“Statistics are like a bikini. What they reveal is suggestive, but what they conceal is vital.” Aaron Levenstein.
The above quote by Levenstein is certainly true. It is becoming increasingly apparent to us that geographical data can have the same effect. Researchers and data crunchers need to be careful. Flawed research is used by policy and decision makers all the time. I have a couple of examples to enlighten you.
Back in 2015, Steiner, Frey, and Hotz wrote an academic paper on European Capitals of Culture and Life Satisfaction. In this paper they did not look at data from cities, they looked at data from Regions (NUTS/Statistical regions, level 2). In the case of big cities this might amount to the same thing. London, Europe’s biggest city, is the equivalent of two regions. Let us consider the Republic of Ireland. A small country. Dublin was European Capital of Culture in 1991 and Cork, in 2005. The Republic of Ireland is split into two regions. Both Dublin and Cork are in IE02 – as are Dundalk and Tralee. According to Google maps, Tralee is 299km from Dublin. It takes 3 hours and 37 minutes to drive and the fastest public transport route takes 4 hours. Dundalk is 336km from Cork It takes 3 hours and 26 minutes to drive and the fastest public transport route takes 4 hours and 26 minutes. The data that Steiner, Frey and Hotz used to assess cities of culture and life satisfaction included data from Dundalk and Tralee. Culture Europe do not claim that the European capitals of culture programme benefits or even has any impact on people living 300 km away. I would argue that the life satisfaction of people in Dundalk, on the Northern Irish border, was not particularly influenced by the fact that the people of Cork, on the south coast, were being excessively cultural in 2005.
In an example closer to home. This is a map which highlights the age profile of wards in Lancaster district. This pale coloured area is big. On the whole it is quite sparsely populated. There are a lot of sheep but for most of it, not many people. It is pale because it designates that the median age profile of the population of this area is 18-32. The median age profile of all the surrounding wards is 39-44. What is going on? The little area ringed in red is the campus of the University of Lancaster. The little blocks are one of our rural venues, Abbeystead, a rural idyll that can’t take shows in the spring because of lambing. If you just looked at this map you might programme to attract an young audience. In reality the young audience live in a small enclosed community in one corner of the ward would never travel to Abbeystead and have probably never heard of it. Abbeystead is where his Grace, the Duke of Westminster hangs out when inspecting his estates in Lancashire. (Though admittedly as the current Duke is only 28 he fits the median age profile of the ward.)
The more metropolitan the area is, the more likely it is that neighbours will share similar characteristics. In London or Manchester, or even some of the large and medium sized towns in Lancashire, there is a high chance that people who share the same postcode fall into the same income bracket, have the same interests, social class, levels of educational attainment and values. This is what the big-data analysis, undertaken through postcode searches by tools such as Audience Finder, relies on. In Abbeystead people from all backgrounds, live cheek by jowl with the same postcode from the Duke of Westminster to the unskilled farm labourer (and they are all involved in lambing). On a national scale, these rural anomalies are statistically negligible but to an organisation promoting directly in these areas they are vital and render the big data analysis largely irrelevant.
Data and statistics can be useful but a degree of scepticism is vital before anyone acts. Otherwise we’d be programming student bands into Abbeystead Village Hall.