著者：張 暁棟 （ZHANG, Xiaodong）
論文審査委員：ジョナサン・ルイス、福富 満久、赤嶺 淳、大坪 俊通
1. Outline of the Thesis
This thesis investigates the political use of social media in East Asia. It starts by reviewing the current situation of social media use in political campaigns across East Asia. It describes the common and different features of social media usage in East Asian states, shows the extent to which East Asian politicians use social media in their political campaigns, and compares the political use of social media in East Asia to that in the United States.
Second, this study investigates the relationship between candidates’ popularity on social media and electoral outcomes, using Taiwan as a case study. The author collected online and offline data and used them as predictors regarding the 2014 local election, the 2015 legislative by-election and the 2016 general election. His results show that candidates receiving more likes on Facebook did tend to get more votes, suggesting that long-term cultivation of support on the Internet can reap benefits at the ballot box. On the other hand, popularity on social media shows more predicting power in the higher-profile 2014 municipal election than in the 2015 by-elections which attracted less public interest. This suggests that further research should be carried out on the usefulness of social media data in predicting the results of elections attracting different levels of public interest. The success of the author’s models in predicting the election outcomes shows that, in the case of Taiwan at least, Facebook data can help us forecast the results of elections on a district-by-district level even in the absence of reliable opinion polls.
Third, this thesis investigates politicians’ posting strategies on Facebook during election periods, using the March 2018 Hong Kong legislative by-elections as a case study. Hong Kong has fair direct elections but fixed opposition (pro-democracy) and ruling (pro-establishment) camps, making it a good case for exploring the difference between ruling and opposition online campaign strategies. Based on a content analysis of Facebook posts, the author finds that the pro-democracy camp had significantly more posts reporting or announcing the live situation or campaign information, while the pro-establishment camp had relatively more posts expressing support from other politicians, elite bureaucrats or social celebrities. On the other hand, both camps engaged in similar levels of negative campaigning.
Finally, the thesis turns to voter behavior, and investigates the clustering of Taiwanese Facebook users with different political orientations in political campaigns. The author finds that the political communities on Facebook perform as echo chambers similarly to political blogs rather than displaying cross-ideological interactions such as those identified in previous research on Twitter mention networks. On the other hand, the author finds no significant differences in hyperlinks shared on the Facebook pages of candidates belonging to the two main parties, suggesting that social media use is not leading to selective exposure.
2. Evaluation of the Thesis
This thesis is extremely ambitious in its attempt to survey the political use of social media across the very different political and social national settings of East Asian countries. Furthermore, it addresses not only politicians’ use of social media but also that of voters.
This study departs from the bulk of research on the political use of social media in its use of Facebook rather than Twitter as the data source. This is valuable because there are many countries where Facebook is more widely used than Twitter, and because Facebook use is high even in countries where Twitter is popular. While many researchers use Twitter because of the availability of large amounts of public data through the Twitter API, this study demonstrates that it is possible to do systematic research on political communication using Facebook data.
In particular, the thesis uses Facebook data as one input for a new method for election forecasting. The author successfully develops models that combine online and offline data to predict the results of legislative elections. These models promise to be useful for forecasting the results of elections where traditional opinion polls are not carried out. The use of page view statistics for candidates’ Wikipedia pages is also a novel way of measuring levels of voter interest.
The selection of Hong Kong as a case study is unusual given the wider political setting within which the territory’s elections take place. It is interesting that, despite the lack of doubt about the overall outcomes of elections and the asymmetrical resources available to ruling and opposition camps, the thesis finds that both sides engage in similar amounts of negative campaigning on line. This offers a promising line for future research.
The thesis does have some limitations. Above all, the author could have discussed at greater length how far the results of the case studies can be generalized to other East Asian territories or beyond. For example, while the Hong Kong case study is valuable in itself, it is not made clear whether the findings could apply to other territories such as Japan where certain parties are in permanent or semi-permanent opposition.
Second, one of the attractions of the election prediction method developed here is that it can offer predictions in the absence of opinion polls. However, the author finds that his prediction models work better in high-profile elections where candidates’ Facebook pages will generate greater user activity. Given that opinion polls will most likely be carried out in such high-profile elections, this suggests that the models developed here are not about to make opinion polls redundant. On the other hand, further refinement of the models may make them especially useful in mid-profile elections where the costs of opinion polling are harder to justify.
On 17 April 2019 we examined Mr. ZHANG, Xiaodong regarding his PhD thesis “More Likes More Votes?: The Political use of Social Media in East Asia.” Mr. Zhang satisfactorily answered all our questions regarding his thesis.
We therefore conclude that Mr. Zhang has achieved the requisite level of academic achievement and ability to be awarded the degree of PhD in Social Sciences from this University in accordance with Article 5 Section 1 of the Hitotsubashi University Degree Regulations.