Do 'most' Chinese really favor outright war with Taiwan? (Sort of, but not really.)

Don't ask how Taiwan should cross the red bridge. Ask whether.

Much fuss was made of a recent study in the Journal of Contemporary China assessing public support in China for various means of annexing Taiwan to China.

The authors don't don't use the term 'annexation', but that's what it would be -- not only does no option for peaceful dialogue leading to mutually-agreed unification appear in their so-called "full range" of policy options, but such an option is indeed unlikely to work. There's no plausible scenario in which Taiwan would want unification with China, and the resultant coercion to achieve that goal is, by definition, annexation.

I'm getting ahead of myself, however. I want to talk about the main finding from this article, published in media such as the South China Morning Post. Their headline reads "Just over half of mainland Chinese people back full-scale war to take control of Taiwan, poll finds". 



                         



I can't read the whole SCMP article as I'm not a subscriber and not interested in becoming one, but it leads with the assertion that "55 per cent in favour of “launching a unification war to take back Taiwan entirely”, with a third opposing it and the remainder saying they were unsure."

That sounds scary indeed. What are we to make of it? To be honest, not that much. I obtained a copy of the original publication -- not necessarily easy as I lost my academic access a few years ago -- and read it to see if it really does support the idea that most Chinese are "in favour" of all-out war to take Taiwan.

Well, that is indeed one possible interpretation of the results. However, I think it's a bit exaggerated, if not outright skewed. I don't just mean that the authors come to more moderate conclusions (though they do), but that I genuinely don't think a careful reading of the study supports the idea that so many are "in favour" of war, so much as they believe war to be one acceptable possibility among many. I'm no master of methodology, but I also think the study has some methodological issues and interpretation bias. 

Let's start with the finding heard 'round the China Watcher Circles. It comes from this figure: 

Yes, it's true that 55% of respondents said they supported "full scale war", with 33% opposed. The opposite -- "separation" -- had 22% acceptance, with 71% opposed. 

I am putting "separation" in quotes because it implies a change in the status quo. However, China and Taiwan are already separated under the status quo -- the PRC does not govern Taiwan -- so there would be no act of separating. It's a telling clue that the authors, despite their best intentions, are biased.

In other words, everything but "separation" was acceptable to the majority. This doesn't imply they are "in favour" of full-scale war. It means they find it one of four acceptable policy options presented, with less-aggressive options also on the table. The exact same number of people support the status quo as full-scale war, with likely quite a bit of overlap.

Although it's impossible to say without more data, one might reasonably infer that many such people would prefer not to resort to full-scale war unless other options are exhausted. Or perhaps many don't actually feel strongly about fighting a war for Taiwan, but feel they have to indicate acceptance for any number of reasons. 

It's still worrying, as when those options fail -- which they will -- these respondents say they find full-scale war an acceptable solution.

However, it's not as scary as 55% of Chinese wanting to go straight to war. I don't know if SCMP clarifies this later in the article as I can't read it, but you wouldn't get that impression from the free-to-read blurb.

The authors, to their credit, do point this out. They even mention that only 19 (not 19%, but 19 respondents total) supported full-scale war as the only option. Not 55%. Not even 19%. 19 people. Here, they explain it a little further: 

A closer look at the data suggest that the respondents largely fall into three categories based on their answers. First, 313 respondents (17.1%) are ‘pacifist’, who found either or both of the two non- aggressive options (‘status quo’ and ‘separation’) acceptable while rejecting the other three options. Second, 572 respondents (31.4%) are ‘bellicose’, who found some or all of the aggressive options (‘sanction’, ‘military coercion’, and ‘full-scale war’) acceptable while rejecting the other two options. Finally, the remainder of the respondents are ‘ambivalent’ (939, or 51.5%), as they endorsed both aggressive and non-aggressive options or were unsure on some or all of the policy options. Importantly, this simple partition of the respondents points to an even smaller share of citizens in support of the aggressive policy options. [Emphasis mine]. In fact, only 19 out of 1,824 respondents (or about one percent of the sample) rejected all but the most extreme option of armed unification.

