One of my colleagues is a leading scholar of the work of an English poet on whom he has written some major studies. Partly as a result of a chance personal connection, in recent years he helped choose the exhibits, write the captions, and make other contributions for a display at a small museum devoted to this writer’s life and work. I spent no small amount of time in 2011 and 2012 chivvying the poor staff at this museum. Could they supply visitor numbers? Sorry, could they please document those numbers in a publicly verifiable form? Did they have evidence of what visitors to the exhibition made of the experience? Did they ask them to fill in questionnaires, did they have a comments book? Sorry, could they provide extracts in a duly authenticated form? What was the evidence of the benefit the visitors derived from their visit? Sorry, I mean evidence of what the exercise calls ‘change in their behaviours’? And so on.

…But in reality these kinds of effects, even if desirable in themselves, as no doubt many of them are, do not testify to the quality of the research at all. My colleague’s scholarship on this poet would still have been of the same high quality whether or not he had happened to be involved with this museum, let alone whether we could demonstrate beyond doubt that a thirteen-year-old visiting with a school party had written in the comments book that the exhibits were ‘ace’.

Stefan Collini, Speaking of Universities (London; New York: Verso, 2017), pp 50-51