2. Elite cues to US public opinion on foreign policy making

Early works debate whether people can collect information and formulate coherent decisions about foreign policies (detailed review in Baum and Potter, Reference Baum and Potter2008). The latest literature on audience costs shows that public opinion influences foreign policymaking in democracies where the general public holds politicians accountable for their policies, foreign policies included (Fearon, Reference Fearon1998; Tomz et al., Reference Tomz, Weeks and Yarhi-Milo2020). Empirical research further confirms that public opinion influenced the US use of force abroad in general (Tomz, Reference Tomz2007; Tomz et al., Reference Tomz, Weeks and Yarhi-Milo2020) and in specific countries, including Nicaragua (Sobel, Reference Sobel2001), Somalia (Klarevas, Reference Klarevas2002), and Iraq (Larson and Savych, Reference Larson and Savych2005; Voeten and Brewer, Reference Voeten and Brewer2006).

The next question is how the general public makes a consistent judgment call on foreign policies remotely related to their daily experiences (Rosenau, Reference Rosenau1965). There is an information asymmetry between elites and the people (Colaresi, Reference Colaresi2007; Baum and Groeling, Reference Baum and Groeling2010). Elite cue-tasking theory suggests that the public takes an information shortcut by receiving cues from the power elites who fill in people with what they would like them to think about foreign policymaking (Brody, Reference Brody1991; Zaller, Reference Zaller1992; Berinsky, Reference Berinsky2007, Reference Berinsky2009).

Based on the logic, I examine whether the US public opinion is in line with power elites' cues, focusing on the China–Taiwan tensions where Taiwan is arguably one of the most likely places where the USA would engage in a war with China (The Economist, 2021).

3. US public opinion on the cross-strait relations

Scholars have paid close attention to the China–Taiwan tension, which has been one of the essential issues in US–China relations, where public opinion in the USA arguably has a role in shaping the formation of the US foreign policy toward cross-strait relations. A few studies account for how US public opinion has exerted such influence.Footnote ¹

The China–Taiwan tension took root in the Chinese Civil War (1927–1949), fought between two belligerents, the Kuomintang-led (KMT) government of the Republic of China (ROC) and the Chinese Communist Party (CCP). The CCP eventually won the war, controlled mainland China, and founded the People's Republic of China (PRC) in 1949, while the ROC retreated to Taiwan.Footnote ² The PRC and ROC governments claimed sovereignty over mainland China and Taiwan in their constitutions.

In 1954, the first Taiwan Strait Crisis broke out when China attacked the Tachen Islands. By heeding the first example of President Truman, President Eisenhower sent the US Navy's Seventh Fleet to blockade the Taiwan Strait to defend Taiwan from China's attack. According to State Department public opinion poll in September 1954, 53% of the respondent favored ‘giving the Chinese Nationalist government in Taiwan all the help it needs,’ while 33% disapproved and 14% had no comments. Of the 53% supporting coming to Taiwan's rescue, 20% favored sending US troops, and 31% supported only some undefined aid (Olivia, Reference Olivia2018: 124). Later, when the Eisenhower administration threatened a nuclear attack on China in 1955, the US public opinion favored a diplomatic and peaceful solution to the China–Taiwan tension (Irish, Reference Irish1960; Kau et al., Reference Kau, Perrolle, Marsh and Berman1978; Soman, Reference Soman1994; Olivia, Reference Olivia2018). According to the Department of State Office of Public Opinion Studies records, Olivia shows that the opposition to the present US policy to intervene in the Taiwan Strait rose from 33% (September 1954) to 50% (March 1955) and 90% (April 1955) (2018: 149). President Eisenhower eventually aborted the use of nuclear weapons and turned to cease-fire negotiations with the PRC and ROC governments.

