Interesting data, poor inferences

Apparently, the UK higher education system is a ‘knock-out’ performer.  Although the UK system is ranked 8th in the world — not bad at all — what is holding it back on that measure is central investment. Per capita, that is, higher education is not well funded in the UK, and the system as a whole pays the price. However, we can compensate for that, correcting for the relative poverty of the country (compared to, say, the US or Switzerland) and thus its relatively low level of public investment. This yields a different ranking, with the UK way up in second place. (Behind Serbia, interestingly.)

All very interesting. It makes you wonder from where all the calls for the UK system to be over-hauled in one way or another are coming (how marking and degrees are scaled being an old chestnut). If it ain’t broke, don’t fix it. On the other hand, is also makes you wonder what would happen if there was some higher level of investment…

The THE article first cited though, goes one step further. The reason why the UK system punches above its weight in this way is the Russell group — i.e. the concentration of funding into roughly 15% of the institutions. (We should also include a few other Universities who are not technically in the RG, but who enjoy big funding and strong research reputations, such as Leicester or Durham.) The article, or rather the chap being quoted, cites the recent REF results as evidence. Evidence for what, exactly? That research is concentrated, to be sure; that funding is concentrated, definitely. But it is certainly not evidence that concentration delivers ‘bang for the buck’. On the contrary, the fact that an army of researchers, working at the other 85% of institutions, are still doing international quality research with bread-crumbs of funding, much less infrastructure, and no endowments to speak of… THAT is bang for the buck. The cost per international-standard research publication at a ‘new’ university is a fraction that of the Russell Group, and I can prove it.

For example, let’s take two random Universities: Newcastle University and Northumbria. Now, Newcastle did pretty well in REF2014, staying in the top twenty nation wide. They submitted a total of 2806 outputs rated 3 and 4*. Northumbria are quite a way down the REF table, but comfortably in the top 100. They submitted 823 outputs at that quality rating. Now, clearly, there is a concentration of research in Newcastle, who produced three and a half times as many quality outputs. However, Newcastle received 35.6 million pounds in recurrent research funding in 2009-10, while Northumbria received … 3.26 million. In other words, with ten times the funding, they only managed three and a half times the number of quality outputs. It’s the same story when we compare Teeside and, say, York. The former required £6440 to produce an international quality paper, while York required £11,900.

Now, I know this is not the whole story. There is a difference between 3 and 4*, and also we may be comparing inherently expensive forms of research (laboratory sciences) with cheaper ones. On the other hand, in the above calculations, I haven’t included other big sources of funding, which are also highly concentrated, such as Research Council grants, European funds, and so forth. In the absence of more study, the evidence clearly goes against the ‘funding concentration’ hypothesis. (Hold on, who wrote this report that the THE is cheering about? — Universitas21 — and who are their UK members: all Russell Group. I’m seeing a pattern emerging.)


Studying in the provinces

Many newspapers ran a story about a study by London First and PWC — here is the THE’s version. The headline number is that between tuition fees and living expenses, international students in London are worth £2.3 billion net. By ‘net’ here meaning subtracting an estimate of the costs to the health service and other such services. What the THE didn’t add, is that fewer than a quarter of the international students in the UK are studying in London: it’s not difficult to estimate then that international students bring in the fat end of 10 billion to the economy. (Source: HESA.) These figures are borne out by a UniversitiesUK report found here (see Annex A). Add to this the more than hundred thousand EU domiciled students nation-wide, some of who to be sure pay less in tuition fees but just as much to live day to day, and the figure is more like £12 billion. Higher education in the UK is a major export industry, something that occasionally gets lost in the various debates.

What should a PhD thesis look like?

Fascinating feature piece in the THE on the PhD thesis as a written document.The basic contrast being made is between the traditional 80 thousand word dissertation and a thesis which is a bundle of peer-reviewed and published papers. The latter is not at all far from the PhD by published work, and as it happens only last week I was part of the examination of just such a PhD.

There are a handful of practical issues here. How long does it take to write up a thesis (time which, it is assumed perhaps wrongly, is not spent in the lab doing real work). If the bundle comprises a set of multi-authored papers, how will the examiners discern the candidate’s contribution? Are examiners able to give as much attention to a long traditional thesis, as distinguished from six or eight papers? Given the lead time in publishing papers, might such an approach result in a delay in a candidate obtaining their doctorate? In some subjects where book publication is valued, the traditional thesis is the obvious practical choice for a young, would-be academic.

But these practical issues are, frankly, technicalities which well-written guidance for both students and supervisors would solve. Tougher are the big issues. First of all, is the writing of papers — and especially of shorter pieces that can be published quickly, and get out there having an impact — profoundly more relevant in the contemporary research environment? That is to say, all other things aside, is a traditional PhD deeply misaligned with how science works?

On the other hand, a PhD is and always has been about a large-scale project — conceiving of it, planning it, and executing it, and doing so with independence and resilience — so how does the set of individual papers demonstrate these important skills?

[Update, May 30th: This report confirms what has been anecdotally known for a while: that most PhDs will not not get permanent academic posts. This means that preparation for work in a research environment outside academia is all the more important, and thus also making the shape of a PhD thesis suitable.]

Review – The Extent and Consequences of P-Hacking in Science

Just a quick blog on this very interesting paper

The Extent and Consequences of P-Hacking in Science by Megan L. Head , Luke Holman,  Rob Lanfear, Andrew T. Kahn, Michael D. Jennions

“A focus on novel, confirmatory, and statistically significant results leads to substantial bias in the scientific literature. One type of bias, known as “p-hacking,” occurs when researchers collect or select data or statistical analyses until nonsignificant results become significant.”

This is a very well written and readable paper which will be of interest to researchers and would be useful for discussion at under graduate level and post graduate level. What I like about this paper is the readability and clarity of the writing (often a rarity in many peer review papers) .

As well as the statistical aspects in the paper, it raises issues that could easily develop into a discussion about ethics in research  and the problems in the publication process.

The lead author recently carried out an Ask Me Anything debate on this paper on REDDIT a website with an enormous number of readers.  


Book review – Advice for new Faculty Members – Robert Boice

Boice’s book was published in 2000, but remains as relevant as ever. It is well known in the United States where some universities used to issue it to all new staff. The book is divided into three sections:

  1. Moderate work at teaching.
  2. Mindful ways of writing i.e. how to produce research papers.
  3. Socialising in your new university.

I am going to focus on the second of these which relates to research and writing of papers.

Boice is a psychologist and was very heavily involved in the induction of staff and follow up processes at his University. This gave him plenty of opportunity to run different experiments and approaches with new staff and to see what worked. He was taking about mindfulness and moderation even back then and despite my strong initial doubts whilst reading the chapters his results are staggering. Ever wondered if you have been doing things wrong? Or rather is there a better way? – well this should make you think.

Boice looked at the exemplars in his study group and compared them against the other staff.  Essentially Boice found that successful academics:

  1. Wrote daily even for a short period (often up to 30 minutes maximum).
  2. Knew when to stop.

These sessions he terms brief daily sessions (BDS) and he contrasts them with binge writing and also discusses the feelings and emotions associated with both approaches. Chapters 10 and 11 give detailed practical guidance on how to get your writing going (and the results at the end of those two chapters will make you think this is at least worth trying). Chapter 14 also deals with letting go of negative thoughts which is useful especially for perfectionist type academics.

What effect has it had on me? Well I am going to try this approach for the next paper I am writing, it is never too late to try a something new .

If you haven’t come across Boice before then I strongly suggest you have a read, especially if you are a new member of staff, or new to research and writing.

Three copies are now in the Staffordshire University library and you can find excerpts online.