In this podcast, Professor Patrick Dunleavy talks about how big data will affect the future of the social sciences. Say goodbye to academic siloes as we enter into a new age of cross/multi/and inter-disciplinary research. In this changing landscape, the old boundaries between physical, social and data science disintegrate. Here Professor Dunleavy talks about the Social Science of Human-Dominated and Human-Influenced Systems given as part of the Annual Lecture series at the Academy of Social Sciences.
Presented by Sierra Williams. Produced by Ewen MacArthur and Cheryl Brumley. Music and sound came courtesy of Podington Bear (By Grace). Image credits for Featured Images: Plugged by Keoni Cabral (Flickr, CC BY); MRI by Daniel Schwen (Wikimedia, CC BY-SA); Stellar classification by Alexander Meleg (Wikimedia, CC BY-SA 3.0). Podcast logo: headphones: gcg2009 (Sennheiser) via Flickr.
More on the Philosophy of Data Science
Mentioned in the podcast, we have been running a series of interviews conducted by Mark Carrigan on the nature of ‘big data’ and the opportunities and challenges presented for scholarship with its growing influence.
Rob Kitchin: “Big data should complement small data, not replace them.”
In this first interview, Rob Kitchin elaborates on the specific characteristics of big data, the hype and hubris surrounding its advent, and the distinction between data-driven science and empiricism.
Evelyn Ruppert: “Social consequences of Big Data are not being attended to”
For the second interview, Evelyn Ruppert discusses creating an interdisciplinary forum to analyse the major changes in our relations to data, as subjects, citizens and researchers.
Deborah Lupton: Liquid metaphors for Big Data seek to familiarise technology
Deborah Lupton talks about how sociologists are involved in making sense of and positioning big data. Also of interest to social researchers are the nature metaphors used to discuss data, such as ‘flows’ and ‘flood’
Susan Halford: “Semantic web innovations are likely to have implications for us all”
Co-director of the Web Science Institute, Susan Halford underlines the necessity of broad interdisciplinarity as well as further technical training to engage in depth with the web as it evolves in different ways.
Video and Report from the Academy of Social Sciences Annual Lecture
Professor Dunleavy’s overview of potential directions for the future of social science research prompted enthusiastic and lively discussion in questions. You can find the video with the slides from the presentation and the corresponding event report from the Academy of Social Sciences below.
In the coming decade the social sciences will become more coherent and unified in their methods and conclusions under the influence of big data analysis innovations, software developments and cross-disciplinary methods inflows, Professor Patrick Dunleavy told the Academy of Social Sciences annual lecture.
“A lot of new methods are coming in from the STEM sciences that are having a great impact on the social sciences,” he said. “I think we’ll move away from social science-specific work to much more use maths and physics-based quantitative work in the next 10 years. Software engineering and IT will become much more important, and stats software for small N datasets, and reliance reactive surveys, will decline.
“We are going to need a lot more systematic reviews [of evidence]. This approach is beginning to spread in from health studies to social policy more generally, and it will have very good effects. We are going to have great new ways of handling text, qualitative evidence and organisational analysis.
“That’s a very big of concatenation of positive developments all happening at the same time and all tending to push towards a pooling of methods and evidence criteria with the STEM disciplines.”
Speaking to an audience of 70 Academy Fellows in London, Professor Dunleavy FAcSS, Professor of Political Science and Public Policy at the LSE, said that the next decade would see three big changes in the social sciences.
“Firstly we will move towards moderate consensus and rapid empirical advances in the social sciences disciplines. I think we have already moved to that extensively. We will have fewer methodology wars –fewer debates where everybody says ‘This whole stream of analysis is utterly without value and should be repudiated’ – the sort of ‘pulling up the roots that kills the plant’ debate.
“I think we are going to enter into much more applied work, accomplished much faster, work that is much more timely and more cumulative. As Randall Collins argued for the physical sciences, prestige will lie more with being at the empirical forefront of ‘discovery’ research, albeit in a more collective way in social science.
“I think academic blogging and related social media (like Twitter) are already playing an absolutely critical role in that.”
Although he didn’t agreed that the days of social theorists were over, “I think there will be a push away from prestige and kudos inside the social sciences sitting with big thinkers and big picture-drawers, towards people doing more applied empirical work”.
“Ten years from now the social sciences will be much more internally unified. It is staggering the extent to which the social sciences at the moment are still run by different single disciplines in terrifically siloed and short-sighted ways, isolated from one another”.
“If we are going to connect more effectively with the STEM sciences – which I’m very hopeful about – it’s got to be at a discipline-group level. It’s got to be a broad-front advance, which I believe is very viable and plausible.”
He gave an example of how STEM subjects could influence social science approaches fundamentally. Actually the methods standards of economics are not all that great – there’s a lot of ‘We have only got a small N dataset (e.g. in cross-national research), but let’s compensate by funking it up with a lots of hi-tech maths’. But this approach is no longer at all convincing, by emerging social science big data standards.
“Big data has changed the whole picture, not just for economics but for large areas of political science and cognate disciplines. People who were at the forefront of their discipline five years ago are not necessarily at the forefront now. They are often a protesting rearguard, hanging on to methods from 20 years ago.”
In Professor Dunleavy’s view the most interesting and innovative research was necessarily being carried out on disciplinary borders, whether in cross-disciplinary, multi-disciplinary or inter-disciplinary ways. “Single discipline work is still a foundation, and a very important foundation. But it is not the be all and end all of academic research any more. The world especially outside universities has moved on from single discipline work.” Social scientists were being brought into research teams by IT companies and government, where a huge amount of work was being carried out in what many commentators now call trans-disciplinary teams.
Professor Dunleavy sits on the Board of the Campaign for Social Science, which was set up by the Academy.
This article is cross-posted on the Academy of Social Sciences website and is reposted with the author’s permission.
Note: This article gives the views of the authors, and not the position of the Impact of Social Science blog, nor of the London School of Economics. Please review our Comments Policy if you have any concerns on posting a comment below.
Patrick Dunleavy is Co-Director of Democratic Audit, Chair of the LSE Public Policy Group, and a Professor of Political Science at the LSE.