05/19: Jennifer Pan

[shevajia]

Comment below with a well-developed question or comment about the reading for this week's workshop. These are individual questions and comments.

Please post your question by Wednesday 11:59 PM, and upvote at least three 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.

[Raychanan]

Hi Professor Pan, thanks so much for sharing your work! I have a few questions regarding the paper.

1. The limitation of the size of the sampled data. Despite the fact that the original datasets from Twitter and Weibo are extremely large (14 million and 6.7 million), the paper appears to sample only the ten most retweeted tweets for each week. For Weibo, the top hundred Weibo posts were retrieved with Word2Vec and USE given the viral tweet, and these form the basis for annotations used to determine whether Weibo posts are relevant. The fact that only the top ten tweets were used was quite surprising to me. Considering that the most popular posts may differ significantly from less popular posts, does the application of such a sampling strategy have a significant impact on your analysis and conclusions?

2. Can you elaborate a bit more on the "ordinary users" on Weibo? I find it interesting to see that ordinary users (or in your words, Weibo users without any affiliation to the media or government) play such an essential role in transferring information from Twitter to Weibo. My prior belief was that those Yellow Verified Users would be playing the major roles. In light of Sina Weibo's recent move to enforce the display of the IP addresses of its users, would it be possible to dig deeper into their characteristics? For instance, are most of them living in China's eastern regions that are much more prosperous? I think it would be interesting to examine whether people from more affluent areas are acting as intermediaries in the information transfer process between China and the rest of the world (80% users are outside the US as you point out).

3. Classification of the contents of the inflow. There has already been a substantial amount of work done by this paper, but I am still curious about the types of tweets smuggled into Weibo. As a worst case scenario, the imported posts are mostly racist criticism of Chinese citizens even if racism does not constitute the popular topic of Twitter, which might have caused major backlash and resentment within Chinese communities during COVID (as implies by your previous experiment on related issues).

[nswxin]

Professor Pan，
Thank you for your time in sharing the research! do agree with you that "governments all over the world impose restrictions on access to digital information". For example, on 5.05, Professor Lazer gave us an excellent sharing about how Twitter suppressed the spreading of fake news during the January 6 insurrection, and there's no clear evidence that government did not have a role in it.
In addition, I think the limited quantity of the inflowed information can also be explained by something other than government intervention. Your data collection happened during the most challenging time in China, and Chinese people simply didn't care that much about what the outsiders were commenting. The most relevant issues are staying at home during the Spring Festival and waiting for the victory of the battle. For example, in Data Index 9, the tweet criticized Wuhan people as "worst passengers, no manners, stubborn, uncivilized and dirty." Apparently, it's not true. And compared to the great battle to fight, including building hospitals and supporting Wuhan, Chinese people simply ignored the non-sense criticism. In addition, delivering such information back would harm the friendship between China and the entire world.

[Thiyaghessan]

Hi Professor Pan,

Thank you for taking the time to share your work with us. I was wondering if you had a little more information on the "ordinary users" who are circumventing barriers to information flows.

1. How does the state interact with such users? Once authorities become aware of their existence, do they move to ban them off the platform? It could be that the users you have identified are those that have escaped being banned or have been let off because the content they spread is not something that directly challenges the state.
2. Similarly for foreign entities, it seems like their ability to facilitate cross-border information flows is definitely one that the state is aware of. Are there then boundaries they cannot cross when posting? Is there evidence of any foreign entity posting anything critical of the Chinese government and getting away with it?

Overall, I am wondering about the effect of possible survivorship bias on the ordinary Weibo users you have identified. Thank you!

[sushanz]

Dear Professor Pan,

Thank you for your time today! It is a really interesting topic. Since you discussed about the different response windows in terms of the news/media types in the research, I wonder how you filtered out those side influences driven behind which may cause bias in the timeline window interpretations. For instance, it is possible that CGTN facilitated media responses to news faster than other topics simply because the posts was made in a special period/ festival such as the National Day of the People's Republic of China. They may want to avoid potential collective actions. Another question I have is about the participants as I did not see their information in appendix. You mentioned those bilingual researchers manually pair the posts/comments on Twitter and Weibo. I am curious if they have similar backgrounds such as education levels and cultural knowledge to distinguish those tweets or weibo posts precisely. Will this cause any possible bias in results?

