Mitigating selective exposure in social media forums

Gao, Mingkun

Permalink

https://hdl.handle.net/2142/113856 Copy

Description

Title

Mitigating selective exposure in social media forums

Author(s)

Gao, Mingkun

Issue Date

2021-11-29

Director of Research (if dissertation) or Advisor (if thesis)

Sundaram, Hari

Doctoral Committee Chair(s)

Sundaram, Hari

Committee Member(s)

• Kirlik, Alex
• Karahalios, Karrie
• Xu, Anbang

Department of Study

Computer Science

Discipline

Computer Science

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Date of Ingest

2022-04-29T21:34:30Z

Keyword(s)

Computer science

Language

eng

Abstract

This dissertation focuses on designing social media interfaces to help people explore diverse social opinions and mitigate selective exposure - a tendency that people actively seek attitude-consistent information and avoid attitude-inconsistent information. Diverse information consumption has potential benefits, including but not limited to helping individuals form accurate viewpoints, facilitating better decision-making processes, cultivating people's tolerance and mutual understanding with others, which is essential for a thriving democratic society. Both actively seeking congenial information (i.e., selective exposure) and passively being in a congenial information environment (i.e., de facto selective exposure) can impair people's exposure to diverse social opinions. Meanwhile, people can play a significant role in shaping others' information environments by sharing information on social media. Thus, we break our general research problem down to two sub-problems: 1) designing interfaces to mitigate selective exposure for individual information consumption, which focuses on the effect of interface design on people's active information-seeking behavior; 2) understanding humans' role as the information filter for others, which is the first step towards a better interface design to tackle potential problems caused by information sharing among people. We first proposed organizing and showing categorized social opinions based on emotional reactions to mitigate selective exposure for individual information consumption. Our evaluation indicated that such a design could motivate people to explore diverse social opinions. Next, we designed and implemented a system that can provide novel visual hints with new recommendation mechanisms to improve people's awareness of diverse opinions and mitigate selective exposure. Finally, we studied the effect of the stance label and the credibility label on people's information selection and perception on a two-column news feed. We found that the stance label can exacerbate selective exposure and make people agree more on fake news. And the credibility label has a limited effect on mitigating selective exposure and combating fake news. Our work expanded the design toolbox of mitigating selective exposure and gave interface/system designers more choices when using these tools. To better understand people's role as the information filter, we conducted a simulated online experiment to figure out how the attitude distribution of the recipient group affects people's information-sharing behavior in the anonymous scenario. We observed that the attitude distribution of the recipient group has an impact on people’s sharing behavior even though various factors (e.g., topics, people's attitudes, etc.) may be related to such effect. People tend to cater to the majority's attitudes by selectively sharing more information consistent with the majority's attitudes in some specific context, which creates the filter bubble for others. This result indicated the necessity to study interface design to motivate people to share more balanced information to help break the filter bubble for those recipients.

Graduation Semester

2021-12

Type of Resource

Thesis

Permalink

http://hdl.handle.net/2142/113856

Copyright and License Information

Copyright 2021 Mingkun Gao

Owning Collections