it is that we will occasionally repeat things that you may already know (“If classification of students’ learning styles has practical utility, it remains to be demonstrated.”) I also have a sneaky suspicion that, like Fu Manchu, the world will hear of learning styles again. Studies of learning styles have tried to teach students in their preferred learning style, assuming they will learn better. But, as Dylan Wiliam suggests, the secret may be to make sure that they have sufficient challenge by also teaching in the style they don’t like. Maybe there’s an optimum balance to be struck here, between challenge and comfort, and once you’ve “diagnosed” a preferred learning style you would know how to mix it up, which would put learning styles back on the agenda…!

  1. Text messaging

I used to work in mail order and loved doing A-B tests of different mailing approaches. Long letter v short letter; brown envelope v white envelope. You can test a lot of these options reasonably cheaply, which is why letters asking for donations are surprisingly long, because the companies know that they work better than short letters. Text messages are relatively cheap and, because they’re discrete, simple interventions, they’re also cheap and simple to do studies on. The Behavioural Insights Team has done many studies on tweaking texts to encourage fine payment, or attendance at job interviews, and it must be ripe for more exploration in education. The quality of information sent to parents by schools is not always, ahem, of absolute top-notch quality, and there must be mileage in testing messages that are carefully thought out and written. This particular intervention did just that, and found a tiny effect. But the great thing is that it doesn’t take a lot of work to implement. It’s not a year-long CPD programme that is really difficult to embed, it’s a series of text messages. It’s a perfect “marginal gain” which will make a small positive difference for hardly any effort.

  1. Baby simulators

Negative results are good, particularly (as with the previous childcare study) when it challenges your assumptions. People can be sniffy about studies that demonstrate the bl**ding obvious, what an old colleague of mine used to call “No Shit Sherlock” studies. And that’s a shame, because if we don’t have evidence about the bl**ding obvious, we could still be wrong. There’s no robust evidence that flossing works, for example, but dentists are all so convinced that it works, that it would be impossible to fund a decent trial. So it’s bl**ding obvious that if you give a teenage girl a crying baby to look after for a week, it will put her off getting pregnant. Won’t it? No, it makes it more likely that she gets pregnant. That’s a good thing to know in itself, and is a useful reminder that the bl**ding obvious doesn’t always work.

  1. Breakfast clubs

This is a simple story with a twist. Before-school breakfast clubs work. That’s great. But wait, it’s not because they’re giving kids breakfast. “… it was not whether more pupils ate breakfast at all that made the difference, but whether more were going to the school breakfast club. It may be that school breakfasts are more nutritious, or that attending the club effectively prepares pupils for learning.” So, although there are surely many children who are going to school hungry, it seems as if the value of breakfast clubs is in giving those (and other children) the time to get “school ready” before the day starts in earnest.

Jonathan Haslam, IEE

Leave a reply

Your email address will not be published.