[LynetteDang]

Thank you Professor Pan for sharing your work! I am wondering if you have looked at the expat community in China (especially in the big cities such as Shanghai, Beijing, Shenzhen and Guangzhou) and their role in facilitating the inflow of global information into China, versus domestic residents. The expat community emigrated to China after living in an environment with easy access to globally-used social media platforms, and thus could be more attached to vpn services. I suspect that they were the ones who use both global social media platforms (such as Twitter) and China-specific social media (such as Weibo) to circulate information.

[borlasekn]

Thank you Prof. for sharing such interesting and applicable work. In your paper, you mention the fact that social media sites restrict content, such as "fake news", but there is no guarantee that these are laissez-faire restrictions. Obviously, there is a lot of hidden information as companies become privatized. In your opinion, in cases such as Twitter where Elon Musk is purchasing Twitter, supposedly for freedom of speech and to ease these restrictions, can the company really be laissez-faire and unregulated, in regards to censorship? Are there any ways that you would imagine one could measure this? Would this idea even be possible under restrictions such as those in China? Thanks

[bowen-w-zheng]

Hi Professor Pan,
Thanks for sharing your work. I am interested in the measure you use to define information inflow given two spontaneously occuring signals. How do you tease apart in-flow information to China from mutual information due to common driver or flow of information in the opposite direction. I am reading something on transfer entropy, which define explicitly directional information flow [https://journals.aps.org/prl/pdf/10.1103/PhysRevLett.85.461]. I wonder if such a measure would be appropriate and benefical to this study. In particular, I am curious if the transfer entropy from other countries to China exihibits dramatic decrease while transfer entropy from state media to popular media plafroms increase after Chairman Xi seized power. Thank you!

[jsoll1]

Hi Professor Pan,

Your study is super interesting! I was wondering whether your wording of your results might be slightly misleading. You phrased it as information making its way into China 'despite' government censorship. I defer to your expertise, but none of the examples you cited making its way into China seemed to be information that the government actively attempted to censor. It seems like from the way you described things, the overall censorship apparatus makes it easier to stamp down on the stuff that they actually want to censor. Is that the case, or am I misunderstanding things? How does information that the Chinese government actually prioritize censoring make its way into China (or not!)?

[afchao]

Thank you or sharing this work with our group! We've seen a fair share of content on the topic of Chinese online censorship, perhaps due to the ease with which it lends itself to the computational social science toolkit. One previous presentation which comes to mind is from a group at UCLA who found, studying the outbreak of COVID-19 in China, that these crisis moments in fact increase popular access to media by means of promoting more risky censor-circumventing behavior. This seems to directly relate to the acknowledgement in your paper that studying specific Chinese media outlets may not offer a full picture of media consumption behavior behind the great firewall. This is a kind of meandering way of asking whether government censorship in China is more about political grandstanding than the actual management of information.

[hshi420]

I have 2 questions:

1. How did you deal with removed content on Twitter? As you mentioned, you used Tweet ID to retrieve the Tweets, but many content has been removed. They can be toxic content or misinformation, and I'm curious to see how misinformation on Tweet 'inflow' into China.
2. For Figure 2, there is a mismatch between the figure annotation and the text description in the article. In the article, it says 'Figure 2 shows how the total number of Weibo posts matched to tweets as a function of K'. The figure annotation says that the y variable is the number of matched tweets. I was wondering if it means that each Weibo can be mapped to multiple tweets, and thus the number of the total number of Weibo posts does not increase linearly with the increase of K.

Thank you!

[hhx2207061197]

Hi Professor,
Thanks for the sharing. I think the limited amount of incoming information can also be explained by reasons other than government intervention. In addition, I am also curious if the transfer entropy from other countries to China will drop sharply when President Xi takes office, while the transfer entropy from state media to mass media platforms will increase.

[Yutong0828]

Hi Professor Pan, thank you very much for sharing your work with us!
My question is about the research design. Do you think it is necessary to apply the similar method on a different topic (e.g., topics not related to politics) as comparison to the results in the current study? This may help us better estimate the Chinese government’s role in the process. Also, adding other countries that does not ban Twitter as comparison may also be helpful, because we don’t naturally know what should be a normal transmission pattern of information between global social media platforms and domestic media platforms without comparison groups.
Thanks!