In 1958, the Second Taiwan Strait Crisis broke out. Americans' preferences again played an essential role in how the US government dealt with when China launched a surprise attack on Quemoy (also known as Kinmen) and Matsu islands, Taiwan's fortified outposts close to China. Based on a Gallop poll, 62% of Americans opposed using nuclear weapons against the PRC (Eliades, Reference Eliades1993). The Eisenhower government ultimately bent on the US public's will to reject the use of nuclear weapons. President Eisenhower sent US aid and small-scale military intervention to contain China's attack. The PRC may not aim to take over the ROC's offshore islands (Christensen, Reference Christensen2011; Taylor, Reference Taylor2011). Ultimately, ‘the US public opinion succeeded in bringing about a major shift in the American policy in the crisis’ (Soman, Reference Soman1994: 374).

In the face of confrontations between China and Taiwan after two Taiwan Strait Crises, the US policy on China has been a hot topic of debate in academic and policymaking circles. Eventually, the engagement camp gained the upper hand in hoping for future democratization in China (Shambaugh, Reference Shambaugh, Shambaugh and Robinson1994). In 1978, China initiated economic reform. Concurrently the Carter administration recognized the CCP-led government of the PRC in 1979 and severed diplomatic relations with the ROC. To counterbalance this, Congress passed the Taiwan Relations Act in 1979, supporting US arms sales to Taiwan for defensive purposes and a peaceful solution to the China and Taiwan tension (Taiwan Relations Act, 1979). Since then, the USA has engaged with China and implemented strategic ambiguity in its security commitment to Taiwan (Campbell and Mitchell, Reference Campbell and Mitchell2001; Zhongqi, Reference Zhongqi2003; Kastner, Reference Kastner2006).

Nevertheless, the US public was polarized about whether the USA should commit to defending Taiwan against a Chinese attack, and their support for militarily defending Taiwan has gradually risen (Kau et al., Reference Kau, Perrolle, Marsh and Berman1978: 135). Based on various polls, Russett and Deluca indicated that the support for using force in defending Taiwan started at 17% in 1974 and rose to 43% in 1980 (Reference Russett and Deluca1981: 387). They concluded that Americans' willingness to defend Taiwan slightly tilted from dovish toward hawkish.

In 1989, the outbreak of the Tiananmen Massacre in China brutally suppressed the Chinese people's demands for democracy, which were growing as they watched the fall of the Soviet Union. The US engagement policy with China encountered a temporary setback due to economic sanctions on China during Reagan's presidency. The next administration, under Bush, soon restored the engagement policy after lifting the ban on China. In 1996, the Third Taiwan Strait Crisis broke out when China held missile drills to intervene in the first presidential election in Taiwan (Scobell, Reference Scobell2000). President Clinton sent two aircraft carrier groups to prevent the tension from escalating (Ross, Reference Ross2002). In a 1996 survey, Tien and Nathan suggested that, due to China's aggressive conduct, approximately 40% of Americans supported the US government militarily defending Taiwan in a Chinese attack (2001: 126). A 1997 Gallup poll showed that 65% of Americans felt that taking a stance against China's forceful takeover of Taiwan was more important than maintaining good relations with China (Newport, Reference Newport1997).

Given that the engagement policy was a significant theme for the Clinton and Bush Jr. governments, some polls found a relatively low supporting rate for defending Taiwan with military force. A 1999 Gallup survey showed that only 27% of Americans favored US troops if China should invade Taiwan (Saad, Reference Saad1999). According to a 2008 survey, only 32% of Americans supported sending troops to Taiwan if China should invade Taiwan (Page and Tao, Reference Page and Tao2010a: 64). In a 2015 survey, the Chicago Council on Global Affairs (CCGA) found a 28% support rate for US troops defending Taiwan (2016).

Previous literature meticulously documents the trend of Americans' attitudes toward defending Taiwan. However, it could be more satisfying in the following regards. First, current research has yet to systematically link China's image to the China–Taiwan tension, instead surveying and reporting on them separately. However, the linkage is significant, as Taiwan has become one of the most likely hot spots for a third world war involving the USA and China (Tucker, Reference Tucker2008; Blackwill and Zelikow, Reference Blackwill and Zelikow2021). This paper intends to fill the gap by presenting survey evidence on such a link.

