Project GoalsThis project seeks to understand how error messages are propagated over large-scale information networks and what content-related cues or network-related features make certain error messages more error-resistant or error-prone than others. The results of the project have the potential to help build a platform that accurately identifies errors being propagated on an information network and effectively manages such error propagation.
Current/Final Results (summary)The results indicate that fear appeals (threat and protection) and risk communication (risk and benefit) lead to a positive effect on unverified message’s propagation signifying error-proneness, while emotion analysis (anxiety and hope) is associated with a negative effect on unverified message’s propagation denoting error-resistance. In addition, health information (medical, strategies and actions) is related to error-proneness, while crisis communication (crisis, accidents and alerts) leads to error-resistance. Furthermore, messages from organizations as well as spatially distant from location of incidence resist the propagation of erroneous information while network-related characteristics (hashtag and followers) increase the propagation.
When reading a text, things can get ambiguous. Reading is different because the context cannot be felt, whereas while listening or viewing one can feel the context much better. We read loads of messages on social media, many of them, ambiguous. Our goal here is to find this ambiguity.Know more about the project.
Twitter is a social media channel, a microblogging service, that depends highly on text and less on media. Hence, we chose twitter for our research.
Ambiguity parser application
The application has been build using Java. We are using the Stanford NLP API to create parse trees from tweets. Each tree represents a meaning and has likelihood score.Use the application.
Data (with documentation)
We have used Mechanical Turk, a service provided by Amazon, to collect tweets.See our tweet collection template
Tweets we collected
Boston rumor data file
Zika rumor data file (with Health variables)
Zika rumor data file (with LIWC variables)
AcceptedDutta, H., Kwon, K. H., & Rao, H. R. (2018). A system for intergroup prejudice detection: The case of microblogging under terrorist attacks. Decision Support Systems, 113, 11-21.
Oh, O., Gupta, P., Agrawal, M., & Rao, H. R. (2018). ICT mediated rumor beliefs and resulting user actions during a community crisis. Government Information Quarterly, 35(2), 243-258.
Volety, T., Valecha, R., Vemprala, N., Kwon, H., Rao, H. R. (2018). Cyber-rumor Sharing: The Case of Zika Virus. Proceedings of American Conference on Information Systems (AMCIS) 2018. New Orleans, Louisiana.
Kwon, K. H., & Rao, H. R. (2017). Cyber-rumor sharing under a homeland security threat in the context of government Internet surveillance: The case of South-North Korea conflict. Government Information Quarterly, 34(2), 307-316.
Valecha, R., Volety, T., Vemprala, N., Kwon, H., Rao, H. R. (2017). An Investigation of Cyber-Rumor Sharing: The Case of Zika Virus. Workshop on Bright Internet and Global Trust Building (BIG) 2017. Seoul, South Korea.
Under Review/Under PreparationKrishnarao, S. S., Valecha, R., Agrawal, M., Rao, H. R. (Under Preparation). Emotional Analysis of Cyber-Rumor Sharing: A Coping Theory Approach. Workshop on Information Security and Privacy (WISP) 2019. Munich, Germany.
Tran, T., Valecha, R., Najafirad, P., Rao, H. R. (Under Preparation). Taxonomy of Misinformation Harms from Social Media in Humanitarian Crises. International Conference on Secure Knowledge Management (SKM) 2019. Goa, India.
Volety, T., Valecha, R., Kwon, H., Rao, H. R. (Under Review). The Effect of Threat and Proximity on Cyber-Rumor Sharing. Conference on Information Systems and Technology (CIST) 2019. Seattle, Washington.