[GabeNicholson]

Hi Professor, thank you for coming to speak with us.

Given the limited flow of information, have there been any studies done that test how informed China's citizens are about global news coverage compared to similar but more open societies? I believe this would be the next step for your research question and this paper could establish the causal link depending on the results. Further, how does your methodology scale to other instances such as Russia's current information environment?

[Jasmine97Huang]

It’s no secret that CCP exerts significant control over Chinese media (mass or social). Nonetheless, this paper that captures such process is thought-provoking. It would be interesting to compare the patterns of information inflow from Twitter to Weibo with that from Weibo to Twitter. That is, what are the differences in the ways COVID-related, event-driven messages diffuse transnationally to a highly moderated versus a loosely regulated platform? Additionally, like my fellow classmate mentioned, characterizing the group of Weibo users who facilitate the flow of information would be worth looking into. Lastly, I am curious about the content that didn’t reach the Weibo platform - are they trivial commentaries or significant events that could have been actively censored? Looking forward to your presentation!

[Jasmine97Huang]

Professor Pan， Thank you for your time in sharing the research! do agree with you that "governments all over the world impose restrictions on access to digital information". For example, on 5.05, Professor Lazer gave us an excellent sharing about how Twitter suppressed the spreading of fake news during the January 6 insurrection, and there's no clear evidence that government did not have a role in it. In addition, I think the limited quantity of the inflowed information can also be explained by something other than government intervention. Your data collection happened during the most challenging time in China, and Chinese people simply didn't care that much about what the outsiders were commenting. The most relevant issues are staying at home during the Spring Festival and waiting for the victory of the battle. For example, in Data Index 9, the tweet criticized Wuhan people as "worst passengers, no manners, stubborn, uncivilized and dirty." Apparently, it's not true. And compared to the great battle to fight, including building hospitals and supporting Wuhan, Chinese people simply ignored the non-sense criticism. In addition, delivering such information back would harm the friendship between China and the entire world.

I would count not "delivering such information back" due to fear of "harming friendship" as government intervention, though. Seems like a strategic move to achieve a political outcome. I do agree that Weibo doesn't have to pick up everything on Twitter. I am wondering if the language barrier would play a role in information flow as the majority of Weibo users do not understand English. The agents that facilitate the transmission are more salient in this case.

[yierrr]

Thank you for sharing the research! I am also interested in why you chose such a small sample and the patterns of tweets that emerge on Chinese social media.

Professor Pan， Thank you for your time in sharing the research! do agree with you that "governments all over the world impose restrictions on access to digital information". For example, on 5.05, Professor Lazer gave us an excellent sharing about how Twitter suppressed the spreading of fake news during the January 6 insurrection, and there's no clear evidence that government did not have a role in it. In addition, I think the limited quantity of the inflowed information can also be explained by something other than government intervention. Your data collection happened during the most challenging time in China, and Chinese people simply didn't care that much about what the outsiders were commenting. The most relevant issues are staying at home during the Spring Festival and waiting for the victory of the battle. For example, in Data Index 9, the tweet criticized Wuhan people as "worst passengers, no manners, stubborn, uncivilized and dirty." Apparently, it's not true. And compared to the great battle to fight, including building hospitals and supporting Wuhan, Chinese people simply ignored the non-sense criticism. In addition, delivering such information back would harm the friendship between China and the entire world.

I do not agree with the above since what is freely asserted can be freely deserted (apparently).

[yutaili]

Hi Professor Pan, thank you for sharing your work! Very interesting topic. One clarification question: how do you count the matched content between tweets and weibo? I see in the appendix that you use the Word2vec model to calculate the cosine similarity between weibo and tweets, but I wonder how the Word2vec works for Chinese and English respectively? What threshold were you used to determine the content is matched? Another question is that your study is during the covid-19 timeframe which has a lot of particularities. Would you anticipate the same limited inflow during another time period? Thanks.

[a-bosko]

Hi Professor Pan,
Thank you for sharing your work with us! It was very interesting to read about government censorship and how information flows despite this barrier. I am not very knowledgeable in this area of research, so it was very interesting and exciting to read about your research!
In the paper "How Information Flows from the World to China", the authors mention about one-fifth of the content that gains widespread attention on Twitter can be found on Weibo. Was this a surprising finding for you? Is this what you predicted, or did you predict that other media outlets would be more dominant? Also, how can the findings of this study be applied to other cultures and other countries? Thank you!

