World University Rankings 2018: the journey in numbers

Browse the full results of the World University Rankings 2018

This year is the third in which Times Higher Education has generated the World University Rankings entirely in-house, and it seems appropriate that we look back over the changes and also peer ahead to the future.

In those three years, the reach of the rankings has expanded significantly – from 400 universities in 2015-16 to 1,000 universities today. In fact, although we publish a table with 1,000 institutions from across the globe, we hold data that would allow us to rank more than 1,100.

At the same time, our methodology has bedded down and now requires fewer annual adjustments. This year, the only changes of note have been a slight improvement to how we handle our papers per academic staff calculation, and an expansion of the number of broad subject areas that we use. The tweak to papers per staff came about directly from a university that contacted us with a suggestion for an improvement (although we do not always act on such suggestions, we do listen carefully). In essence, it provides a fairer way of dealing with orphan papers – those that sit in a subject where the university has no staff – or at least where it does not report any staff.

The expansion in subjects sees psychology, education and law become separate from social sciences. Psychology always sat uneasily within social sciences – it could equally well have been part of life sciences or clinical and health; allowing it to stand alone resolves a dilemma. Education and law also make sense as individual subjects, in terms of size and in terms of career choices. Looking at the size of remaining subjects, I do not think it likely that we will expand the range of subjects in the future. Our experience is that a more detailed ranking is hard to justify from a data perspective – and anecdotally we have come across some very bizarre consequences when other organisations have attempted it.

Some of the other initiatives that we have enjoyed this year have included the expansion of our Asia and our Latin America rankings, as well as a novel exploration of our Young University Rankings that divided the ranked institutions into generational cohorts.

However, the major step forward this past year was the launch of our new series of teaching-focused rankings. We have always been clear that although the World University Rankings include some teaching metrics, their primary focus is unashamedly on research-intensive universities. We believe that this will rightly remain our premier ranking, but now it is complemented by its new cousins.

Our first teaching ranking, the Wall Street Journal/Times Higher Education US College Rankings, covered more than 1,000 US higher education institutions – ranging from traditional marquee names such as Harvard University, the Massachusetts Institute of Technology and Stanford University, to some outstanding institutions that few outside the US may have heard of, such as Dordt College in Iowa.

These new rankings consider four areas:

• Resources available for teaching
• Engagement with learning
• Outcomes for students
• Environment for learning

The College Rankings were followed by our Japan University Rankings, which will, in turn, be followed by our European Teaching Excellence Rankings as well as by a Global Business School Rankings (again in association with The Wall Street Journal).

Looking slightly further ahead, we are also working on a few changes.

First, we are exploring how we can better reflect the achievements of specialised universities. At the moment, these are excluded from all our rankings. We would like to be able to include them in the subject rankings, even though this can cause difficulties. One hurdle is that these more specialised institutions frequently do not teach at an undergraduate level. Although that is not a problem for the institutions, two of our metrics use the number of undergraduate degrees awarded in their calculation.

We also intend to expand the number of universities that we list in each of the subject rankings. Our previous limit of 100 was initially set by print requirements: extending this will let some excellent universities shine more clearly.

Although I have said that we will not look to rank universities at a more detailed level than our 11 broad subjects, we do want to find a way of providing prospective students with greater insights. So we hope to explore an alternative approach that will be more robust and plausible than a traditional ranking. We will trial this later this year.

Download a copy of the World University Rankings 2018 digital supplement

But on to the elephant in the room. I have discussed the longer-term vision for the World University Rankings at many of THE’s data masterclasses (which are certainly worth attending – we would love to see you at one).

This year we have begun to gather opinions on what I describe as “WUR 3.0”. Looking at the history of our methodology, we launched Version 1.0 in 2004 (and if you really want to see it in action, there is still a version out in the wild). When we shifted production to Thomson Reuters in 2010, we adopted Version 2.0, and you could argue that after the move of our bibliometric supplier we are now on Version 2.1.

But that still leaves us with a methodology that is approaching eight years old. Not long by university standards, but a lifetime in the world of big data. This year’s new undergraduates were in Year 5 at primary school when it was introduced.

We have already received some great input and some useful ideas – and I will be following these up in blog posts and articles throughout the year. I still expect the World University Rankings 2019 to use our current approach, but 2020 seems a reasonable target for the new approach.

Duncan Ross is data and analytics director at Times Higher Education.