In a new paper, we show experimentally that anti-Chinese discrimination boosts support for the Chinese regime among a new generation of Chinese students in the US who may otherwise be most sympathetic to democracy
w/
@xuyiqing
,
@yingjie_fan
, Zijie Shao
For the past few weeks,
@YingdanL_kk
@AnfanChen
& I have systematically collected half a million Weibo posts about Ukraine. It’s true that a large portion of Weibo posts attribute cause of war to US/NATO/“Western” aggression or blame Ukrainian government for conflict.
My book WELFARE FOR AUTOCRATS is now available on Amazon.
What happens when a country becomes obsessed with order? The book shows how China's pursuit of order/stability turned an effort to alleviate poverty into a tool of surveillance and repression
@OUP
Super exciting to receive Advances in Experimental Political Science (
@CUP_PoliSci
)! Learned a ton from Jamie Druckman, Don Green, and everyone who contributed. Useful for anyone doing & interested in doing experiments in social science.
Paper "Strategies of Chinese State Media on Twitter" analyzing CGTN, Xinhua, People's Daily, China Daily from 2013-2020 out with
@yingjie_fan
, Jaymee Sheng
But not all hold this view --- only ~50% of posts. ~10% see Russia as aggressor/blame Putin, 15-20% express other opinions (e.g., sympathy for Ukrainians, criticize Putin *and* Zelensky), and 20-25% share info/facts unrelated to either view (e.g., # of deaths).
Our experiment shows anti-Chinese discrimination (seeing racist comments, Treat B) reduces students' belief that political reform is desirable for China and increases support for authoritarian rule. Non-racist criticisms of Chinese gov't (Treat A) don't have this effect.
In a new (draft) paper with
@xuyiqing
"Gauging Preference Stability and Ideological Constraint under Authoritarian Rule," we show that China is not by any stretch of the imagination a monolithic society that believes in one ideology.
Next step is to train models to classify all posts in our data to see time trends, and analyze Chinese language tweets we've gathered. Our work in
@journalqd
monitoring Weibo during early days of covid laid ground work for this:
Based on a survey of hundreds of Chinese first-year undergraduates in the US and in China, we find that Chinese students who study in the US are more predisposed to favor liberal democracy and less nationalistic than their peers in China.
"CASM: A Deep-Learning Approach for Identifying Collective Action Events with Text and Image Data from Social Media" with Han Zhang in print: . We found >135k protests from 2010-17 in China. Here's the data (posts + events): link:
Effects of discrimination are most pronounced for students who are more pre-disposed to *reject* Chinese nationalism. Results are unlikely explained by relative evaluations of US and Chinese government handling of covid-19.
Our paper "How government-controlled media shifts policy attitudes through framing" w/
@xuyiqing
, ZijieShao is out. We show experimentally how framing policy issues differently allows government-controlled media to move public opinion toward opposing sides of the same issue.
New pub w/
@debutts
in
@Journal_Of_Comm
"Reporting after Removal: Effects of Journalist Expulsion on Foreign News Coverage". What happened to coverage of China after a large portion of the reporting corps of
@nytimes
,
@washingtonpost
,
@WSJ
were expelled from China in 2020? 1/3
We estimate % based on sample of 1000 posts randomly selected from the half a million posts (stratified by day), classified by RAs (trained by
@YingdanL_kk
&
@AnfanChen
around the clock to achieve high intercoder agreement); thanks also to
@debutts
for checking coding rules
New APSR paper "Concealing Corruption: How Chinese Officials Distort Upward Reporting of Online Grievances" with
@Chen_Kerry
is out
, link to replication data at
Huge congratulations to Dr. Yingdan Lu
@YingdanL_kk
on successfully defending her dissertation "Performative Propaganda Engagement: How Celebrity Fandom Engages with State Propaganda in China”! Check out more of her work at
“How Chinese Officials Use the Internet to Construct their Public Image” in print. This came out of my first solo research project as a grad student, where I first learned Python back in 2010
Contrary to common belief, chronological feed did not significantly alter polarization, politics knowledge, or other survey-based outcomes, even though chronological feed led users to spend much less time on FB and Insta + changed what content they saw.
"The Pervasive Presence of Chinese Government Content on Douyin Trending Videos" w/
@YingdanL_kk
shows >40% of trending videos on Douyin (Chinese version of TikTok) come from Chinese gov't, CCP, & official media accounts (what we call regime-affiliated accounts). 🧵
Excited to publish my first image-as-data paper with
@jenjpan
! We combine analysis of video, text and metadata to examine 50K Douyin Trending videos, and find the pervasive presence of the Chinese regime to compete for attention. Also enjoyed all amazing works in this SI, yay!
So many great new articles authored or coauthored by women who also know. This one, "China's Ideological Spectrum," by
@jenjpan
and Yiqing Xu should be on your reading list.
#womenalsoknowstuff
My chapter (Ch29) "Experiments on Political Activity Government Want to Keep Hidden" talks about the design and ethics/legality/safety of experiments in authoritarian context including work by
@david_yang
,
@kgmichelitch
, Jaimie Bleck,
@laurenyoung231
,
@XuXuPoliSci
.
