Making Inferences About Elections and Public Opinion Using Incidentally Collected Data

17 Pages Posted: 28 Jun 2016 Last revised: 1 Nov 2016

Date Written: June 23, 2016

Abstract

Elections and public opinion scholarship has traditionally relied on datasets designed and collected for the specific purpose of measuring individuals’ political attitudes and behaviors. The Internet has greatly increased the availability of new non-purposefully or incidentally collected data (ICD), i.e. data that arises from citizen daily interactions on social media platforms, with search engines or civic technology platforms. ICD is now increasingly used in studies of public opinion, as well as to forecast elections and test causal relationships. While this work has shown that ICD can provide a valid basis for inference, it is clear that researchers need to treat such data with considerable caution in order to avoid making false claims. In this chapter we review the range and robustness of current research conducted with ICD and describe the main analytic challenges that researchers face in using ICD. Finally we show how, when properly handled, ICD can yield important new insights into political phenomenon.

Keywords: internet data, big data, public opinion, google trends, twitter, censorship, social media, election forecasting

Suggested Citation

Mellon, Jonathan, Making Inferences About Elections and Public Opinion Using Incidentally Collected Data (June 23, 2016). Available at SSRN: https://ssrn.com/abstract=2799658 or http://dx.doi.org/10.2139/ssrn.2799658

Jonathan Mellon (Contact Author)

University of Manchester ( email )

Oxford Road
Manchester, M13 9PL
United Kingdom

Register to save articles to
your library

Register

Paper statistics

Downloads
66
rank
321,189
Abstract Views
476
PlumX Metrics