TY - GEN
T1 - Analyzing polarization in social media
T2 - 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL HLT 2019
AU - Demszky, Dorottya
AU - Garg, Nikhil
AU - Voigt, Rob
AU - Zou, James
AU - Gentzkow, Matthew
AU - Shapiro, Jesse
AU - Jurafsky, Dan
N1 - Funding Information:
Acknowledgements. We thank Jure Leskovec and Adrijan Bradaschia for data, and Cleo Con-doravdi, Chris Potts, Linda Ouyang, David Ritz-woller and Frank Yang for helpful feedback. We are grateful for the support of the Stanford Cy-ber Initiative, the Melvin and Joan Lane Stanford Graduate Fellowship (to D.D.), NSF GRF DGE-114747 (to N.G.), the Michelle and Kevin Douglas Stanford Interdisciplinary Graduate Fellowship (to R.V.), NSF grant CRII 1657155 (to J.Z.), the Stanford Institute for Economic Policy Research and the Knight Foundation (to M.G.) and the Brown University Population Studies and Training Center (to J.S.).
Funding Information:
We thank Jure Leskovec and Adrijan Bradaschia for data, and Cleo Condoravdi, Chris Potts, Linda Ouyang, David Ritzwoller and Frank Yang for helpful feedback. We are grateful for the support of the Stanford Cyber Initiative, the Melvin and Joan Lane Stanford Graduate Fellowship (to D.D.), NSF GRF DGE-114747 (to N.G.), the Michelle and Kevin Douglas Stanford Interdisciplinary Graduate Fellowship (to R.V.), NSF grant CRII 1657155 (to J.Z.), the Stanford Institute for Economic Policy Research and the Knight Foundation (to M.G.) and the Brown University Population Studies and Training Center (to J.S.).
Publisher Copyright:
© 2019 Association for Computational Linguistics
PY - 2019/1/1
Y1 - 2019/1/1
N2 - We provide an NLP framework to uncover four linguistic dimensions of political polarization in social media: topic choice, framing, affect and illocutionary force. We quantify these aspects with existing lexical methods, and propose clustering of tweet embeddings as a means to identify salient topics for analysis across events; human evaluations show that our approach generates more cohesive topics than traditional LDA-based models. We apply our methods to study 4.4M tweets on 21 mass shootings. We provide evidence that the discussion of these events is highly polarized politically and that this polarization is primarily driven by partisan differences in framing rather than topic choice. We identify framing devices, such as grounding and the contrasting use of the terms “terrorist” and “crazy”, that contribute to polarization. Results pertaining to topic choice, affect and illocutionary force suggest that Republicans focus more on the shooter and event-specific facts (news) while Democrats focus more on the victims and call for policy changes. Our work contributes to a deeper understanding of the way group divisions manifest in language and to computational methods for studying them.
AB - We provide an NLP framework to uncover four linguistic dimensions of political polarization in social media: topic choice, framing, affect and illocutionary force. We quantify these aspects with existing lexical methods, and propose clustering of tweet embeddings as a means to identify salient topics for analysis across events; human evaluations show that our approach generates more cohesive topics than traditional LDA-based models. We apply our methods to study 4.4M tweets on 21 mass shootings. We provide evidence that the discussion of these events is highly polarized politically and that this polarization is primarily driven by partisan differences in framing rather than topic choice. We identify framing devices, such as grounding and the contrasting use of the terms “terrorist” and “crazy”, that contribute to polarization. Results pertaining to topic choice, affect and illocutionary force suggest that Republicans focus more on the shooter and event-specific facts (news) while Democrats focus more on the victims and call for policy changes. Our work contributes to a deeper understanding of the way group divisions manifest in language and to computational methods for studying them.
UR - http://www.scopus.com/inward/record.url?scp=85084290619&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85084290619&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:85084290619
T3 - NAACL HLT 2019 - 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies - Proceedings of the Conference
SP - 2970
EP - 3005
BT - Long and Short Papers
PB - Association for Computational Linguistics (ACL)
Y2 - 2 June 2019 through 7 June 2019
ER -