Basically, to say more than half of respondents "favor full-scale war" is not quite right. More accurately, about half of all respondents favor some move toward unification, but it need not be war and indeed, might be peaceful. One-third view it as unacceptable, and the same percentage who do find war acceptable also 

The authors removed respondents who found all five options acceptable or unacceptable, for a total of 259 removals: 70 who found all options acceptable, and 189 who found all options unacceptable (I'm very curious about what this latter group tends to think are optimal policy options instead, but doubt I'll ever know). That's more people removed for finding both "separation" and "full-scale war" acceptable than people who favored only full-scale war. I do understand that such a response means they probably didn't take the survey seriously, but that it outweighs the number who only support the most bellicose option is, I believe, telling.

Frankly, this isn't surprising. In fact, given the extremely skewed and brainwash-prone education system in China, I would have thought support for military coercion, if not all-out war, would have been higher and support for Taiwanese independence would have been lower. That about one-third of respondents are fine with Taiwan maintaining its independence is actually more hopeful than I would have predicted! 

The authors point this out, too: that support for full-scale war is not as clear-cut as it may seem, and Chinese leadership should understand this. 1/3 of Chinese accepting Taiwanese independence whereas only half wanting war -- and almost no one wanting war as the only option -- is not the solid wall of public support that the CCP needs to attack Taiwan. I can imagine that quite a few Chinese who find war an "acceptable" last resort would nonetheless be very angry if a war were launched without attempting more "peaceful" means first. 

I also want to point out that the authors did not, in fact, include "all" policy options. Although I don't think unification would come of it -- frankly, none of these options would bring about "peaceful" unification, as Taiwanese simply do not want it, and being sanctioned isn't going to change that -- they fail to include "dialogue", "persuasive incentives" or "supporting preferences for unification among Taiwanese" (that is, helping out the unificationists in the KMT or more radical parties such as the New Party), or any truly peaceful means of pushing their agenda. This really isn't a full slate of policy options! 

They further separate "the status quo" and "separation". I could call this bias, but I'll be generous: perhaps they rightly believe that Chinese view these two concepts differently. Perhaps they view "separation" as China officially giving their blessing for Taiwan's sovereignty, and 'the status quo' as China neither recognizing Taiwan's independence nor doing anything about it. While I don't think there's a big difference -- the status quo is that Taiwan is not governed by the PRC, period -- I can understand that Chinese citizens might. 

The authors are indeed fairly reasonable in their interpretations of the data gathered. I'll quote at length here as this study is not readable to all: 

Conventional wisdom holds that the call for armed unification has been ramping up in mainland China in recent years,60 setting the stage for ‘an all-out war . . . devastating to all’.61 Despite the media hype, there is scant empirical evidence indicating the extent to which the Chinese public would support such a war rather than non-violent means to unification. Understanding mass support for the different policy options Beijing could adopt to ‘resolve the Taiwan Question’ is important because we know this is one issue about which public opinion holds sway over Chinese leaders....

We find that at the aggregate level, only a slim majority of the respondents are explicitly supportive of waging a unification war, which has been the focus of current policy debates and academic research, and a third of them are explicitly opposed to it. These numbers are consistent with a survey conducted on an urban sample in 2019,62 thus bolstering our confidence in the external validity of the findings. This also suggests that public support for armed unification has remained relatively stable, despite the rapid deterioration in Beijing’s relations with both Washington and Taipei....

Our study has important policy implications. Both pundits and policy makers who sound the alarm for an imminent or inevitable war in the Taiwan Strait, one that likely would involve the US and its allies, implicitly assume that Beijing’s hands are tied because most Chinese support ‘wutong’ and the public’s patience is wearing thin. An ambitious paramount Chinese leader who cares about his domestic audience can only make things worse. Our findings suggest that this pessimistic outlook may be based more on myth than on reality.

That said, I am not at all sure that they controlled for "social desirability bias" despite claiming to attempt to control for this (however, at least they mentioned it!) Having respondents analyze each option isn't a bad idea, but avoiding social desirability bias first requires guaranteed anonymity -- something that is impossible online in China and that any thoughtful respondent would realize was not necessarily a given. They might have opted in, but also been thinking (not without reason) that someone, somewhere was monitoring the answers.

