The last forty years have seen a decline in the attention paid to quantitative methodology and social statistics in university social science degree courses. This has happened despite changes in technology that have revolutionised the availability and quality of data, and made its analysis easier. The arrival of ‘big data’ has made it more important than ever to reverse this decline. The ESRC, British Academy and others are now determined to meet this challenge. John MacInnes reports.
In the 1960s, UK social science had an enviable reputation for innovation in the survey and other quantitative methodology led both led by academics (such as the work of John Goldthorpe and others on social mobility) and by the forerunners of the ONS. Their achievements, especially impressive considering the technology at their disposal, made the UK a world leader in longitudinal birth cohort studies and established a remarkable data infrastructure that has evolved into the suite of economic and social surveys undertaken by ONS and NatCen, successive and birth cohort studies and the Economic and Social Research Council’s (ESRC) new flagship ‘Understanding Society’ project with a panel of some 40,000 households.
Over the half century since then, however, a curious division has emerged between disciplines such as economics, which have continued to use and develop quantitative methods, and others, such as sociology, human geography or politics and international studies, where qualitative methods have become much more prominent. Unfortunately, the relative neglect of quantitative methods in these disciplines has become self-reinforcing. Fewer staff interested in this methodology has meant less attention being paid to quantitative methods and social statistics in undergraduate and postgraduate curricula, which in turn means that fewer younger academic staff have become knowledgeable about them or their potential.
This is not a healthy situation for any discipline, and international benchmark reviews have drawn attention to the UK’s deficit in this area. Moreover, it comes at a time when technology has revolutionised the potential of quantitative data. The UK’s remarkable data infrastructure is literally a couple of mouse clicks away, while software such NESSTAR allows novice users to explore the descriptive statistics of a dataset online without any need to download the data or read it into an analysis package.
The ESRC has developed a Quantitative Methods (QM) initiative to tackle this deficit, with programmes aimed at improving the reach and quality of QM training at undergraduate and postgraduate levels. As reported in the October RSS NEWS, the latest stage in this initiative was the launch with the Nuffield Foundation and the Higher Education Funding Council for England of a £15.6m programme to established centres of excellence in undergraduate QM teaching. A key feature of this programme will be the recruitment of up to 50 lecturers across up to 15 university centres. This programme will draw upon a suite of 20 projects in UK universities, developing curriculum innovation in QM teaching and sharing knowledge amongst teachers of best international practice, which started earlier this year. A new website will shortly be launched, quantitativemethods.ac.uk, which will act as a portal to the growing volume of high quality QM training material available on the web.
The British Academy has established a High Level Strategy Group to bring key stakeholders together to ensure that UK higher education is able to produce both a larger cohort of social science graduates with ‘STEM skills’ – a robust background in QM and their potential application – and graduates in all disciplines with a surer grounding in statistical literacy. The RSS getstats campaign is represented on the group by David Walker. The Academy has also released a position statement, ‘Society Counts’, calling on universities to ensure that all undergraduates in humanities and social sciences are given proper training in quantitative skills (available at www.britac.ac.uk/policy/Society_Counts.cfm). Significantly, every major social science learned society supported the statement.
Reversing the trend for decline in QM in UK social science will not be easy, but it is an urgent priority, not only to make good use of the UK’s data infrastructure, but also to rise to the rapidly growing challenge of harnessing the potential of ‘big data’ and digital society. Virtually every human activity now leaves some kind of digital trace, whether that is through social media, administrative data records, transactional data stored by private enterprises or geospatial data recorded by phones, tablets or satellites. This data not only has vast scientific potential; its use also raises profound ethical questions about privacy and disclosure. New techniques will be required to exploit its potential. For the last half century, inference from samples to populations has been at the heart of social statistics. In the next 50 years data mining will come to the fore.
Half a century ago, probably a majority of social scientists in the UK would have had sufficient competence in social statistics to understand the theory and practice of surveys or inference from samples to populations. Often the private sector would have turned to them for advice. Today, many social scientists might understand the potential of ‘big data’, but only a tiny fraction have the technical skills to exploit it. It is almost exclusively in the private sector that such skills are found. Raising the standard of social science QM training will be a vital first step in addressing this serious weakness. It will also reinforce the efforts of getstats to improve graduate statistical literacy, and help UK higher education to supply the growing number of data scientists that the digital economy will demand.
* John MacInnes is professor of sociology, University of Edinburgh, ESRC strategic advisor on quantitative methods training and a member of the getstats campaign board and British Academy High Level Strategy Group on quantitative skills.