Social science and value judgements
I had a nice conversation with one of my favourite philosophers at UCL recently, a man I respect greatly. His main research area has been the philosophy of mathematics and he is a pure and honest thinker driven, it seems to me, only by the highest intellectual ideals and integrity. We talked about reason and rationality and the human difficulty of reacting to statements about probability.
And yet I found myself disagreeing with him over a fundamental principle of social science and where we are in the study of politics, economics, sociology and the like.
The issue hinges around the relationship between quantitative and qualitative analysis: how both are currently used; how they should be used; the problem of collecting, analysing and interpreting data and the significance of both sorts of analysis in the final reckoning for what is significant in social science.
I had recently read a book – Ferraris for All – which is an attack on those who espouse green economics and sustainable economics, at least in so far as those approaches might halt a growth in GDP – as currently defined. The book is odd in places and a little shrill in others, but it is well researched and packed with supporting data from serious academic studies to show that, overall, growing GDP is a good thing and standing in the way of GDP growth is a bad thing.
So, as with so many books of this nature and a similar calibre, one reads the material, believes the data and the academic studies cited, and ends up in a position of being fairly convinced by the arguments of the book. But, hang on! You then read another book – by, say, Herman Daly or Tim Jackson – and are equally convinced by the data, the seriousness and the arguments – but these guys arrive at completely different conclusions. How, then, can anyone possibly decide on these issues? One data set contradicts another, the greatest experts in the world disagree on all aspects of the data,and before long we are arguing over minutiae of sample size, methodologies of data collection and skewing of statistics, completely missing the big picture we are trying to see and the problems we are trying to address.
I tried to tell my philosopher friend that a quantitative analysis of many situations in social science can therefore be far less helpful than it appears. Indeed, the appearance that the data ‘supports’ a position may be illusory. My philosopher friend’s position was that ‘one cannot escape the numbers’ – and this, I believe, is the current orthodoxy. However, I would like to resist this, at least in part.
At the foundation of social sciences is, indeed (!), the study of society. And society is bound together by values – human qualitative things. Traditionally many of these values had a religious sanction and, more recently, evolutionary biologists have attempted to explain most of them (the incest taboo, outlawing of murder etc) in terms of the benefit to the individual, group or species. But whether they come from religion, from evolutionary pressures or from elsewhere, the success of the values depends on them being adopted by the societies in which they originate. And that process, that adopting of a position on a value judgement, is largely independent of whether the position is supported by any kind of quantitative data or statistical analysis.
To take a positive example, the idea of the Separation of Powers, usually regarded as proposed by John Locke, developed by Montesquieu and systematized in a pragmatic fashion by the Founding Fathers of the United States, is not an idea that could conceivably have been developed by quantitative analysis. Perhaps a modern social scientist could reply that we could have reached this position, by running various computer models, perhaps defining characteristics of individuals and watching how their lives evolved in various political structures which we have managed to symbolise in code. But this is not realistic. There is currently no conceivable way to encode something of such complexity. And even if somehow the many possible models could be captured in code, this still would not guarantee, or perhaps even give a good idea as to, the political system that would be most desired by any given population. No, the only test for such an idea of the Separation of Powers is that it works in practice – and that depends on it being accepted and sanctioned by society in which it functions, and that, in turn, depends on the values held within that society.
To take a negative example, we could look at the current debate in economics over growth, sustainability, fossil fuel consumption and the related nexus of debates around capitalism, socialism and the role of finance. As I say above, the data here is, if not fundamentally contradictory, then certainly bewildering, even for the best living experts. What then can we do to adopt a position which we feel is right and which we are confident to defend? This is where we should resort to values, ways of guiding our actions which we are prepared to defend in the face of competing and contradictory data. One value might be thrift, another might be sufficiency, another might be parity of economic reward, another might be the avoidance of waste. (I recall a section from Obama’s book: ‘when your drapes cost the same as the yearly wage of the average American – you have enough.’)
This is not the place for a detailed defence of such a philosophical position on values – for that see, e.g., Elizabeth Anderson’s excellent Value in Ethics and Economics. But we should be aware that at bottom it is values which drive our societies and the economic structures on which they are built. Values are qualitative things; there is currently no conceivable way to quantify such value judgements.
Values, as qualitative things, are best defended by qualitative arguments; arguments which appeal to our humanity, our rationality and our common sense. Sometimes these arguments may be well supported, but when they are not, we should not give up such arguments as hopeless. Rather we should work harder to make the case, to convince our fellow human beings and – if we are arguing for a new outlook – that it is an outlook based on new values worth committing to.
The too-hasty flight to data, statistical analysis and a quantitativist view of nature can be seen as part of the reductionist scientistic attempt to reduce all argument to simple, general forms, based on accepted scientific formulae, mathematics and computer code. There is no denying that such approaches constitute a vital aid to understanding and progress in many aspects of life and study. This is not a manifesto to give up on such crucial enterprises. The gathering and analysis of data has become a pivotal part of modern life. But we need to remain circumspect. In the social sciences the quantitative approach can only ever go so far. In the end the efficacy and success of the results of research in the social sciences depend on the humans that use those results and the values they espouse. These values are inherently qualitative things. We should be prepared to defend them as such.