Second, existing studies have yet to account for how the general public in the USA relates China's image to the China–Taiwan tension for two reasons. The US China policy was one of the most critical themes in the 2016 US presidential campaign (Etzioni, Reference Etzioni2016). A drastic downturn in the US China policy took place under the Trump administration. Since the Republicans and the Democrats have gradually come to a consensus on viewing China as a threat, they would continue to support a hawkish China policy (Leung and Depp, Reference Leung and Depp2019). The latest survey by Page and Tao (Reference Page and Tao2010a), with its relatively low public support for sending troops, was implemented before the first term of Obama's presidency. Thus, it is essential to observe whether the Americans followed suit after President Trump's unprecedented rhetoric about China's threat to the USA. Public opinion was arguably consequential to the USA's use of force to confront China to defend Taiwan. There is a need for an updated study on US public opinion to investigate how China's image affects the tension between China and Taiwan.

4. Elite cues for public opinion on China's threat and the China–Taiwan tension

In this paper, I argue that given power elites' loaded cues about China's threat to the USA, the military threat enhances Americans' attention to the China–Taiwan tensions. After decades of economic prosperity, China became more open and liberal during General Secretary Jiang and General Secretary Hu. When Xi took the helm, China drastically regressed by tightening up all China's social, economic, and political controls and lifting the President's term limit instead of becoming more free and liberal (Economy, Reference Economy2014, Reference Economy2018). Next, Xi Jinping ended the rules and norms of the two-term limit for the last two decades and declared to serve the third term as General Secretary at the 20th National Congress of the Communist Party of China from October 16 to 22, 2022. The rise of an authoritarian leader like Xi in China proves that the USA failed to produce a more open and liberal China through an engagement policy. The bankruptcy of the engagement camp in fostering a democratic China has given rise to the containment camp, which urges the USA to take preemptive actions to contain China, now seen as a revisionist state challenging the interests of the US hegemon (Roy, Reference Roy1994; Broomfield, Reference Broomfield2003; Friedberg, Reference Friedberg2005; Zbigniew and Mearsheimer, Reference Zbigniew and Mearsheimer2005; Kirshner, Reference Kirshner2012; Allison, Reference Allison2017).

Despite the debates about China's threat to the USA, Americans' attitudes toward China were polarized. In 2001, Americans generally saw China as a threat to current and future American interests but not a major threat (Tien and Nathan, Reference Tien and Nathan2001). Fourteen years later, most Americans still felt the same way. According to a 2015 survey, China is a minor threat to 41% of Americans, with 31% believing that China's military and economic power constitute critical threats to the USA (Chicago Council on Global Affairs, 2015).

Trump, his campaign team, and later his administration repeatedly emphasized that Chinathreatens to the USA and even the world on military and economic fronts (Stracqualursi, Reference Stracqualursi2017; Rappeport, Reference Rappeport2019). He painted China as a major threat to the USA via social media, TV shows, and campaign rallies (Stracqualursi, Reference Stracqualursi2017). He even doubts the necessity of the One-China policy (Bohan and Brunnstrom, Reference Bohan and Brunnstrom2016). Before Trump took over the presidency, most Americans (55%) held an unfavorable view of China, the USA's largest Asian competitor (Wike and Poushter, Reference Wike and Poushter2016). At the beginning of the Trump presidency, the Pew Research Center fielded a survey in January 2017 showing that ‘65% said China is either an adversary (22%) or a serious problem (43%), while only about a third (31%) said China is not a problem’ (Manevich, Reference Manevich2017). Such a change in Americans' view of China is understudied for how it influences Americans' evaluation of the China–Taiwan tension.

In this paper, I argue that China's threat influences Americans' perception of the China–Taiwan tension concerning its level and the issues seen as relevant. Specifically, Americans are wary of China's military strength, aggravating the perceived severity of the China–Taiwan tension. Given the geographical significance, Taiwan is one of the flashpoints where China would most likely engage in war with the USA. When China looms large as a threat to the USA, its tension with Taiwan makes an alarming signal of China's intention to shake up regional stability for Americans.

