Consistent with other opinion polls in Taiwan, party affiliation has a lot to do with views about credibility. DPP supporters are more likely to find 🇺🇸 credible & KMT supporters are more likely to find 🇨🇳 credible. We did not ask about the relative credibility of 🇺🇸&🇨🇳. 3/-

Identity is correlated with views on the credibility of 🇺🇸&🇨🇳. Those who identify as Taiwanese are more likely to find 🇺🇸 credible and less likely to find🇨🇳 credible. The reverse is true of those who identify as exclusively Chinese. Those who identify as both fall in between. 4/-

Most polls, such as the one by the Election Study Center at NCCU, show ~60% of respondents identifying as exclusively Taiwanese, ~30% as both Taiwanese & Chinese & ~2+% as exclusively Chinese. 5/- esc.nccu.edu.tw/PageDoc/Detail…

Preferences on cross-Strait ties is also correlated with the views on the credibility of 🇺🇸&🇨🇳. Those who prefer independence (now or later) are more likely to find🇺🇸credible and less likely to find🇨🇳credible. The reverse is true of those who prefer unification (now or later).6/-

So what can 🇺🇸 do to make the Taiwan public believe that it might support Taiwan’s defense? We asked in Wave 2. High-level visits, statements of commitment by POTUS & arms sales all bolster this belief. However, air & naval patrols by the 🇺🇸 military have the largest effect. 7/-

What does all this mean? In short, credibility of both 🇺🇸 & 🇨🇳 may be limited & falling among the Taiwan public, but 🇨🇳 has it much worse. The Taiwan public may be unable to trust any 🇨🇳 proposal, esp without credible commitments. 8/- cdn-japantimes.com/wp-content/upl…

🇺🇸 does better, but has work to do. Taiwan’s public would like to see 🇺🇸 put skin in the game. Dual deterrence (“strategic ambiguity”) may be confusing Taiwan’s public, which is pro-status quo rather than reckless. 9/- esc.nccu.edu.tw/PageDoc/Detail…

In a democracy like Taiwan, public opinion is probably the best restraint on politicians. For 🇺🇸 cross-Strait policy to be effective, it must sufficiently account for ties with both Taiwan & 🇨🇳. For Taiwan, this must include the public. 10/- freiheit.org/sites/default/…

Please see the note below on our sampling strategy & other methodological approaches. 11/-

Fascinating work by Wen-Chin Wu , @hhp_tw , Chien-Huei Wu & James Lee on Taiwan public opinion re: credibility of 🇺🇸&🇨🇳! Some info is already out there. Respondents sees credibility of 🇺🇸 as low, and are even less trustful of 🇨🇳. 🇺🇸actions can affect trust.

Some of the study already received some attention, thanks to Bonnie Glaser / 葛來儀, Kathrin Hille, Demetri & Financial Times.

Also, from the FT piece, it seems that 81% of respondents think the state of Taiwan-🇺🇸 relations are about right (42%) or can get stronger (39%). Wonder how subsequent waves of 🇨🇳 intimidation via military exercises will affect opinion.

Might be a useful reference following all the excitement with visits & military exercises.

大家好！We’re delighted to see American Portrait get traction! We’ve been working hard on the survey & to get some initial findings out. Given that there’s always some risk of misinterpretation & 🍒picking, we thought we’d share an explainer. A 🪡🧵 1/

Check out how our team member, James Lee, talks about our survey results: ucigcc.org/podcast/dispat…

Excited to share our study examining the reactions of the Taiwanese public to the military flybys by and the US in the Taiwan Strait. @WENCHENGFU, Mao-Chia Sun, Wen-Jian Huang The China Quarterly doi.org/10.1017/S03057… 1/3

We conducted a survey experiment in Taiwan in October–November 2020. We find that Taiwanese are less sensitive to the Chinese military presence in Taiwan's ADIZ but have become more supportive of de jure independence after seeing the US aircraft in the area. 2/3

The findings may add to the ongoing debates over China's deterrence strategy towards Taiwan and the US policy of strategic ambiguity towards the Taiwan Strait. 3/3

The new paper with Tae Jun Seo (my former student at Dartmouth) is now published with page numbers in the Journal of Conflict Resolution (doi.org/10.1177/002200…). 1/5