Avoiding social desirability bias also requires evaluating questions carefully so as not to induce answers that respondents might think are "acceptable". The authors claim to have done that by avoiding the term "Taiwan independence" (smart), but as above, they did not offer all the possible non-military policy options, instead presenting two non-coercive and three somewhat or very coercive choices. That is, more coercive choices than non-coercive ones. Respondents might therefore feel a nudge that it's "better" to choose some form of coercive method, even if it's just sanctions, rather than none at all. Given the chance to say any or all given option is acceptable, why not also choose the one that the government so plainly wants you to support (war), as you can also choose the one you might think is more reasonable (such as sanctions)?

Regarding that bias, the authors do point out the role of the education system in China. They're clearly aware of the role of oft-repeated propaganda. That said, I'm still not sure they truly understood the impact this might have on the results. When people are fed fairly simple slogans (think "national rejuvenation of the great Chinese nation!" or, in Taiwan years ago, "long live the Three Principles of the People!") they might repeat those slogans in a "normal" and peaceful environment. It's so normalized that there probably isn't much of a second thought -- you repeat the slogan without much thought for the actual policy underpinning it or its details. How many Taiwanese once chanted "long live the Three Principles!" without thinking or caring much about what the Three Principles actually were?

I don't see a way to control for this: as with polls showing most Taiwanese are willing to fight to defend Taiwan, we can only gauge what respondents say they believe about war in the moment, not what they'd think or do under actual wartime circumstances. However, Taiwanese aren't pushed by society, education and the government every day to engage in sloganeering and not think too much: everyone is free to say they'd fight or not. You won't be put on a watchlist. That is to say, it's more difficult if not impossible to truly control for social desirability bias in an authoritarian regime. Offering five different options for analysis is insufficient to counter such inculcation. Why not acknowledge this?

Although I applaud where the authors attempt to either reduce or confront bias regarding outcomes, some parts of the introduction and analysis have me scratching my head. Towards the end they reference "The Taiwan Question" -- a heavily loaded term that has connoted genocide when used to describe other groups such as Jewish people and Armenians -- seemingly without much thought. They treat the 1992 Consensus as a real agreement between the two sides, when it was not a consensus and not even called such until 2000. Taiwan/China history is presented in a way that makes it seem as though Taiwan actually was fully controlled by China before 1895, when it wasn't. It was considered and treated as a colony, and for most of the centuries China "controlled" Taiwan, they only really controlled the western third or so. Total Chinese governance of all of Taiwan didn't last long at all. 

To their credit, they do not use the term "reunify" except in quotes, although they pepper the term "mainland" a little too liberally and do call it "returning" and "national unification", forgetting that, of course, Taiwan is not part of the current PRC "nation" and that Taiwan does not have a mainland. Taiwan "returning" to China is used only in reference to what respondents might believe; nevertheless, it's not in quotes -- it's treated as the appropriate term. While overall the authors do take a moderate tone, little things like this worry me.

They spend a lot of time going over China's position, and the US's as well. Taiwan's position doesn't get much space, however. It's limited to Tsai's "refusal" to accept the "1992 Consensus" (which is not called out as fabricated at any point), and this half a paragraph:

In Taiwan, the Election Study Center of National Chengchi University, which has been tracking the unification versus independence stances of the Taiwanese since 1994, reported the smallest pro- unification margin in 2022. In the same year, as much as 73% of Taiwanese explicitly expressed their willingness to fight should the mainland use force.

Cool, but they don't actually say that unificationism enjoys support that is more or less tied with the margin of error, and they don't mention that most Taiwanese identify as solely Taiwanese, not Chinese at all.

Towards the end, the authors imply (although don't outright say) that unification would be a good thing. They do directly state that Taiwan's annexation would be a "return": 

One contribution of our survey is that we move beyond the focus on armed unification. In doing so, we provide a more nuanced understanding of public preferences for a broad range of peaceful and non-peaceful policy options that Beijing could adopt in achieving unification during Tsai’s second term: military coercion short of full-scale war, economic sanctions, and embracing the status quo and waiting patiently for Taiwan’s return.

