Feature | Data to Measure Diversity

Results from Audience Finder survey raise important questions about the inclusivity of arts organisations...

Audience Finder has two main aims: the first is to help individual cultural organisations use visitor data in their audience development work, and the second is to create a useful picture of audiences at a national, regional and sector level.

In both cases, Audience Finder sets out to help the arts sector find audiences rather than simply monitor them as a bureaucratic bean-counting exercise. This principle is really important when it comes to describing the diversity of our audiences.

We were determined that Audience Finder should use demographic data to contribute to making positive change. This consideration shapes what data we ask organisations to collect, and how we layer, analyse and present it. It draws on data from a number of different sources to help us understand the diversity of our audiences.

Socioeconomic diversity

We get a good picture of the overall socioeconomic diversity of audiences by looking at Audience Spectrum segmentation tool, which assigns one of ten lifestyle profiles to every household in the UK. It gives us a good idea about the outlook, income levels and cultural engagement of collective audiences and shows starkly what we might expect: that those most highly engaged in cultural activity tend to be the most privileged, well educated and relatively well-off.

But with over a third of households in England now tagged as arts-bookers, we can see that a considerably wider group of people do engage, albeit less frequently. Audience Finder tells us about these people – what they do, when and where – and proves there is the potential for organisations to build on those relationships.

From a national perspective, it pinpoints the gaps in provision that need to be closed if we are to see a real change in building more socially diverse audiences.

Monitoring the protected characteristics

But what Audience Spectrum cannot tell us about is age, ethnicity, sex or disability. Like all profiling systems, it gives us an idea of the likelihood of people to live a certain lifestyle. General assumptions about socioeconomics based on address are valid, but this method is not an appropriate way of monitoring protected characteristics (age, race, religion, disability, sexual orientation, etc).

For one thing, inferences about personal identity on the basis of address are not statistically accurate. More importantly perhaps, this approach is not considered ethical, as it is important that people are able to self-identify their ethnicity, disability and other personal information.

So the other key source of insight about diversity in Audience Finder is via a standard survey, asking audience members the same questions directly about their cultural engagement: basic behaviour, motivation, opinions and identity. Organisations using the Audience Finder survey can be confident that they are monitoring the protected characteristics of their audiences in a robust and ethical way (exceeding the accepted standard required by Arts Council England of larger National Portfolio Organisations).

New insight

The survey data is the key to understanding the demographic diversity of our audiences. This is the first time we have been able to look at a whole year’s worth of data across so many venues to see a genuine national picture. There are many interesting themes emerging around audience diversity that demand further exploration. Here are three in outline.

Are our audiences becoming more culturally diverse?

White people are over-represented in the overall data set. 92% of all audience members who completed the questionnaire were white (86% of people in England and Wales are white). Asians and British Asians are under-represented at 3.2% of respondents compared with 7.5% of the population. People describing themselves as Black or Black British are even more under-represented, at just 1% of respondents compared with 3.3% of the population.

But this gloomy picture may be changing. Although small, the most over-represented group is also the fastest growing – the multiple ethnic background group.

More remarkably, if we cross-reference age with ethnicity, we see that younger audiences are significantly more ethnically diverse. The ethnic mix of the under-35 audience looks far more like the mix of the whole under-35 population in England and Wales – a balance even more marked for the 16 to 24s.

How does disability affect people’s engagement?

Just as with the stats on age and ethnicity, the scope of the Audience Finder data can help sharpen thinking on disability issues. In line with the Office of National Statistics census, our survey asks people if their day-to-day activities are limited because of a health problem or disability. If the answer is yes we ask respondents to specify ‘a lot’ or ‘a little’. 9.4% of the population answered the census question ‘yes, a little’, compared with 8% of Audience Finder respondents.

While we clearly cannot be complacent with this kind of result, it is encouraging that disabled audiences in this category are reasonably well provided for. But the picture looks very different for those answering ‘yes, a lot’. Nearly 8.5% of the population falls into this group, but only 2.1% of our audiences.

This wide disparity suggests that our access policies and practices are scarcely reaching people in this category, many of whom are in need of regular social care support and have distinct needs from those whose needs we may be more familiar with.

Do we serve everyone equally?

The most important reason to monitor the protected characteristics is not as a tiresome funding obligation but because it helps us understand whether we are treating everyone fairly, and to adjust our activities accordingly.

This is a vital aspect of diversity policy, so it’s important to find out about perceptions of how people are treated alongside the personal information we collect. In Audience Finder we measure this in a simple, universal way called the net promoter score (NPS) – a standard measure of how likely someone is to recommend something to someone else.

Without going into lengthy detail here, black, Asian and minority ethnic (BAME) audiences, younger audiences (especially under 25s) and disabled people in the ‘activities limited a lot’ category all give much lower promoter scores than other audiences. If we want to do something more active about the flat-lining diversity of our audiences than wait for population change, probing the underlying issues here could make a world of difference.

How do we respond?

More generally, these stats clearly show that some organisations are leading the way, achieving far more diverse audiences than the national benchmark, and indeed than the population in their catchment. Galleries, for example, have considerably more ethnically mixed audiences than other artforms.

Some organisations have very high net promoter scores for all their audiences. As well as celebrating their success, we need to understand more about what such organisations are doing today, and about their journey of audience development.

About the Audience Finder Survey Sample:

  • Analysis is of data collected between April 2015 and March 2016.
  • 276 organisations did the survey – 145 Arts Council England NPOs and 135 other cultural organisations. Last year’s sample was skewed towards the museum sector, making up 34% of the sample, although in future years it will give a more balanced picture across artforms. The other types of organisations submitting data were galleries, multi-space venues, performing arts, touring companies, outdoor arts festivals, concert halls and a small group of independent cinemas.
  • There is a good regional mix, so you can be assured this is telling a broadly national story. The best represented regions are London and the north west. 18% of the venues in the sample are based in each, while the north east is least represented with only 4% of its venues carrying out the survey.
  • There were a total of 104,000 questionnaires conducted by museum and heritage organisations within the period. This is a huge sample and gives us reassurance that results are robust and representative of real attenders. Taking Part is often quoted as the main source of information about engagement but is based on a survey of 15,000 members of the public.
  • We use the same categories as the Office for National Statistics, as these have been extensively tested with members of the public and using them means we can always make an accurate comparison with the population.

First published by Arts Professional, 4 November 2016, artsprofessional.co.uk