The data trail

John Davies, English Heritage’s economist, shows how we can use data to find new ways to understand the impact of cultural and sporting investments.

Advances in statistics are creating new ways to understand the impact of cultural and sporting investments.
Traditionally, assessing the impact of cultural and sporting facilities has involved collecting survey data and estimating economic impact using a set of assumptions (e.g. how much spill-over economic activity is created per pound spent). This can be time consuming and using assumptions to assess impact is not entirely satisfying. In its report The art of the possible, the CASE programme therefore undertook research into how analysis of existing data can be used to assess the effects of cultural and sporting investments.

Through time

A key ingredient in statistical assessments of impact is having data from before and after investments were made. This helps establish causality as it allows you to compare how the impacts of interest were affected by the change in (creation of) the facility, as opposed to other factors that might affect things.
As a simple example of the kind of data that can form the starting point for more detailed analysis, the picture left shows the creative and cultural businesses around the Sage concert venue in 2009 and you can see how this compares to 1999 (before the Sage’s 2004 opening) on page 75 of the report(PDF 2.2mb).
The growth in the number of creative and cultural businesses in Gateshead between 1999 and 2009 was 86 per cent: higher than that for the UK as a whole (79 per cent) and Newcastle (60 per cent). Factors besides the Sage’s opening could have contributed to this e.g. other investments in the area, and there are other areas which have had higher growth rates than Gateshead. More detailed analysis of the drivers of business numbers would therefore be needed to establish the extent of the connection. Statistical techniques used in this area often go under the label of regression analysis, with Differences in Differences and Panel data [PDF 156kb] being types of regression that are particularly relevant in evaluating impact.

Into space

Data doesn’t just have to be analysed over time, it can be analysed over space too. Technologies such as Geographical Information Systems (GIS) are allowing spatial questions to be examined that, until recently, it would have been impossible to answer. For example, we can find out the extent to which people value living near to cultural or sporting facilities.
House prices provide a measure of how much people value living in a place. Obviously many factors affect prices, besides being close to a cultural or sporting facility e.g. house size, location, the characteristics of the local area, state of the economy etc. Regression analysis can control for these – i.e. factor out their effects on house prices to isolate the effect of living closer to the facility. To do this the analysis requires sufficiently rich data on these factors (before and after the opening) to be available.
A recent example of this kind of work is research by Ahlfeldt and Kavetsos on the effects of the new Wembley stadium. As you can see in the diagram below. this found that out to 5km there was a premium associated with living closer to the new stadium (controlling for other factors). The interpretation being that people see a benefit from living closer to the stadium, and are prepared to pay more for it. This premium peaked with the completion of the stadium arch, but was still found to exist after stadium completion.

Boundaries

Statistical analysis of existing data may not always be appropriate to assess impact. Smaller cultural and sporting sites are arguably less suited to this approach. Undertaking this type of analysis also requires specialist skills. With public investments it is important to assess whether any activity/benefits have been displaced from elsewhere (for example, to what extent is the growth in businesses around the Sage due to new start-ups or relocations from elsewhere), and how to analyse this statistically is an area which is still developing.
This kind of analysis also depends on sufficiently rich data being available at a local level on the factors you want to assess and control for. Excitingly, a wealth of new information is becoming available that has the potential to greatly expand the impacts that can be examined.

Enter the data world

There is a revolution going on. Large volumes of information on people’s activities and perceptions are flowing in from social media and mobile devices: tagged photos, twitter traffic, search terms etc. Surveys can also be done through mobiles, allowing spatial-survey data to be collected; for example, the happiness data collected by LSE’s mappiness project.
This kind of information allows a whole new set of questions to be investigated: how much happier are people made by cultural and sporting activities, or by being surrounded by historic buildings? What are the buildings people are most interested in taking photos of? What kinds of things do people talk about near cultural sites, as opposed to other places?
Whether it will be possible to answer these questions robustly (and what the resulting implications will be), remains to be seen. But this is an interesting area for future exploration which gets to the heart of what CASE is about; allowing us to obtain a deeper understanding of the benefits that culture and sport bring to society.

Follow case

• Subscribe to the CASE newletter


• On twitter using the #CASEprog hashtag


• Read more CASE blogs