"Does ideology influence hiring in China?" with
@ttzhang107
is out. For the first time since the 1970s, employees of China's public & private enterprises are required to study political ideology. How does this renewed emphasis on ideology affect economic opportunity?
Dibao: The priority of order seeped into Dibao, China’s major social assistance program (中国最低保障). Dibao benefits became used in preemptive policing and surveillance, to deter people from activities deemed by the CCP to threaten order/stability such as protesting.
Stability: Everyone knows “stability” (order) is important to the CCP. Fewer know that what stability means to the regime changed after Falun Gong protests in 1999. Instead of a means of achieving economic modernization, stability increasingly became the goal itself.
Repressive Assistance: The resulting Dibao distribution boosted surveillance and demobilized people (through bonds of dependence and obligation). This erases the delineation between repression and concessions.
Seepage: Prioritizing order led to a process I think happens often (seepage) when one government priority alters the allocation of resources and goals of unrelated policy areas.
In the 2nd paper, we removed reshared content from Facebook feeds. This decreased political news participants saw in feeds, reduced their clicks on partisan news sources & reduced news knowledge but didn’t affect polarization or other attitudes.
Does it work or does it backfire? It’s true recipients are demobilized. But its implementation is error-prone and it undermines policy promises the government has made (increasing discontent and lowering perceptions of government competence and legitimacy)
However, preferences are relatively stable over time (a month). Those with higher education and political knowledge have more stable, more extreme preferences
2 more papers that are part of this effort also out today: Asymmetric Ideological Segregation in Exposure to Political News by
@sgonzalezbailon
@davidlazer
Strikingly, exposure to different frames of same issue moves preferences regardless of predisposition. (There's strong correlations btw predisposition & post-treatment preferences, suggesting preferences are relatively coherent & results not due to demand effects).
@christinelu
@YingdanL_kk
@AnfanChen
We didn't limit collection by geography. Agree it would be great to compare. Hard to do because few users (~5%) enable Weibo geolocation or include geography in profile. We've collected Chinese language tweets and aim to use that for comparison (though it's also imperfect).
Exposure to treatments moves respondents toward policy position espoused in the video, regardless of what that position is. Many effects persist up to 48 hours.
Do we see so much non-propaganda content because this is a proactive strategy (i.e., such videos are boosted to Trending) or because users like this content more (i.e., videos trend organically)? Early evidence suggests the former.
Instead, videos contain human interest stories, stories about positive energy (ordinary people doing good/moral stuff), entertainment, & breaking news completely unrelated to any aspect of the Chinese gov't or CCP.
Share of stories about China did not change, nor did audience engagement with stories; sentiment of stories did not become more positive or negative; outlets did not become more likely to write stories because of actions or pronouncements by the Chinese government. 2/3
@noUpside
@chris_bail
Domestically, where the government can censor as well as produce content, trying to persuade runs the risk of exposing the public to issues you want off the table. Many of the “propaganda” strategies we’ve been studying is about changing behavior, not changing beliefs/attitudes
@GsTs_pl
Our sample is only undergrads. And yes, they are demographically, socioeconomically, and ideologically different from undergrads in top schools in China (Table 1, p17)
@tombschrader
The data we found are social media posts made on Chinese social media sites, e.g., Weibo. I just doubled checked the raw data (publicly available on Harvard Dataverse) and there are no posts made on Twitter or Facebook
Trending videos from regime-affiliated accounts look and feel more like videos from celebrities & influencers (e.g., contain attention-maximizing feature like super short duration, high entropy) than videos from non-official media.
What appears to have changed was the production processes of outlets with expulsions--more articles were written collaboratively (increase in bylines and contribution credits). Paper at: . Replication . 3/3
>65% of these regime-affiliated trending videos have nothing to do with the Chinese gov't or CCP's actions / policies / programs / outcomes / achievements / ideas / orgs / people (let's label this "Propaganda").
We designed a survey of 63 policy questions spanning 6 issue dimensions: political liberalism, market economy, nationalism, traditionalism, social equality, and ethnic accommodation. We fielded the survey with two longitudinal online samples.
@GroseTimothy
I really look forward to reading! I don’t have much empirical data from Xinjiang for the book but I talk about some implications for Xinjiang in the conclusion.
@gpanger
Thanks! The chrono feed ran from Sept 24 to Dec 23 so it continued after the election, which we wanted for the exact reasons you state. We thought things would be settled by the holidays (wrong).
@jugander
That's interesting. I felt like one of the reasons I was asked to join the project was because of my work on China. Change may reflect huge shifts in US-China relations between 2010-2014 and 2020.
Treatment: domestic & foreign video news segments in style of CCTV. Domestic: SOE reform should be *Market* vs. *State* led. Foreign: China should take *Dovish* vs. *Hawkish* stance in South China Sea disputes w/ Philippines. *Control* segments show no policy stance.
What types of firms are most likely to penalize non-conformity? Exploratory analysis suggests it's firms designated as CCP strategic priorities with high innovation requirements (e.g., AI, 5G, biotech, new energy, smart/nanomaterials, geospatial tech).