How exposure to different opinions impacts the life cycle of social media

Abstract

As a lot of communication and media consumption moves online, people may be exposed to a wider population and more diverse opinions. However, individuals may act differently when faced with opinions far removed from their own. Moreover, changes in the frequency of visits, posting, and other forms of expression could lead to narrowing of the opinions that each person observes, as well as changes in the customer base for online platforms. Despite increasing research on the rise and fall of online social media outlets, user activity in response to exposure to others’ opinions has received little attention. In this study, we first introduce a method that maps opinions of individuals and their generated content on a multi-dimensional space by factorizing an individual–object interaction (e.g., user-news rating) matrix. Using data on 6151 users interacting with 287,327 pieces of content over 21 months on a social media platform we estimate changes in individuals’ activities in response to interaction with content expressing a variety of opinions. We find that individuals increase their online activities when interacting with content close to their own opinions, and interacting with extreme opinions may decrease their activities. Finally, developing an agent-based simulation model, we study the effect of the estimated mechanisms on the future success of a simulated platform.

This is a preview of subscription content, log in to check access.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7

Notes

  1. 1.

    Mapping the opinions in lower (than influential factors) dimensional spaces forces the factorization process to mix the effect of multiple factors on a single dimension. Yet, the computational expense of the optimization increases with higher dimensions. Thus, setting the number of dimensions is a tradeoff between capturing the (most important) influential factors in separate dimensions and keeping the optimization process feasible.

  2. 2.

    The number of ‘between-weeks’ was optimized using the F-test on the first regression model (on posting), and we chose the number of in-between weeks based on the p-value. Note that since the number of variables (user and time-fixed effects) and the number of datasets change with different ‘between weeks’ values, we cannot use \(R^{2}\), Akaike information criterion (AIC), or log-likelihood value for this comparison. \(\textit{sigmoid}\,(K=0)=0.5\) implies that the story is not biased toward any of the groups (supporter/opposition and yellow/green), and the story’s attractiveness is zero: The story is neither attractive nor unattractive in obtaining votes.

  3. 3.

    \(\textit{sigmoid}~ (K=0)=0.5\) implies that the story is not biased toward any of the groups (supporter/opposition and yellow/green), and the story’s attractiveness is zero: The story is neither attractive nor unattractive in obtaining votes.

  4. 4.

    We assumed that users will not publish more than 100 stories per day, post more than 500 comments, or visit the website more than 5 times. We conducted a sensitivity analysis and the overall dynamic behavior of the results is not sensitive to these assumed values. These rates cannot be negative, either.

  5. 5.

    Due to the short life cycle of stories (i.e., one day) and the range of online rate of users in our case study platform (i.e., less than one time per day), releasing the memory does not have any significant effect; however, in other platforms this assumption may be violated.

  6. 6.

    Hence, for all the opinion groups, gaining utility by reading stories with the same point of view increases the posting rate.

  7. 7.

    Note that a greater difference between the opinion of the user and that of the stories she reads results in a further decrease in her utility and posting rates. For instance, when a User-II reads a Story-III, her posting rate decreases more than when she reads a Story-III which is more neutral toward opposing the government.

  8. 8.

    In other words, reading green stories (Story-II or Story-IV) that are somehow more neutral (i.e., not extreme in opposing/supporting the government) increases the posting rate of green users (User-II or User-IV). However, reading green extreme opponent stories (User-IV reading Story-II or User-II reading Story-IV) reduces user posting rates, implying that they become discouraged about sharing their opinions when they read stories with opinions differing widely from their own with regard to the government (i.e., when they feel the opinions of their audiences are very different from theirs).

  9. 9.

    Unlike in the previous cases, here Users-III confront opinions of Users-I who are extreme in supporting the government by posting more stories.

  10. 10.

    As noted earlier, we had raw data for comments and posts but estimated the time of online visits—see Ashouri Rad (2016) for more information.

  11. 11.