The more Americans pay attention to the China–Taiwan tension, the more likely they perceive China as a major threat to the USA. However, China's intention to take down Taiwan has been a consistent policy since the founding of the PRC. It is China's military ambition that has changed in recent years. About six months after Trump swore into the presidency in July 2017, PRC gradually increased its military presence near Taiwan (Gady, Reference Gady2017). By the end of 2017, Reuters reported that high officials of the PRC claimed an ‘overwhelming advantage’ in bringing Taiwan to heel (Reuter Staff, 2017). Such aggression raised awareness and wariness of cross-strait tensions worldwide, including among the general public in the USA. China's threat induces Americans to take China's tension with Taiwan more earnestly than the other way around.

The first hypothesis is derived accordingly to empirically check whether Americans take the China–Taiwan tension more seriously when they perceive China as a grave threat to the USA.

Next, what aspect of China's threat draws Americans' attention to the China–Taiwan tension? I argue that concerns over China's military rather than economic strength induce Americans' perception of the China–Taiwan tension as crucial to the USA. When China's military strength concerns Americans, it makes Americans uncertain about the chance for the USA to win a war with China. It induces them to care about the China–Taiwan tension, one of the most likely reasons the USA and China would engage in war. Essentially, the USA and China disagree on how to resolve the Taiwan issue.

Poll results show that many Americans are worried about China's growing military power. In a 1997 survey, 72% of Americans said that opposing China's sale of nuclear weapons worldwide is more important than maintaining good relations with China (Newport, Reference Newport1997). In a 2008 US survey, Page and Tao found that a substantial minority saw China as a ‘critical threat’ developing into a world power, focusing more on its military than economic power (2010a). Also, more than 70% of Americans are worried about China becoming a future military threat to the USA (Page and Tao, Reference Page and Tao2010a).

China claims that it would only use force against Taiwan. China has upheld the One-China principle with a focus on Taiwan being part of China since the foundation of the PRC. The CCP's General Secretary Xi Jinping initiated military and economic actions in response to the US balance of power against China (Dian, Reference Dian2015). At the 19th National Congress of the Communist Party of China, General Secretary Xi Jinping mentioned the importance of ‘Two Centenaries’ in his speech. The year 2021 is the first centenary of the founding of the CCP, and 2049 is the second centenary of the founding of the PRC. It is debatable whether they are deadlines for unifying Taiwan by force (Lams, Reference Lams2018; Bush, Reference Bush2019; Mazza, Reference Mazza2021; Sacks, Reference Sacks2021). China's aggression against Taiwan gradually looms with a higher frequency of military presence near Taiwan.

Nevertheless, the US government clearly states that it considers ‘any effort to determine the future of Taiwan by other than peaceful means, including by boycotts or embargoes, a threat to the peace and security of the Western Pacific area and of grave concern to the United States’ (Taiwan Relations Act, 1979). To further emphasize the role of Taiwan in the US grand strategy, the Trump government and Congress reinforced the importance of Taiwan's capacity for self-defense against China by bypassing and signing the 2018 National Defense Authorization Act and Asia Reassurance Initiative Act into law (Asia Reassurance Initiative Act of, 2018; National Defense Authorization Act for Fiscal Year, 2018; Wu and Hetherington, Reference Wu and Hetherington2020). In the annual reports to Congress from 2016 to 2018, the Department of Defense repeatedly mentioned China's threat to the USA and its implications for US interests in the event of a Chinese attack on Taiwan (Office of the Secretary of Defense, 2016, 2017, 2018).

