02/04: Yinxian Zhang

[smiklin]

Comment below with questions or thoughts about the reading for this week's workshop.

Please make your comments by Wednesday 11:59 PM, and upvote at least five of your peers' comments on Thursday prior to the workshop. You need to use 'thumbs-up' for your reactions to count towards 'top comments,' but you can use other emojis on top of the thumbs up.

[kthomas14]

Thank you for sharing your research with us! My question is, do you have any plans to apply these research methods to other areas and explore how social media influences people's view of their own government? If so, what might you expect to see?

[william-wei-zhu]

Thank you very much for your paper. I wonder among the users on Weibo who are disillusioned by democracy, how many percentage of them hold a rational calculative attitude (i.e. current political system is better than democracy because the current system brings more economic and social benefits), and how many percentage of them have internalized the commitment to the party (i.e. democracy is bad because it harms the interest of the party. I am committed to the prosperity of the party)? Thank you.

[luyingjiang]

Thank you for sharing. My question is how you test the model on different platforms?

[xxicheng]

Hi professor Zhang,

This is quite interesting, especially as a student curious about the recent Chinese political environment. I have a question about the tf-idf method you used in this research. It seems the word frequencies are counted and compared without their context, or in other words, are done after tokenizing the text. How can we tell the hidden attitudes (positive, negative, or neutral) of posters when they mention these words? Or do you think it is necessary to think about their attitude?

Another question about social media platform researches is: how do you deal with the possible censorship on Weibo influence your study? Thank you.

[WMhYang]

Thank you very much for sharing the paper with us. It investigates an interesting and controversial topic. In my opinion the trend coincides a little bit with the construction of confidence of Chinese people in its own development strategies, theories, and cultures. Is it possible to distinguish the trend? Thanks.

[XinSu6]

Thank you for this fascinating paper draft. Like many of my peers above, I am also wondering why can democracy-related opinions on social platforms like Weibo represent the trend for the whole people's attitudes toward democracy? Wouldn't the result be not representative enough?

Looking forward to your speech.

[ginxzheng]

Thank you for coming! Just wondering is there any comparison across different social media platforms, and would it be possible to incorporate these differences when we measure public opinions?

[mikepackard415]

Thank you for sharing your work, I'm really looking forward to Thursday. You close your paper with a notion of responsibility for democratic countries like "the U.S. or U.K," citing the information age and the influence these countries may have for democracy in other parts of the world. It seems there is a general trend toward authoritarianism worldwide, and I've heard theories that relate this trend to climate change and the uncertainty/fear it evokes. First I'd like to ask, what is your take on how these trends interact with climate change? And second, given recent events in the U.S. (culminating in the January 6th insurrection), how hopeful do you feel for democracy around the world in general?

[ddlee19]

Thank you for sharing your paper! I was curious about the reasons behind political commentary being more prevalent on text-based social media platforms (e.g. Twitter, Weibo) compared to image-based or video-based social media platforms were (Instagram, YouTube).

[97seshu]

Thank you for the presentation. My question is: How do you think your research findings can contribute to improving the relationships between China and other democratic countries?

[MegicLF]

Thanks for presenting the research! Could you discuss more about the bias and the validity of the data you use here?

[ghost]

How did you get interested in this topic?

[Qiuyu-Li]

Thank you for sharing with us. This is really an excellent research! I guess my only concern is just, how would you separate the effects of public opinion shifting with changes in censorship. Some of my classmates have mentioned opinion leaders on Weibo, and I think their posts are closely related with how the government is tightening the censorship.

[anqi-hu]

Thank you for sharing this work with us! In the presence of censorship on social media platforms, especially with regard to political opinions, how would you approach representing those with more polarized voices?

[YanjieZhou]

Thanks very much for your presentation! I really suspect the criterion of chooseing sample opinion leaders in this research, and whether the results have external validity.

[chiayunc]

Thank you so much for sharing this wonderful work. From the paper, I picked up something I find interesting in regards to the top words throughout the years. We see that the word "reform" has always been a front leader, and appearing as salient using LDA. However, from the TF_IDF count, we see that "constitutionalism" and "rule-of law" plunged quite significantly in recent years. I wonder, what might be some of the substantive arguments/ aspects for the "reform" people have been appealing? This probably comes from an out-of-context stream of thinking, and I apologize for it, but I am curious if the plunge of "constitutionalism" and "rule-of law" is the real depiction (ignoring the possibility that the number plunge because these words are communicated in other forms to dodge censorship), what type of reform are the people voicing? Since "constitutionalism" and "rule-of law" are not democracy itself, and they could possibly be ideal tools to substantiate arguments for a form of government that has check and balances, without resorting to democracy, I wonder why those concepts are not more popular with the public.

