Reframing models of arts attendance: Understanding the role of access to a venue. The case of opera in London

Arts attendance in England has, in recent decades, been the subject of several surveys focusing on how individual factors such as socio-economic status, education, ethnicity and age influence attendance. These surveys have been used to create small area estimates of arts attendance. But other studies of the use of public facilities suggest that access to a venue would be highly predictive of attendance. This paper compares administrative data on opera attendance in London with small area estimates of opera audiences, and finds a systematic geographic bias in the errors of the predictions, related to a lack of information about the location of venues. It demonstrates that a model using 2001 Census data and a simple accessibility index better predicts attendance, and more accurately locates audiences. It concludes that, by focusing on individual-level explanations in order to understand cultural engagement, funders have failed to examine the effect of their own investment.