On the other hand, China's economic power may induce public concern over the China–Taiwan tension. China achieved remarkable double-digit GDP growth from 1978 to 2007. In 2010, China surpassed Japan and became the second-largest economy (Barboza, Reference Barboza2010). The size of China's economy allegedly would surpass America's by 2035 and double it by the mid-twenty-first century (Keidel, Reference Keidel2008). While most American corporations see China as an opportunity and hope for increased exports to China after China enters the WTO, Americans' opinions split on how they see China's economic power as more of a threat or an opportunity to the USA, with about 45% on each side (Newport, Reference Newport1997). After China entered the WTO, more Americans saw China as an opportunity rather than a threat. In total, 48% believed that more trade with China would benefit the US economy, whereas 37% saw it as detrimental to the US economy (Tien and Nathan, Reference Tien and Nathan2001). According to a 2016 Gallup survey, ‘by 50% to 37%, more Americans say China rather than the US is the top economy,’ however, ‘more now believe the US economy will overtake China in the future’ (Saad, Reference Saad2016).

Such sentiment may spill over how Americans see the China–Taiwan tension. Taiwan's trade dependence on China's market makes Taiwan more pro-China (Wong and Wu, Reference Wong and Wu2016). It can potentially eclipse the US influence on Taiwan, home to major semiconductor producers crucial to the technology industries in Silicon Valley. However, when the engagement was the focus of US policy on China, the China–Taiwan tension was marginalized to clear the US commercial engagement with China. Based on this argument, I derive the second hypothesis as follows.

5. Research design

5.2 Variables

5.2.1 Measuring the perceived China–Taiwan tension

To verify Americans' perception of the China–Taiwan tension, Pew 2018 asked the question, ‘I am going to read you a list of things that may be problems for the US. For each one, please tell me if you think it is a very serious problem. [Tension between China and Taiwan].’ Answers like ‘Refused’ and ‘Don't know’ are dropped out of the observations. As Figure 1 shows, 7.9% of US respondents think that the China–Taiwan tension is ‘Not a Problem,’ and 19.7% answered ‘Not Too Serious,’ while 47.1% and 25.3% of them respectively consider the tension ‘Somewhat Serious’ and ‘Very Serious.’ Next, I compute a binary variable of whether the perceived China–Taiwan tension is serious, where 0 denotes ‘No’ (combining the 27.6% for ‘Not a Problem’ and ‘Not Too Serious’) and one is ‘Yes’ (72.4% for ‘Somewhat Serious’ and ‘Very Serious’).

Figure 1. US public opinion on the China–Taiwan tension.

5.2.2 Measuring China's threat

I measure China's threat by its level and nature in the US public opinion. To capture the respondents' perceived level of China's threat, Pew 2018 asks, ‘Do you think that China is a major threat, a minor threat, or not a threat to the US?’ where 0 refers to ‘Not a Threat,’ 1 to ‘Minor Threat,’ and 2 to ‘Major Threat.’ Observations with ‘Don't know’ are left out of the analysis. On the left-hand side of Figure 2, the percentage of respondents considering the China–Taiwan tension a serious problem for the USA goes up from 51.8% (Not a Threat), 67.1% (Minor Threat), to 80.6% (Major Threat) as the level of China's threat goes up.

Figure 2. China's threat and the China–Taiwan tension.

As for the nature of China's threat, respondents are presented with the question: ‘Which concerns you more about China?’ I recode the variable such that the value of ‘Economic Strength’ is 0, ‘Military Strength’ is 1, and ‘Both’ is 2. I drop ‘Neither’ (3.1%), ‘Don't know,’ and ‘Refuse.’ In the right-hand graph in Figure 2, higher percentages of respondents take the China–Taiwan tension seriously when they have concerns about China's military strength (77.5%) and both economic and military strength (80.8%) in comparison with those concerned only about economic strength (70.2%) (Figure 2).

5.2.3 Control variables

I use demographic, political, and attitudinal variables as control variables. For the demographics, I include gender, age, and education. First, I code the male as a dichotomous variable, where females are coded as 0 and males as 1. Previous literature indicates that women are marginalized in security issues (Huddy and Terkildsen, Reference Huddy and Terkildsen1993), although some have shown otherwise (Ono and Yamada, Reference Ono and Yamada2018). I do not predict the gender effect on the perceived China–Taiwan tension.