[chentian418]

Hi professor Zhang, thanks for presenting your fascinating paper, and I am interested in your study of opinion leaders.

On the one hand, a concern about the Weibo data is that opinion leader posts may still be censored or self-censored, and that in some cases their accounts may be entirely removed, which you also mentioned. As this would constitute some selection bias in selecting the opinion leaders, how do you overcome it or gauge this impact of state censorship on the trend of online discussions on democracy?

Secondly, you identified the opinion leaders of two types. The first type is identified through measuring their popularity and influence among users who are interested in political discussions on Weibo. While the other type is identified from users who may be popular but rarely contributed to public discussion of political and social issues. I am curious that, as the second type is selected using local knowledge when these users don't often publicly post politically related posts, how would you considered their posts to be useful for further data collection and processing, where you used a word-embedding model to expand the single keyword "democracy" to 17 terms that are semantically relevant to it?

Thanks!

[adarshmathew]

Thank you sharing your work, Dr. Zhang! Along with your takes on the questions of @wanxii, @jsoll1 , @Rui-echo-Pan @yierrr and @j2401, I have two questions:

1. How you would defend the stability/external validity of your interpretations when political rhetoric is constantly shifting (like @bjcliang-uchi mentions), especially in a centralized top-down manner as is the case with China?
2. Do you think your methodology would be able to capture the democratic decline of rhetoric and the turn to the 'illiberal, through shifts in meaning and usage? You could take the case of India and the US as examples of authoritarian creep.

[chun-hu]

Thank you for sharing! I'm also wondering if we can apply network analysis to this research?

[jinfei1125]

Thanks for sharing this interesting paper with us! One thing I felt could lead to the bias of this work, is that I think in China's online social media (or even in world-wide social media), only people with strong opinions speak loud. Do you have any idea of some possible biases and how to control them?

PS. Feel really happy to read a paper that studies China friendly ;)

[luckycindyyx]

Thank you so much for sharing! My question is that, how to distinguish the effect of censorship with self-censorship and "The Great Firewall of China"? Thank you!

[qishenfu1]

Thank you for sharing! Your research topic is very interesting and realistic. However, I wonder whether the data is biased -- social media in China is strictly regulated, and what many people say may not be what they really think. What do you think about this issue?

[SoyBison]

Hello Dr Zhang, I'm showing up a tad late to the discussion this week I'm afraid, so I'll just echo a lot of what I'm seeing from other students, especially as to the concern about the source of these trends. We know that propaganda and political pressure can change the way a country's popular opinion appears online, even absent of these sources, cultural waves and quirks of the technology can change these appearances, for example, surveys of opinion on Twitter/Reddit have vastly different results from NORC surveys, even on things that appear to be universal. On the other hand though, we also know that the Chinese Communist Party has employed pre-censoring on social media posts to prevent people from hosting public political gatherings.[^1] So what are your thoughts on these confounds?

[^1]: (King, G., Pan, J., & Roberts, M. E. (2014). Reverse-engineering censorship in China: Randomized experimentation and participant observation. Science, 345(6199), 1251722.)

[Raychanan]

My biggest concern is still about whether we can extend the scope to a larger audience/dataset.

[fyzh-git]

Thank you for sharing this interesting topic. I'm wondering whether the construction of the list of opinion leaders based on the '20 wide-acknowledged opinion leaders' would introduce implicit bias. As mentioned later, you also include people such as Han Han based 'your prior knowledge of Chinese opinion leaders', which would have not been included using the selecting method you developed.

[RuoyunTan]

Thanks for sharing your work. I am curious about how certain shocks affect public opinion in China. For example, in your opinion, how would the pandemic influence social attitudes towards democracy in China?

[goldengua]

Thank you for sharing this paper! I wonder how the text processing might affect the results. As people might use some word replacement strategies to avoid censorship, I suppose this is particularly challenging for the topic of your interest.

[FrederickZhengHe]

Thank you for sharing this interesting work. My question is: how did you identify the sentiment of Chinese netizens' posts? What were the criteria?

[mintaow]

Thanks for sharing your work with us! My question is more on the result interpretation side after semantic analysis and keywords summary. When you conclude "the very process of increasing exposure to international affairs has significantly influenced Chinese attitudes toward democracy and concurrently, toward the Chinese government" on page 23, I am curious about how to isolate this influence from the exposure to international affairs from the other interweaving effects?