    Basically, based on the structure of our simulated social media website (in which stories have a one-day life cycle and are sorted based on their time of publication), for consistent results we need to have enough new stories for each user to read in each online session. The \(R_{Read}\) and \(D_\textit{Online}\) of users determine the number of new stories we need in the system for consistent simulation results and, considering parameter values (Table 3), 500 users generate enough new stories for all users to read. Communities may fizzle out due to lack of content for significantly smaller user bases.

  12. 12.

    GWI Social report 2015: A typical internet user on average spend 1.77 hours per day on social networks, while younger generations spend more than that (2.68 h for 16–24 years old and 2.16 for 25–34 years old).

References

  1. An, J., Quercia, D., Cha, M., Gummadi, K., & Crowcroft, J. (2014). Sharing political news: The balancing act of intimacy and socialization in selective exposure. EPJ Data Science, 3(1). doi:10.1140/epjds/s13688-014-0012-2.

  2. Ashouri Rad, A. (2016). Three essays on dynamics of online communities (Doctor of Philosophy), Virginia Tech.

  3. Ashoury, A., Herrera-Restrepo, O., & Zhang, H. (2013). The diffusion of a petition in a social network. In Paper presented at the 63rd annual conference and expo of the Institute of Industrial Engineers.

  4. Bakshy, E., Rosenn, I., Marlow, C., & Adamic, L. (2012). The role of social networks in information diffusion. In Paper presented at the proceedings of the 21st international conference on world wide web.

  5. Boyd, D. (2007). Why youth (heart) social network sites: The role of networked publics in teenage social life. MacArthur foundation series on digital learning–Youth, identity, and digital media volume, 119–142.

  6. Burke, M., Marlow, C., & Lento, T. (2010). Social network activity and social well-being. In Paper presented at the Proceedings of the SIGCHI conference on human factors in computing systems.

  7. Cannarella, J., & Spechler, J. A. (2014). Epidemiological modeling of online social network dynamics. arXiv preprint arXiv:1401.4208.

  8. Deters, Fg, & Mehl, M. R. (2013). Does posting Facebook status updates increase or decrease loneliness? An online social networking experiment. Social Psychological and Personality Science, 4(5), 579–586.

    Article  Google Scholar 

  9. Dwyer, C., Hiltz, S., & Passerini, K. (2007). Trust and privacy concern within social networking sites: A comparison of Facebook and MySpace. AMCIS 2007 proceedings, 339.

  10. Ellison, N. B., Steinfield, C., & Lampe, C. (2007). The benefits of Facebook “friends:” Social capital and college students’ use of online social network sites. Journal of Computer-Mediated Communication, 12(4), 1143–1168.

    Article  Google Scholar 

  11. Garcia, D., Mavrodiev, P., & Schweitzer, F. (2013). Social resilience in online communities: The autopsy of friendster. In Paper presented at the proceedings of the first ACM conference on online social networks.

  12. Giliette, F. (2011). The rise and inglorious fall of myspace. http://cli.gs/94lb3yf.

  13. Gose, C. (2013). Internal communications: The rise and fall of internal social networks. Retrieved from http://www.rmgnetworks.com/blog/bid/215425/Internal-communications-the-rise-and-fall-of-internal-social-networks.

  14. Gottfried, J., Barthel, M., Shearer, E., & Mitchell, A. (2016). The 2016 presidential campaign—A news event that’s hard to miss. Pew Research Center.

  15. Greenwood, S., Perrin, A., & Duggan, M. (2016). Social media update 2016. Pew Research Center.

  16. Jalali, M. S. (2014). How individuals weigh their previous estimates to make a new estimate in the presence or absence of social influence. In Social computing, behavioral-cultural modeling and prediction (pp. 67–74). Springer International Publishing. doi:10.1007/978-3-319-05579-4_9.

  17. Jalali, M. S., Ashouri, A., Herrera-Restrepo, O., & Zhang, H. (2016). Information diffusion through social networks: The case of an online petition. Expert Systems with Applications, 44, 187–197.