[FranciscoRMendes]

Thank you for coming Professor Pan! It seems like your work has sparked a pretty intense discussion among my fellow classmates. That's always a good sign.
I am curious, why did you choose the COVID crisis as the sole time frame within which to study censorship? I understand that this is probably the best time to test censorship, however, having some baseline level of censorship to act as a control would have probably made more sense. Censorship of online platforms during COVID was actually quite common and it's hard for me to distinguish between censoring a medically unsound opinion circulated on social media from a politically motivated removal of a social media post. Also, could you walk us through the reaction your research has had among Chinese researchers? Could you characterize their response to your work?

[javad-e]

Thank you for presenting your research at our workshop!
How do you expect the flow of information to be different for topics other than COVID-19? In particular, how are entertainment-related information and contentious information on political matters expected to differ?

[isaduan]

Thank you for sharing your research with us! I am curious how you think of indirect and nuanced ways of speaking that Chinese Weibo users have picked up quickly to talk about sensitive things despite censorship - what proportion of that information would be compared to explicit one? how could we account for the indirect way of communicating ideas? Thank you!

[NikkiTing]

Thank you for sharing your work! I was wondering if you observed any patterns in terms of the types of content that are flowing into China that were facilitated by Weibo users. In your paper, two of the examples you gave were more satirical, rather than providing new knowledge. In addition, all three examples seem like the state has little reason to censor them. Considering this, could it be that although non-state-controlled outlets are contributing to the inflow of information, they are not actually able to contribute substantive information--that “matters” or that can “set the agenda” (i.e., whether because of state- or self-censorship)?

[LFShan]

Thank you for coming Professor Pan, I am very interested in China's censorship and how censorship changes over time. Also, could it be any method to distinguish between self-censorship and government-imposed censorship? (Eg. some content may not appear on Sina Weibo because user chose not to post on it compared to content that is posted and later censored)

[AlexBWilliamson]

Thank you for sharing your research with us - I found it very interesting! I have one main question. You mention that despite the high levels of censorship in China, the vast majority of people living there don't make serious attempts to evade the censorship. How has the Chinese government has gone about censoring the media while avoiding a significant backlash?

[chrismaurice0]

Thank you for coming to our workshop Professor Pan! I am curious about what happens to information inflows when censors are overwhelmed with a plethora of domestic content. Meaning, that if there is a lot of discussion about a topic occurring on Weibo, say about recent lockdowns in Shanghai, are censors overwhelmed, allowing for greater inflow from outside sources?

[awaidyasin]

Thank you for sharing your work, Professor. I had two queries regarding your method:

1. Based on my understanding, the current way of defining inflows does not consider any feedback effects (or a bidirectional relation). It seems entirely plausible that a less viral tweet made its way into Twitter (from Weibo), which gained traction there and, then, got transmitted back to Weibo (with an amplified intensity). The current method would filter out such instances but intuitively these could be considered inflows, given the 'net' inflow effect.
2. Based on your sample tweets, it looks like most of the inflow occurs through influential (or blue-ticked entities), who have the potential to generate viral tweets. Would it be fair to say that person-to-person transfer of information is quite limited?

[luckycindyyx]

Thank you for sharing such interesting work with us. Since this research on information inflow was based on Covid-19 and was to some extent similar to an event study, I was wondering if the conclusion can be generally applied on other events. If so, is there any evidence for the robustness? If not, what could be the potential differences? Thank you so much!

[DehongUChi]

Thank you for sharing your work with us, Professor Pan!
I am wondering if you have done any research regarding account banning due to political reasons on Chinese social media platforms. What are the effects of the bans? How is account banning related to smuggling information from social media platforms outside of China?
Another question is do you worry that your research might be used by other governments who wish to impose control on information flow like China?

