Changes in the THE World University Rankings, 2027 and 2028

Congratulations to all the universities that saw a rise in the 2027 World University Rankings. Today is a day to relax and enjoy your recognition.

But alongside the release came some interesting news from Phil Baty, Danae Banette and the THE data team. Let’s look at what this could mean, first for this year’s rankings and then for next year’s.

2027 changes

Today’s ranking doesn’t have any major methodological changes. It does, however, have a worrying data change, and an interesting presentational change.

On the data side we have been told that the number of votes in the THE reputation survey had dropped by over a third.

Reputation in trouble?

This drop seems to have been caused by challenges in populating the master contact list, and by the long-term drop in response rates that everyone working with surveys is facing.

Another less obvious issue is around the process of balancing the survey, as the dataset that THE traditionally used to understand global academic distribution was discontinued several years ago.

This is now making a large proportion of the ranking less stable, especially for universities lower down the ranking, where there were fewer votes anyway.

Expanded visibility

THE has also decided to give exact positions to universities in the top 300 rather than just the top 200. Below this, universities are put into bands of 25, then of 50 and so on.

There has always been significant pressure from universities to give exact positions, and it certainly opens the door for THE’s marketing opportunities. But the reason that the bands are there still holds, and is arguably even stronger now that the ranking has more than 2,000 ranked universities.

Think of it this way. With 2,000 ranked universities and only 1,000 unique scores available (100.0 to 0.0), there are always going to be ties for a position. In practice the distribution of scores is far from even, and they tend to cluster lower down in the ranking, which is why we see four universities tied in =299th place (Ahlia University, the Norwegian University of Science and Technology, Rutgers University–New Brunswick and the University of Hawaii at Manoa).

And the universities ranked 200–300 sit in exactly the area where any rankings changes, including the issues with reputation, will be most visible. Previously at least some of the noise would have been hidden by the banding.

2028 changes

I’m very glad that Phil Baty has been open about proposed changes to the 2028 rankings so early. These changes (at the moment) relate to the measurement of academic outlook through bibliometrics.

• Restricting self-citations. The exact methodology for this isn’t clear yet, and it has been described as a cap. However, with the increasing focus on academic integrity this is certainly going to be a welcome move.
• Rebalancing the four citation-based measures. The oldest, FWCI (known as Citation Impact in THE speak), is going to be cut back from 15 per cent of the ranking, and the other three measures will presumably be increased as a result.

Limitations and effects

The move to cap self-citations will probably be well received, and for most universities will have little impact. It’s not clear, however, how this change will address more sophisticated citation cartels. It will catch the unsubtle cheats only. It also won’t have any effect on the other hot topics of over-publishing and predatory publishers (and Research Productivity is still worth 5.5 per cent of the rankings).

Reducing the proportion of FWCI in the ranking will increase long-term stability. But at a short-term cost.

The three additional citation measures were initially brought in by David Watkins and Billy Wong several years ago, with the intention of fully replacing FWCI. There was even a draft ranking produced on that basis. The decision to stick with FWCI as the majority citation measure then was down to the instability that this produced in the rankings.

What does this mean for universities?

The combination of these changes with the weakening in the reputation survey is likely to create quite a lot of movement, especially in the newly visible 200s. Universities should expect to see variation in their apparent performance as a result, and would do well to think about how to discuss this with their stakeholders and the people they report to.

Of course, if you have questions about rankings and reputation, get in touch.

First published on LinkedIn in two parts: part 1 and part 2.