Examining the Effect of IP Location Display on Cross-cutting and Like-minded Discussions

To combat misinformation, the Chinese government has implemented a new policy for social media platforms, allowing all users to identify the province or country where the poster is located through IP locations (Reuters, 2022). IP location serves as a form of social cue, signifying users’ geographical affiliations. Such social cues are crucial for online discussions. Positive cues, such as expert status and positive emotions, can promote online discussions (Himelboim, 2008). Prior research has predominantly focused on social cues that stem from individuals’ voluntary selfdisclosure. People tend to self-disclose on social media platforms to craft their personas and foster social interactions (Bazarova & Choi, 2014), leading to a situation where most online social cues result from meticulous curation and selection. However, the display of IP location is not optional but compulsory. We speculate that this mandatorily displayed social cue will have a distinct impact on online discussions, especially cross-cutting discussions.

We designed a randomized controlled experiment to explore our research question. We recruited students (n=119) from the School of Film and TV as participants at a college in Guangdong Province, China. Participants were randomly assigned to three experiment groups. They were asked to read a post on social media about a controversial issue. According to our filtering question, all participants did not agree with the post's content. In other words, the experiment materials presented cross-cutting content for all participants. The content read by the three groups was the same, except for the IP location of the poster. The content read by one group did not show the IP location, while the content read by the other two groups showed the IP location within the province and the IP location outside the province, respectively. After reading the material, they were asked to complete a questionnaire to measure their intention to discuss the given topic with the poster of cross-cutting discussion (i.e., cross-cutting discussion intention) and to discuss the given topic with like-minded people (i.e., like-minded discussion intention). The ANOVA results showed that there were significant differences in cross-cutting discussion intention between different groups F (2,116) = 4.193, p < .05. The results of post hoc comparisons showed that compared to IP locations within the province (M = 2.08, SD = 1.02), cross-cutting discussion intention was significantly higher for no IP location group (M = 2.69, SD = 1.41) and remote IP location group (M = 2.90, SD = 1.27). In addition, although the overall result of like-minded discussion intention was not significant F (2,116) = 2.168, p = .119, the results of the post hoc comparisons showed that like-minded discussion intention from the remote IP location group (M = 4.14, SD = 1.41) was significantly higher than that of the no IP location group (M = 3.56, SD = 1.30). In summary, IP location display significantly affected participants’ discussion intention, and the remote IP group had higher crosscutting and like-minded discussion intention.

We speculate that this may be the result of group pressure. The geographical origins of individuals are significant elements of their social identity (Gao, 2015). IP location displays serve as a social cue indicating users’ geographical group affiliation. Close IP locations can be regarded as in-group, while remote IP locations are out-group. When cross-cutting content originates from in-group, individuals may hesitate to debate due to the fear of isolation. When cross-cutting content originates from out-groups, out-group pressure may provoke individuals to engage in argumentative defense of their beliefs. Therefore, our participants' high cross-cutting discussion intention may not necessarily signify a rational democratic discussion but a polarized argument defending their own beliefs.

The results of like-minded discussions corroborate this interpretation. Like-minded discussions are often associated with the polarization of opinions (Strandberg et al., 2019). When individuals are exposed to out-group content that contradicts their views, they are provoked not only to argue in defense of their beliefs but also to reinforce their opinions through discussions with like-minded individuals.

Our findings suggested that the display of IP location may intensify group pressure and exacerbate polarization in online cross-cutting discussions. This finding contrasts with previous research that suggested social cues could foster constructive discussions, indicating that only self-disclosed social cues might have such an effect. Our study offers insights for further exploration in the field of intergroup communication. As a new form of social cue indicating users’ geographical group affiliation, IP location may uniquely influence inter-group discussions. Future research could delve into its ramifications in various inter-group communication scenarios. Finally, our study fills the methodological gap of cross-cutting discussion research, which primarily relied on surveys to explore factors influencing cross-cutting discussions. On the contrary, the current study operationalized cross-cutting discussion as the willingness of individuals to engage in discussions with users expressing opposing views in experimental settings that better examine the causalities behind it. This operationalization also reflects the real-world scenarios where social media users participate in cross-cutting discussions, providing a valuable direction for future research.

Our study also has limitation. To set up a stable cross-cutting scenario for all participants and ensure the participants have similar views on the topic, we recruited homogeneous participants, which overlooked potential confounding variables, such as differing levels of cross-cutting views or demographic diversity. This limitation could be addressed in future research by incorporating repeated tests and comparisons across various populations. However, it is important to note that recruiting representative participants for a single experiment is challenging, as truly cross-cutting content (i.e., content that is universally disagreed upon) is unlikely to emerge in real-world social media discussions.