756.2
Online Randomized Experiment on Social Influences upon Behaviors in Web Forums
Online Randomized Experiment on Social Influences upon Behaviors in Web Forums
Monday, 16 July 2018: 17:45
Location: 706 (MTCC SOUTH BUILDING)
Oral Presentation
The (right-wing) political extremism in online news discussion forums has been reported and attracted much attention in Japan (Kimura 2017). The observation studies, however, have difficulty in clarifying the mechanism of the political extremism. There are two possible mechanisms: selection and social influence (cf. Lewis et al. 2012). Selection mechanism explanation is as follows: people have the determined opinion before entering the forum. Once the atmosphere of the forum is determined, those who have the similar opinion remain, and otherwise exit from the discussion, thus leading to radicalization of the existing opinion. On the other hand, social influence mechanism explains that the majority of opinion in the forum does make an influence on the participants and change their behavior in congruence with majority behaviors, thus leading to the extremism.
Here, we conducted an online randomized experiment in web discussion forums for clarifying the mechanism( For the relevant studies, see Alvarez and Winter 2017; Cheng et al. 2017). We used a crowdsourced recruitment for our experiment. Participants were instructed to read the article reporting the regulation of hate speech. Then, they were randomly assigned to two different experimental conditions to see how the majority of opinion affects participants’ behaviors. In the “positive comment” condition, the majority of the comments on the news was positive, supportive for the policy. In the “negative comment” condition, the majority of the comments on the news was negative, critical on the policy. Subsequently, the participants were asked to give up-votes or down-votes for each comment, write down their own comments and decide whether to display preferentially or non-display each commentator. We found the clear evidence that participants’ behaviors are influenced by the major opinions in the forum. This result suggests social influence mechanism does operate in the radicalization of opinion in online news discussion forums.
Here, we conducted an online randomized experiment in web discussion forums for clarifying the mechanism( For the relevant studies, see Alvarez and Winter 2017; Cheng et al. 2017). We used a crowdsourced recruitment for our experiment. Participants were instructed to read the article reporting the regulation of hate speech. Then, they were randomly assigned to two different experimental conditions to see how the majority of opinion affects participants’ behaviors. In the “positive comment” condition, the majority of the comments on the news was positive, supportive for the policy. In the “negative comment” condition, the majority of the comments on the news was negative, critical on the policy. Subsequently, the participants were asked to give up-votes or down-votes for each comment, write down their own comments and decide whether to display preferentially or non-display each commentator. We found the clear evidence that participants’ behaviors are influenced by the major opinions in the forum. This result suggests social influence mechanism does operate in the radicalization of opinion in online news discussion forums.