Okay, but why would anyone outside China want to "achieve unification"? Yes, the point is to see how Chinese feel about these policies the CCP "could adopt", but something about the tone is off. The authors are careful academics who do not say that the less violent but still coercive options might be good choices for China, but it sure reads as though they assume that putting some short-of-war pressure on Taiwan might not necessarily be a bad thing.  

Even where they don't offer a fully China-centric perspective, their word choices give questionable vibes:
 

Furthermore, we find all of these policies receive levels of support similar to that for full-scale war. And quite surprisingly, about one fifth of the respondents even find acceptable the unthinkable option of allowing the two sides to go their separate ways, which is tantamount to de facto independence for Taiwan.

What does "de facto" mean here? Taiwan is already de facto independent; that's a present reality. Why treat it as a future outcome? "Go their separate ways"? China and Taiwan are already governed separately! And sure, Taiwanese independence is "unthinkable" to the Chinese government. But why is Taiwan independence unthinkable in any broader sense that justifies its use without quotation marks? It is indeed very thinkable -- I think about it all the time! 

Although the authors do state a Chinese victory would be "pyrrhic", their choices regarding what background to include, as well as their word choice, indicate to me that they view the somewhat-coercive policy options as part of a potentially reasonable, justifiable endgame.  It all sounds very neutral on the surface, but I have to ask whether the language choices reveal a potential bias.

Of course, there is a difference between gauging what respondents think and what is actually true. For example, a discussion of what Chinese responders might think of the KMT and its ability to "facilitate the peaceful resolution of a cross-strait crisis" might differ markedly from what the KMT could actually do: that is to say, not a lot without general public approval. They might win in 2024, but that does not mean they have a mandate from the Taiwanese people to negotiate away Taiwan's sovereignty. They simply do not.

Also consistent with existing survey findings about Chinese public attitudes towards the western world,53 respondents with better knowledge about PLA development are more likely to prefer the more aggressive policy options. And those who believed that a KMT government could better facilitate the peaceful resolution of a cross-strait crisis in the future were more ambivalent, possibly hoping the KMT will win the 2024 election after Tsai’s second term. 

It's telling, however, that they don't differentiate much. They call it a "peaceful resolution of a cross-strait crisis" as though that's what selling Taiwan to China would actually be; they are not clear that this is how Chinese respondents might view the situation. They do not examine the possibility -- dare I say likelihood -- that a KMT win does not mean that Taiwanese are receptive to unification. There is difference to be explored here, and the authors do not explore it. They seem to equate "cross-strait peace" with the pro-China leanings of the KMT, as though the only obvious way to ensure "peace" is to move toward unification. The opposite is true: a decisive move toward unification is just as likely to precipitate war.

All in all, I do believe the researchers had good intentions. They don't seem to be unificationists even though some of the language employed and assumptions made were at times questionable. There was insufficient differentiation between language used to describe general sentiment in China (not "the truth", but Chinese perspectives), and the actual situation between Taiwan and China. 

Some flaws in the study, e.g. the difficulty if not impossibility in guaranteeing confidentiality, without which controlling social desirability bias is impossible, were not discussed from a methodological perspective. 

While the authors were circumspect and careful in their own interpretation of the findings, SCMP's portrayal of them, at least the lede presented in the 'free' nubbin of text, is highly questionable. I may have questions about the study, but the media is the bigger problem here. Can we really say that 55% of Chinese "favor" full-scale war when the exact same percentage can be said to "favor" the status quo?

Although I have some questions about the study itself, the overall findings don't fill me with concern, and they shouldn't worry you all that much, either. 

Although SCMP may not agree, if Liu and Li's research should keep anyone up at night, it's CCP officials who do need solid, large-majority public support for a full-scale war for Taiwan. Without it, everything from protests to difficulty conscripting soldiers who will fight fiercely for Taiwan will be more difficult and internal governance will be far more challenging. Right now, it seems they don't actually have the support they truly need. 

Perhaps they should heed the 55% who find the status quo acceptable, not the 55% who find war acceptable. If they're not the exact same 55%, it's probably pretty close.

And no one at all should heed the South China Morning Post.