Where does reputation come from?

THE announced the latest edition of its reputation rankings last October. The amount of information is quite limited: scores are given for only the top fifty universities. But even that provides a few interesting insights.

First, there is really no point in providing separate data for teaching and research reputation. The correlation between the two for the top fifty is .99. This is unsurprising. THE surveys researchers who have published in Scopus indexed journals and so there is a very obvious halo effect. Respondents have no choice but to refer to their knowledge of research competence when trying to assess teaching performance. If THE are going to improve their current methodology they need to recognise that their reputation surveys are measuring the same thing. Maybe they could try to find another source of respondents for the teaching survey, such as school advisors, students or faculty at predominantly teaching institutions. 

Next, after plugging in a few indicators from other rankings, it is clear that that the metrics most closely associated with teaching and research reputation are publications in Nature and Science (Shanghai), highly cited researchers (Shanghai), and papers in highly reputed journals (Leiden).

The correlation with scores in the RUR and QS reputation rankings, citations (THE and QS), and international faculty was modest.

There was no correlation at all with the proportion of papers with female or male authors (Leiden).

So it seems that the best way to acquire a reputation for good teaching and research is publish papers in the top journals and get lots of citations. That, of course, applies only to this very limited group of institutions.