Beyond Binary Labels: Political Ideology Prediction of Twitter Users

Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics, pages 729–740 Vancouver, Canada, July 30 - August 4, 2017. c⃝2017 Association for Computational Linguistics https://doi.org/10.18653/v1/P17-1068 Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics, pages 729–740 Vancouver, Canada, July 30 - August 4, 2017. c⃝2017 Association for Computational Linguistics https://doi.org/10.18653/v1/P17-1068 Beyond Binary Labels: Political Ideology Prediction of Twitter Users Daniel Preot¸iuc-Pietro Positive Psychology Center University of Pennsylvania danielpr@sas.upenn.edu Ye Liu∗ School of Computing National University of Singapore liuye@comp.nus.edu.sg Daniel J. Hopkins Political Science Department University of Pennsylvania danhop@sas.upenn.edu Lyle Ungar Computing & Information Science University of Pennsylvania ungar@cis.upenn.edu Abstract Automatic political preference prediction from social media posts has to date proven successful only in distinguishing between publicly declared liberals and conserva- tives in the US. This study examines users’ political ideology using a seven- point scale which enables us to identify politically moderate and neutral users – groups which are of particular interest to political scientists and pollsters. Using a novel data set with political ideology labels self-reported through surveys, our goal is two-fold: a) to characterize the po- litical groups of users through language use on Twitter; b) to build a fine-grained model that predicts political ideology of unseen users. Our results identify differ- ences in both political leaning and engage- ment and the extent to which each group tweets using political keywords. Finally, we demonstrate how to improve ideology prediction accuracy by exploiting the rela- tionships between the user groups. 1 Introduction Social media is used by people to share their opin- ions and views. Unsurprisingly, an important part of the population shares opinions and news related to politics or causes they support, thus offering strong cues about their political preferences and ideologies. In addition, political membership is also predictable purely from one’s interests or de- mographics — it is much more likely for a reli- gious person to be conservative or for a younger person to lean liberal (Ellis and Stimson, 2012). ∗ Work carried out during a research visit at the Univer- sity of Pennsylvania User trait prediction from text is based on the as- sumption that language use reflects a user’s de- mographics, psychological states or preferences. Applications include prediction of age (Rao et al., 2010; Flekova et al., 2016b), gender (Burger et al., 2011; Sap et al., 2014), personality (Schwartz et al., 2013; Preot¸iuc-Pietro et al., 2016), socio- economic status (Preot¸iuc-Pietro et al., 2015a,b; Liu et al., 2016c), popularity (Lampos et al., 2014) or location (Cheng et al., 2010). Research on predicting political orientation has focused on methodological improvements (Pen- nacchiotti and Popescu, 2011) and used data sets with publicly stated dichotomous political orien- tation labels due to their easy accessibility (Syl- wester and Purver, 2015). However, these data sets are not representative samples of the entire population (Cohen and Ruths, 2013) and do not accurately reflect the variety of political attitudes and engagement (Kam et al., 2007). For example, we expect users who state their political affiliation …

Content truncated to 500 words