Does site architecture matter? The political implications of public- versus private-oriented social network sites in China

ABSTRACT

Political communication researchers have shown that social network site (SNS) use and online network characteristics can impinge on people’s political attitudes and behavior. Nevertheless, individual SNSs have varying site architecture, basic design features, and functionalities, which may influence usage and online network characteristics. This study thus examines whether site architecture matters for understanding the political implications of SNSs. Specifically, this article conceptualizes the distinction between public-oriented and private-oriented SNSs. It contends that the two types of SNSs would have varying impact on political behavior through the kinds of online networks that they tend to sustain. Analysis of a survey of university students in Guangzhou, China (N = 897), shows that respondents using public-oriented SNSs tended to have a relatively more ‘public’ online network – larger, involving more strangers, and with more connections to public actors. Some of these network characteristics in turn affected political discussion behavior. The analysis thus demonstrates how SNS architecture can indirectly impinge on individual-level political outcomes.

KEYWORDS:

• Social network sites
• site architecture
• public-oriented versus private-oriented SNSs
• online network characteristics
• political discussion
• China

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Chris Chao Su is a Ph.D. student at the School of Journalism and Communication, the Chinese University of Hong Kong. His main research interests reside in mobile media, media use, and digital media consumption.

Francis L. F. Lee is Professor and Head of Graduate Division at the School of Journalism and Communication, the Chinese University of Hong Kong. His main research interests reside in political communication, journalism studies, public opinion analyses, and media and social movements.

Gongcheng Lin, Ph.D., is an assistant professor at the School of Communication and Design, Sun Yat-Sen University. His main research interests reside in communication theories, media economics, and media effects.

Notes

1. According to CNNIC’s 2014 report, 77.8% of Internet users have used QQ during the six months prior to the survey. The corresponding percentages for WeChat and Weibo are 65.0% and 43.6%, respectively.

2. Given the procedure, we are including respondents who live together in the same room. So one might question if these respondents are ‘independent’ from each other. Our judgment is that the issue of independence should not be a serious one in the present study: although the students live together, they are not family members having very close ties with each other. Nor is our survey concerned with topics that are directly related to dormitory environment. The data set does not contain information about which four respondents belong to the same room, so there is no basis for us to conduct the analysis using the robust standard error approach. This is one limitation of the current data set and analysis.

3. Concerns with space in the questionnaire forbade us to ask the respondents about their usage of the third or even fourth most frequently used SNSs. It should be noted that the respondents spent an average of 2.8 and 1.3 hours per day on the most and second most frequently used SNSs. The various SNS-related variables, therefore, have captured the characteristics of, on average, more than 4 hours of SNS use of the respondents. We believe that the variables created based on how the respondents used the two most frequently used SNSs should capture their overall SNS usage reasonably well.

4. The questions did not specify the exact type of contents they shared or received; hence the contents could be either people’s personal views or materials derived from other sources. But the distinction between one’s own views and derivative contents is not important for the present conceptual concern.

5. The low reliability coefficient suggests that news consumption behavior can be channel-specific. Removing specific items would not enhance the reliability coefficient though. The index is used to avoid the inclusion of too many individual items into the analysis.