    Article  Google Scholar 

  18. Kim, J.-G., & Lee, K.-S. (2014). Predicting political orientation of news articles based on user behavior analysis in social network. IEICE Transactions on Information and Systems, E97D(4), 685–693. doi:10.1587/transinf.E97.D.685.

    Article  Google Scholar 

  19. Lampe, C., Ellison, N. B., & Steinfield, C. (2008). Changes in use and perception of Facebook. In Paper presented at the proceedings of the 2008 ACM conference on Computer supported cooperative work.

  20. Lee, J. (2014). Are some people less influenced by others’ opinions? The role of internal political self-efficacy and need for cognition in impression formation on social networking sites. Cyberpsychology Behavior and Social Networking, 17(9), 571–577. doi:10.1089/cyber.2013.0713.

    Article  Google Scholar 

  21. Liu, D. C., & Nocedal, J. (1989). On the limited memory BFGS method for large scale optimization. Mathematical Programming, 45(1–3), 503–528.

    Article  Google Scholar 

  22. Madejski, M., Johnson, M., & Bellovin, S. M. (2011). The failure of online social network privacy settings. Department of Computer Science, Columbia University, Tech. Rep. CUCS-010-11.

  23. Morahan-Martin, J., & Schumacher, P. (2003). Loneliness and social uses of the internet. Computers in Human Behavior, 19(6), 659–671.

    Article  Google Scholar 

  24. Muchnik, L., Aral, S., & Taylor, S. J. (2013). Social influence bias: A randomized experiment. Science, 341(6146), 647–651. doi:10.1126/science.1240466.

    Article  Google Scholar 

  25. Rad, A. A., & Rahmandad, H. (2013). Reconstructing online behaviors by effort minimization. In Social computing, behavioral-cultural modeling and prediction (pp. 75–82). Springer. doi:10.1007/978-3-642-37210-0_9.

  26. Tanford, S., & Montgomery, R. (2015). The effects of social influence and cognitive dissonance on travel purchase decisions. Journal of Travel Research, 54(5), 596–610. doi:10.1177/0047287514528287.

    Article  Google Scholar 

  27. Torkjazi, M., Rejaie, R., & Willinger, W. (2009). Hot today, gone tomorrow: On the migration of MySpace users. In Paper presented at the proceedings of the 2nd ACM workshop on online social networks.

  28. Trusov, M., Bucklin, R. E., & Pauwels, K. (2009). Effects of word-of-mouth versus traditional marketing: Findings from an internet social networking site. Journal of Marketing, 73(5), 90–102.

    Article  Google Scholar 

  29. Tucker, C. E. (2014). Social networks, personalized advertising, and privacy controls. Journal of Marketing Research, 51(5), 546–562.

    Article  Google Scholar 

  30. Viégas, F. B. (2005). Bloggers’ expectations of privacy and accountability: An initial survey. Journal of Computer-Mediated Communication, 10(3), 00–00.

    Article  Google Scholar 

  31. Wadee, Z. (2013). Facebook your boss: Using social media in internal communications. Retrieved from http://www.theguardian.com/careers/careers-blog/facebook-employers-encourage-social-media.

  32. Wu, S., Das Sarma, A., Fabrikant, A., Lattanzi, S., & Tomkins, A. (2013). Arrival and departure dynamics in social networks. In Paper presented at the proceedings of the sixth ACM international conference on web search and data mining.

Download references

Acknowledgements

This study is based in part upon work supported by the National Science Foundation under Grant Numbers SES-1027413. Any opinions, findings, and conclusions, or recommendations expressed in this study are those of the authors and do not necessarily reflect the views of the National Science Foundation.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Armin A. Rad.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary material 1 (docx 557 KB)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Rad, A.A., Jalali, M.S. & Rahmandad, H. How exposure to different opinions impacts the life cycle of social media. Ann Oper Res 268, 63–91 (2018). https://doi.org/10.1007/s10479-017-2554-8

Download citation

Keywords

  • Social media
  • User activity
  • Opinion measuring
  • Agent-based simulation