Viewpoint: Wither the survey?

Mike Savage and Roger Burrows

It is commonplace to argue that the proliferation of new kinds of data and information has created huge social changes that we still do not really understand. One interesting example is the worry of social scientists that their preferred data sources and modes of analysis are being challenged by the rise of new digital data sources. These debates are causing considerable reflection about the very rationale of the social sciences in a digital age.

It is worth noting that the academic social sciences have enjoyed 50 years of remarkable growth. Just after the Second World War, only around 3% of academics were social scientists, and they were dwarfed by those working in the natural sciences, medicine, and in the arts and humanities. Now, if one includes those working in the ‘applied social sciences’, such as business or educational studies, they make up nearly half the total academic workforce. This expansion went hand in hand with their successful use of two related research methods: the national sample survey and the interview. The Government Social Survey began during the Second World War, and in subsequent decades social scientists played a key role in developing survey methods, notably in the introduction of longitudinal and panel surveys, such as the British Household Panel Study, which began in 1991. Until the 1950s, qualitative (or ‘in-depth’) interview methods were mainly regarded as being appropriate only to psychotherapists and related disciplines, but have become increasingly widely used in the social sciences as it was seen how they could provide insights into people’s identities, values and experiences.

During these glory years of the expansion of academic research, social scientists were rather aloof to those working in the commercial sector, notably in market research. They saw their methods as superior in terms of providing a detached and rigorous account of social relations. This led to a situation in which the worlds of academic social science and commercial research were largely detached from each other – incidentally, a far cry from the situation in the middle 20th century when market researchers such as Mark Abrams influenced all kinds of social research, and Claus Moser was President of the Market Research Society (and Ralf Darendorf as late as the early 1980s). It is this stand-off that is challenged by the proliferation and research potential of transactional data. Traditionally, social scientists have emphasised how their methods allow them to delineate a range of social variables, notably social class (measured in different, more ‘accurate’ ways than in the market research categories), which could be seen to predispose people to act in particular ways. This was then seen as ‘causal’ analysis, seen in opposition to ‘descriptive’ studies concerned with the clustering of particular actions.

However, in the current situation, where data on whole populations are routinely gathered as a by-product of institutional transactions, the sample survey seems a poor instrument. To give a simple example of the merits of routine transactional data over survey data, Amazon.com does not need to market its books by (a) predicting on the basis of inference from sample surveys the social position of someone who buys any given book and then (b) offering them other books to buy which it knows on the basis of inference people from that social position also tend to buy. It has a much more powerful tool. It knows exactly what other books are bought by people making any particular purchase, and hence can immediately offer such books directly to other consumers when they make the same purchase. Hence the (irritating, though often tellingly useful) screens offering versions of ‘Other people who have bought x have also bought y’, which confronts the Amazon customer. Similar principles are used by supermarkets through data gathered by their loyalty card schemes, which enable them to identify for any given customer – without knowing anything very much about their personal, ‘social’, characteristics – what other kind of goods they might be liable to buy if they buy, for instance, organic bananas. They can hence bypass the principles of inference altogether and work directly with the real, complete, data derived from all the transactions within their system. In this framework, the value of analysing the ‘causal role’ of social variables becomes altogether more uncertain, and the appeal of comprehensive descriptive analyses becomes more apparent. Some leading sociologists, such as Andrew Abbott from Chicago or Bruno Latour from Paris, have explicitly called for a new kind of ‘descriptive sociology’.

We should not exaggerate the novelty of this situation. Survey-based market researchers also saw the threat from the development of customer databases as early as the mid-1980s, yet predictions of the end of survey-based research proved unfounded because specific databases provided information only on actual customers, and the data also lacked profile and contextual information. However, the potential to link data sets to provide complete maps of the population, through the postcoding of data, and methods derived from social network analysis might have more profound implications.

These debates have encouraged the main public funder of social science research in the UK, the Economic and Social Research Council, to make new efforts to gain access to commercial data, and are also leading some academic social scientists to reflect on the value of working more closely with those in the private sector, to share expertise and reflect on the potential of new data sources for academic concerns. It is not yet clear what this might lead to, as there are issues of confidentiality, market sensitivity of data and different working practices that need to be addressed. So, is it possible that a closer alliance between academic social science and commercial research might now be firmly back on the agenda, to build links between surveys and new sources of data for research purposes? If not, does the traditional social survey face extinction? What do IJMR readers think the future holds for the ‘traditional’ survey?

Professor Mike Savage is Directorof the ERSC Centre for Research on Socio-cultural Change.

Roger Burrows is Professor of Sociology at the University of York & Co-director of the Social Informatics Research Unit (SIRU).

International Journal of Market Research 50(3), 2008

Would you like to respond to this Viewpoint? Or perhaps you have an idea for another? Responses and new submissions are welcome. They should be emailed to the IJMR, where they'll be considered for publication.

What's New - Membership - Company Partner Service - Members' Area - Code/Guidelines - Qualifications - Training - Awards
Events - Networking - Publications - Media Info - Market Research - Search - Site Map - A-Z Directory - Contact Us - Home