[PAHADRIANUS]

Thank you for sharing your study that is particularly unique in this structure. It is a truly daunting to conduct a cross-lingual analysis using computation methods, and your project managed to complete a splendid job with w2v and USE tools, preventing loss/miscommunication of information due to translation and transcription issues. The project is truly enterprising in its advancement towards looking into the entire world ethernet as a whole.
Still, I have following worries and think it may be made more delicate and robust:

1. I am concerned that the 150 tweets sample may not generate representative information: the selection process from the massive original dataset measured in millions was an impressive job, but reduction of this scale also greatly reshaped the sample environment and must be done with some degrees of subjectivity. In addition, the eventual result based on the analysis of such a small sample cannot really be considered a computational method and really forgoes the advantages of the large dataset.
2. I think the explanation on how Chinese authorities control news inflow is too simplified. Using Weibo is a great option since it really comes to personal account to transcribe foreign news outlets while state medias just neglect them. However, there are multiple layers of censorship mechanisms that not only come from governmental intervention, but also autonomous abridgement by posts' authors and the social media platform. In addition, using the compiled historical Weibo dataset could miss a substantial amount of information: (based on my personal observation) during a daily period from midnight to about 4 am, Weibo users tend to communicate in large numbers on matters that would normally be censored, using phrases that could circumvent automated filters, and such records would be deleted by the platform's human supervisors once they come online in the morning. The traces of these communications would not be found in the dataset, but in fact a lot of news inflow occurred and masses of Chinese people were informed. I wonder if using a real-time collected dataset (definitely many more times more work) could evade the issue.

[egemenpamukcu]

Hi Professor, thank you for sharing your work. I found the results on "whose Twitter content happens to appear on Weibo," but are there any clues on what kind of content happens to make its way into Weibo as well? Would it be possible for these deep models to infer what characteristics of a Twitter content, besides its creator, make it more likely to flow into Weibo? What topics and what kind of language is tolerated by the enforcers of censorship for instance, and how has it evolved over time? I guess these questions can be answered without deep learning and through subject matter expertise as well, but I think actually understanding how and in which contexts censorship is enforced by looking at empirical data could be illuminating as well.

[xzmerry]

Dear Professor Pan,

Thank you for sharing such interested research related to the media censorship in China. I have several both conceptual and technique questions regarding cultural and language differences for across-border research.

1. It is found in this paper that the rate of across-border information flow between Twitter and Weibo is one fifth. Whether the rate is high or low when considering different interests and languages between users on Twitter (mainly English speaking) and Weibo (Chinese users)? What is natural rate of across-border information flow between English-speaking vs non-English social media? For example, is there a comparison between research regarding Japan and Korea?
2. I am curious about how the semi-automated system deal with different languages in this across border research. How accurate could it be?
3. My third question is about the types of information flows from Twitter to Weibo as well as the direction of information flow. For the across border information flow, I guess not only the rate but also it matters that what types of information could flow, as well as the direction.

• For example, there is one fifth information flow across border from Twitter to Weibo, but what about the ratio flow from Weibo to Twitter? If the information flow for the different direction is even lower than one-fifth while there is less censorship on Twitter for the known higher transparency in the U.S., then what could be the reason?
• In addition, the rate could be less informative or over-evaluate the information flow if, for example, only good information could flow from Twitter to Weibo.

4. Moreover, I wonder how you interpret the rate of "one-fifth", do you think it is reasonably high or low? And why?

Thank you!

[linhui1020]

Thanks for your share of research! I wonder is there any deplatform strategy for social media platform such as Weibo, and how Weibo adjust their popularity scoring algorithm to serve for the purpose of government as well as oppressing the collective actions?

[xxicheng]

Hi Professor Pan,

Thank you for sharing your work with us! Your early work on Weibo censorship, using observational and experimental designs is one of the first social science articles that I have read. From then on I found this is a fascinating area, so technically speaking, I am a fan :) So excited to have you at our workshop tomorrow! My question is: do you think the Chinese government's control of social media has been changing from 2013 up to now? Are we now in a different world? The present Chinese social media is more divided to me, full of anger between different subgroups (male vs. feminists, nationalism vs. cosmopolitans, etc.), as well as more overconfidence in China. This is very different from how things were 10 years ago, I think. I remember more self-reflection on China or even flattery on the Western world around topics like democracy when I was in middle school. Would love to hear your thoughts! Thank you, and I look forward to your presentation tomorrow!

Best,
Xi

[mdvadillo]

Hi Professor Pan,

Thank you for your presentation today. I found the methodology very ingenious, and was wondering if you were able to get some characteristics about the information that was flowing into China. Would it be the case that the information has certain common characteristics (aside from just relating to COVID-19), or maybe do the opposite analysis and see if the information that did not go into China had a different set of characteristics and features than the set of information that made its way into the country.