[Panyw97]

Thank you for sharing! I'm wondering about the external validation of your model?

[hesongrun]

Thanks a lot for your wonderful presentation! How do you think this analysis can be extended to other disciplines of social science such as economics? Thanks!

[shenyc16]

Thank you for sharing this inspiring research with us. Focusing on Chinese people's perceptions of demorcracy instead of government performance is really meaningful. My question is on opinion leaders you select from Weibo or other online platforms. As far as I am concerned, while they express their sincere feelings and rational thoughts in most cases, there definitely exists some posts that are merely unbridled release of emotions or even deliberate solicitation for their private purposes. How will you distinguish and treat those posts? Thank you!

[zixu12]

Thanks for your sharing your work. I am interested in your topic on public opinions. After reading your relevant research, I am wondering what does the democracy usually look like in Chinese people's views? For some people, they regard democracy as more dynamic and they does not necessary think that they are living outside the democracy.

[heathercchen]

Thanks for presenting this interesting paper with us! In this paper, you utilize data from 2009 to 2017 to see Chinese people's attitude towards democracy. I am wondering do you have any plan to update this paper with the latest data, I mean data in 2021? Do you expect the pandemic will impose some shocks on people's perceptions? If there is such kind of effect, what direction might it be?

[luxin-tian]

Thank you very much for sharing. I believe that Chinese public opinion towards democracy is undergoing a dynamic shift as generations are influenced differently by education and the external environment. It must be interesting to extend the research in the most recent years.

[weijiexu-charlie]

Thanks for your presentation! I have similar concerns as other people above: how would you address the problem of fake accounts on Weibo?

[romanticmonkey]

Thank you for your presentation! I was wondering if Weibo participants only consists of relatively small portion of Chinese population. Would you say that ideologies on Weibo reflect that of the whole Chinese population?

[Qlei23]

Thank you for sharing this inspiring research! I have similar questions metioned above. First, what do you think of the effetcs of censorship? Second, I think the public opinoins on Weibo are somehow biased as its user group is a biased sample of the population (younger and more females). Thank you!

[cytwill]

Thanks for your presentations here. I am wondering how you recognize Chinese people's attitude on the issue of democracy from the text on social media. Also, considering the existence of censorship in Chinese social media, how would you verify the reliability of your findings?

[minminfly68]

Thanks for your presentation. This is a super interesting topics and I am very looking forward to hearing your methodology and apply it to my own research.

[ttsujikawa]

Thank you very much for sharing your interesting work. I also have a question regarding censorship like other students. Among social media platforms, I assume that there exists a certain difference in terms of the level of censorship. In this context, how would you expect this difference affect your conclusion?

[ziwnchen]

Thanks for the presentation, this topic is super interesting! As many have mentioned, one of the largest difficulties in using Weibo to study public opinion is dealing with the censorship issue. The censorship introduces a lot of bias in the data collection process. For example, survival bias in choosing the opinion leaders (It might be possible that after 2013 a lot of democracy-related opinion leaders' accounts are removed more often). You mentioned in your paper that "this methodological challenge serves the purpose of the paper, that is to gauge the impact of state censorship on the trend of online discussions on democracy." Could you elaborate more on that?

[Yiqing-Zh]

Thank you for your presentation! I am wondering whether the selection of the social media platform would impact the conclusions since the demographic distributions are very different for different social media platforms.

[timqzhang]

Thank you for your paper. I have the similar questions to my peers before, as the social media platform's true reflection to the opinion of people? How to measure it?

[yongfeilu]

Thank you very much for your presentation! I wonder how you quantify people's opinions on the social media platform. How do you think NLP can help with your research?

[chuqingzhao]

Thank you for sharing. I am also wondering whether there is any selection bias of the social media platform? Thanks

[TwoCentimetre]

I wonder if I want to conduct a similar research and do some comparison across different countries, could the methodology be transplanted? I do not see how we can use this method to compare different countries and I want to know what the baseline in this research. I mean we can obtain some data and generate some results, but how can we interpret them. Are they too high or too low? Second, it seems different countries have different form of democracy. Or do you believe there is a standard form of democracy? If not, how can we do such researches?

[AlexPrizzy]

Interesting work, do you think the perception of democracy in Chinese online discourse is affected by fear of the communist regime?

[Yaweili19]

I am very excited that our workshop finally has a researcher interested in the development of Chinese politics. Your empirical method is very clever. Looking forward to your speech!

[k-partha]

Thank you for your paper! Looking forward to your presentation.