For the authors of “The Spirit Level”, not one, but two books have been published just to argue how wrong they have got things. The essence of the original case is that in countries where there is less difference between the best and worst off (like Japan, Sweden and Norway), everybody is healthier and has fewer social problems. In other words, if the rich shared their wealth out more, they would benefit as much as those lower down the social scale.
Naturally this does not go down well with right-wing commentators, especially in the United States, where anything that smacks of redistribution is identified with communism (or as it is spelled in American-English, treason). So counter-arguments have been pouring off the presses and through the media to try and quash this dangerous idea.
At the heart of “The Spirit Level” is a set of data about wealth, health and social issues. Again and again, the authors found a correlation between low levels of social exclusion and higher life expectancy and the like. (If you want to get a sense of the original argument, you can find the key slides at www.equalitytrust.org.uk/resources/slides.)
For data analysts working on marketing problems, finding repeated correlations between independent variables is the breakthrough moment in a project. Once you see how several factors work together to produce a result, it starts to become predictable. Then you can use a range of stimulii, in the form of marketing activity, to ensure you get the desired outcome each time.
The mathematical principles in use by both marketing and policy analysts are the same, even though there are major differences between the data sets they use. Most public sector information is at population level and lacks much depth. Marketing data is far richer, but less widespread.
Fortunately for the marketing analyst, the outcomes being pursued are narrow – increasing sales of a product or service or retaining customers. The political analysts are aiming at major changes in society which would often require vested interests to give something up.
Yet there is less separation between the two areas than might appear. Every decision about consumption has a social implication, from how it is paid for (for example by borrowing to buy it), to where the product is manufactured (what were the working conditions in the factory?) and how it will eventually be disposed of (recyclability of packaging and materials).
You may not be at risk of vilification by lobbyists. But correlation in marketing is applied, not pure mathematics and is therefore about the real world where actions have consequences.