[YaoYao121]

Hi Prof. Pan, thank you for sharing such a interesting paper. I like the topic to study the flow between the rest of world and China. While, I wonder whether the finding that approximately one-fifth of content with relevance for China that gain widespread public attention on Twitter appear on Weibo indicates the flow from the World to China. Some topics and opinions might just originate from the worries and focus of domestic people themselves in China. The evidence could show a correlation between Twitter topics and Weibo topics, but may not be the causality.

[fiofiofiona]

Hi Professor Pan, thank you for sharing your research in our workshop. Chinese governmental censorship over information and social media has been a sensitive topic that many researchers have been carefully investigated. As co-occurrence of information on both Twitter and Weibo can be caused by information flow from one to the other platform, could the content originally be generated on a third platform? E.g. a short video clip first being generated on YouTube or Tiktok, and users spread it out to both Twitter and Weibo. In this case, how would you decide the direction of information flow? Thank you.

[mikepackard415]

Hi Professor Pan,
Thank you for coming to the workshop and sharing such an interesting paper. I'm interested in the method of using a word2vec model to first find tweets and weibo posts that are close in vector space, Did you train the word2vec models yourself, and if so, how did you go about aligning them between the two languages? The cross-language nature of this project seems like it must have been a tough methodological challenge. I'd be interested to hear about how you went about developing the method. Thanks again!

[edelahayeUChicago]

Hi Professor,

I've got a question about the way in which tweets are matched to each other. Given that the text need not be the same on both platforms are there any issues on how a text might evolve to include/exclude information that wasn't in the original post. Obviously you use human raters to classify matches but could a tweet be "the same" whilst excluding an important dimension that has implications for the extent to which particular types of information within a tweet makes it from one platform to another?

Many thanks!

[NaiyuJ]

Hi Jen, thanks for discussing this interesting paper with us! Among all sources of information inflows, I find the most interesting case is how the commercialized media in China facilitates this sort of foreign information mobilization. I think it would be really fun if you can discuss more the motivation/ incentives of the commercialized media for disclosing information, the tradeoff between profit and state penalty, what kind of topics they will talk about, and what kind of topics they will not.

[atowey-uchi]

Hi Professor Pan! Thank you for coming to speak with us and sharing your work! I am curious about the role of rare events in your analysis. To focus on a time period in which a great amount of social upheaval is occurring may be a more extreme example of how the government interacts with censorship, especially when the pandemic was not a predicted or planned event. Do adverse, unexpected events curate different censorship patterns than run-of-the-mill day to day?

[cgyhumble0612]

Hi Professor Pan! Thank you for presenting us such an interesting paper and raise our attention on social media censorship in China. Actually, I'm very curious about the financial influence of Chinese censorship on these internet giants such as Sina and Tensent. Can we find other perspective to value the influence of Chinese censorship policies on information flow? Maybe the stock price changes of these social media companies?

[jinfei1125]

Hi Prof Pan, thank you so much for coming to our workshop and a very welcome from Chicago! My question is similar to Jasmine's, how do you separate the effect of language barriers from the government intervention?
Another interesting thing is that I notice recently all social media suddenly support showing the IP address of the users who post and comment, for example, Beijing, Shanghai, Canada, and the US. Do you think this is a government intervention to make Chinese users actively refuse opinions from oversea users?

[zbchen0129]

Hi Professor Pan! Thank you for coming to speak with us and sharing your work! In your paper, you mention 'co-occurring content'. I just wonder how you decide which contents are co-occurring. Besides, what does the 'media or government affiliation' mean? Can users without it escape from the censorship?

[Tanzi11]

Looking forward to your work Professor Pan. I am interested in how social media spaces are treated as the right of the government--in that they can do things such as censorship. Do you think there may be positive repercussions to this as well? (Obviously, censorship is bad, but what about government regulation that can for instance, prevent hate or violence?) Thanks!

[wanxii]

Thanks so much for sharing your work. I wonder if the research could have some more profound implications, e.g. whether the flow of censored information might relate to some economic activities or group attitudes/ideologies. Many thanks!

