Computational Social Science: The Future of Social Science?

Last week I attended The Eric Mindich Conference on Computational Social Science at Harvard, organized by David Lazer (Harvard) and Sandy Pentland (MIT), and co-sponsored at Harvard by the Institute for Quantitative Social Science and the Program on Networked Governance, and at MIT by the Legatum Center for Development and Entrepreneurship and the MIT Living The Future project. The conference explored the current and future developments in conducting social science research made possible by the increasing availability of digital pervasive data collected through new technologies such as the Internet, mobile telephony, RFIDs, CCTV cameras, administrative records and GPS information.

A series of brilliant presentations (agenda available here) explored the opportunities and the challenges facing social scientists working in this emerging field, which are raised by the collection of this new type of data. As survey analysis is becoming increasingly fraught with problems, due to the high costs of recruiting simple random samples and generally low response rates, social scientists are beginning to explore how digital data can be used as a new exploratory tool. The main opportunity provided by this approach clearly lays in the unprecedented access to large amounts of detailed data about respondents, which can be linked with their physical/geographical location. This of course raises also several challenges:

– the privacy issues involved in collecting pervasive user data and the legal implications for storing such databases, in terms of intellectual property
– the informatic challenges for social scientists who are engaging in such research
– the challenge of developing new statistical tools and methods of inferential analysis for analyzing the data collected

While at present the research is very much at the descriptive level of analysis, employing social network analysis tools and visualization techniques of social and geographical data, the potential for developing theory and hypothesis testing from such data is enormous (for example, the data lends itself particularly well to investigating how social phenomena from information to disease spread amongst a population). The challenges highlighted above should not be seen as barriers but rather as fostering some very interesting future developments in social science research which will shape the discipline in years to come. This in turn will push towards more inter-disciplinary research which will bring together social scientists, computer scientists and legal scholars – thus providing exciting new insights both on the social phenomena being studied and on the social implications of carrying out such research.

Author: Corinna

Corinna di Gennaro (BSc LSE; MPhil, DPhil, Oxon) is a sociologist and a professional in the printing industry.

Leave a Reply

Your email address will not be published. Required fields are marked *