Big Data Meets Survey Science

5Measuring the Strength of Attitudes in Social Media Data

Ashley Amaya¹, Ruben Bach², Frauke Kreuter^2,3,4, and Florian Keusch^2,3

¹RTI International, Research Triangle Park, NC, USA

²Department of Sociology, School of Social Science, University of Mannheim, Mannheim, Germany

³Joint Program in Survey Methodology, University of Maryland, College Park, MD, USA

⁴Institute for Employment Research, Nuremburg, Germany

5.1 Introduction

The amount of stored data is expected to double every two years for the next decade (Quartz 2015). The advances in social science research will be enormous even if only a fraction of these data can be used. Official statistics could be produced more quickly and less expensively than currently possible using surveys (Kitchin 2015). Big Data may also provide statistics for small areas to make informed policy and program decisions (Marchetti et al. 2015). Researchers have already begun to harness Big Data from social media for a variety of purposes, including research into disease prevalence, consumer confidence, and household characteristics (Butler 2013; O'Connor et al. 2010; Citro 2014).

In researchers' excitement to capitalize on Big Data, many have published papers in which they attempt to create point estimates from Big Data. Unfortunately, there are as many examples of biased estimates created from Big Data as there are success stories. For example, Twitter data has successfully been used to predict the 2010 UK, 2011 Italian, and 2012 French elections ...