Issue Salience and Issue Ownership Online and Offline: Comparing Twitter and Survey Data

Abstract

Social media, in particular Twitter, become an increasingly attractive tool for political scientists for the study of political attitudes and behavior because of its potential for tracking public opinion with minimal costs. Yet extracting reliable, valid and precise measurements of politically relevant concepts from these data sources still forms a major challenge. The present research is part of the project “Transforming Social Media Contents to Political Data” (funded by the Netherlands Organization for Scientific Research NWO), which aims at addressing this challenge by developing new tools and methods for harvesting, coding and analyzing messages from social media. Using data collected from Twitter during the last three weeks of the Dutch parliamentary election campaign 2012, we demonstrate in this paper how two prominent concepts of research on voting behavior – issue salience and issue ownership – can be measured online, and how these measures compare to traditional survey data. The preliminary results obtained by this study show that measures of issue salience and issue ownership using data collected from Twitter are comparable with the survey data to a considerable extent. However, we have also found some discrepancies between the two data sets on key issues, which suggest that our online measurements of the two concepts need to be further developed for a more sound validation.

Keywords: Twitter, Elections, Issue salience, Issue ownership, The Netherlands