[FrederickZhengHe]

Thanks very much for this interesting working paper, and I am very eager to listen to the presentation tomorrow. My question is: Why do you choose the +-five-day period of the timestamp of a Tweet instead of +- 3 days or +- 7 days? I think the reason here is not so developed. Many thanks!

[YileC928]

Hi Professor Pan, thank you so much for joining the workshop! The paper nicely depicts the big picture of cross-broader information dissemination, and the methodology part is especially intriguing.
I am interested in learning if your team has attempted to investigate what kind of content is more likely to be transferred from the globe to China. Besides, as censorship is mentioned several times in the paper, have you explored how governmental censorship played a role here (i.e., the counterfactual - what would the sample and dissemination mechanism be like if there were no censorship)?

[Qlei23]

Hi Professor Pan,
Thank you for sharing the research! The idea of co-occuring content between Twitter and Weibo is really interesting. Likewise, I am also interested in why you chose such a small sample and the patterns of tweets that emerge on Chinese social media. From my perspective, the selection bias of global information inflows and the number of "retweets" on Weibo may be another issue as some content may be censored even if it is posted on Weibo in order to limit the number of persons who can view it. I'd like to hear your opinions on such issues. Thank you!

[BaotongZh]

Hi Professor Pan, thank you for sharing such a great work, And as you mentioned in the last section of the paper you only concerned the Weibo and Twitter, I was just wondering how do you think of the information transmission between video platforms(like from TikTok to Douyin(Chinese TikTok)).

[kuitaiw]

Dear Prof Pan，
Thanks a lot for sharing your work. I found your research is covid related. So I wonder if the study has general applicability? That is to say, can this be applied to other events?

[kthomas14]

Thank you for sharing your research with us professor Pan! It is interesting to read about the flow of information into China. I would be interested to hear about an analysis of topics that are not as closely related to events surrounding the pandemic. The topics that people feel compelled to share on a social media from outside sources may often include sensitive topics. I would also be interested to hear about an assessment of the spread of misinformation through inflow.

[y8script]

Hello Prof. Pan, thank you for bringing up this interesting study! I am curious about whether it is possible to identify the characteristics of information inflow through government/state media, overseas entities or individual Weibo users? Do you have hypotheses about which type of information is prioritized by each of the sources? Moreover, why certain information(from non-government/state media) are allowed by the censorship system while others didn't?

[97seshu]

Thanks for sharing your work, Prof. Pan!
I wonder do you notice any regional differences in the intensity of government censorship inside of China?

[Toushirow1]

Hi Professor Pan,
Thanks for sharing this research with us. I am interesting in the language processing for this research. Since there is a huge gap between the meaning of slangs in different culture context, how do you capture this difference in your research. And, do you think the information flowing from world to China affect on the policy making?

[sudhamshow]

Thanks for sharing your research Professor Pan!
I am curious to know more about the dataset you used to train the word2vec model. I ask this because you collect data from January to April and depending on when the data was sampled, the context (and sentiment) of certain word usage might change over time.
I am still unclear about how you would account for the messages on Weibo that would have been removed (since the messages are collected post hoc in April). If the rules or level of moderation changed (adapted) over time wouldn't this be a confounder when trying to study the flow of information influence to/from Weibo?

[siruizhou]

Thank you for presenting this interesting research Professor Pan. I'm interested in knowing how innovative language intentionally curated by Chinese netizens to avoid censorship would impact the information flow.

[zixu12]

Hi Professor Pan, thanks for coming to our workshop! I do agree with some questions posted with other classmates. E.g. during the pandemic, the less inflow of the information might be because people care more their own life.

[ttsujikawa]

Hi profesor Pan,
Thank you very much for presenting your research. I am wondering if there is any application of this research in any other country?

[k-partha]

Thank you for presenting Professor Pan! I was wondering if you have any ideas on how your methods could be used to study political censorship in the US - particularly in quantifying the degree to which either side of the political spectrum is censored, seeing as it is a hotly debated question in mainstream news.

[Hai1218]

Thank you for joining our Workshop, Professor Pan.
As you might have already known, all social media platforms within China are now required to show the geotag of the user as they post, repost, or comment on the platforms.
How do you think the geotagging could alter the current paradigm of information transmission into China? I also feel that future research could pay some attention to comparing the effect of information flow before and after the geotagging mandate was imposed.