Second, I recode six age groups, with a 10-year gap between the age groups. Age is closely related to the effect of living memories on how Americans see the China–Taiwan tension. Senior Americans have witnessed the US confrontations with the Communist camp, including China during the Cold War and the US engagement policies during China's rise. In contrast, younger Americans have only experienced the latter. Given the differences in historical experiences, the older generation is more likely to have concerns over the China–Taiwan tension than the younger generation. Therefore, I expect age to be positively associated with the likelihood of seeing the China–Taiwan tension as a severe problem for the USA.

Third, I compute a binary variable where 0 refers to ‘No’ for respondents without a college education, and 1 is ‘Yes’ for those without a college education. Respondents with higher education are presumably better at processing detailed information about international affairs and making comprehensive evaluations. Thus, they are more likely to be aware of what the China–Taiwan tension might imply for the well-being of the USA.

As for political controls, I take political identification and political ideology into consideration. First, Party Identification is coded as 1 ‘Republican,’ 2 ‘Democrat,’ and 3 ‘Independent.’ Since Trump was a Republican president, his China policy was considered a more hawkish line, so Republicans were assumed to take the China–Taiwan tension more seriously than Democrats. Meanwhile, Independents have different foreign policy stances. I do not make assumptions about their perceptions on this issue. Second, Political Ideology is measured on a five-point scale from 1 ‘Very Liberal’ to 2 ‘Liberal,’ 3 ‘Moderate,’ 4 ‘Conservative,’ and 5 ‘Very Conservative.’ Previous literature indicates that conservatives focus on domestic affairs and have great reservations about intervening in international affairs, while liberals emphasize otherwise (Fordham, Reference Fordham1998; Desch, Reference Desch2008). Thus, liberals presumably take the China–Taiwan tension seriously, as they essentially are more internationalists than conservatives.

Next, I control for a battery of variables about the most controversial and well-known issues constantly brought up in Trump's rhetoric about China's threat, which entails consequences for how Americans see China–Taiwan relations. The variables include the Loss of US Jobs to China, US Trade Deficit with China, and Cyber Attack from China. I also include President Trump's controversial policy of not pressing China's Policy on Human Rights. Suppose China's threat should remain significant after including these variables in the models. In that case, I can confirm that China's threat is vital for Americans' attention to the China–Taiwan tensions. I will elaborate more on the issue variables as follows.

Loss of US Jobs to China and US Trade Deficit with China. In Trump's framing, China's economic strength was believed to be ‘raping the US with unfair trade policy’ (Zurcher, Reference Zurcher2016) and stealing jobs from the USA (BBC News, 2019). Previous polls show that ‘economic competition from low-wage countries has regularly been treated as a critical threat to the vital interests of the United States’ by a substantial minority. It was mentioned by 38% of respondents in a 2008 CCGA survey, up slightly from 32% in 2006 to 35% in 2004. The figure was 40% in 1998 (Page and Tao, Reference Page and Tao2010b: 25). Given that China accounted for one-third of the total US trade deficit, most Americans believe that ‘reducing our trade deficit with foreign countries’ should be ‘very important’ (Page and Tao, Reference Page and Tao2010b: 22–23). The two control variables are crucial to Americans' evaluation of the China–Taiwan tension. When Americans pay more attention to the loss of US jobs to China and the US trade deficit with China, they are more likely to be wary of the tension between China and Taiwan. The outbreak of war in the Taiwan Strait would drastically impact US economic interests regardless of the US role in the war.

China's Policy on Human Rights. China has been known for its human rights abuse, from organ procurement practice (Paul et al., Reference Paul, Caplan, Shapiro, Els, Allison and Li2017), to religious persecution of minority groups (Potter, Reference Potter2003; Clarke, Reference Clarke2010), and suppression of political dissent (Conrad and Ritter, Reference Conrad and Ritter2019). In 2017, Secretary of State Rex Tillerson publicly condemned China for having the lowest ranking on human trafficking issues (Gramer and Allen-Ebrahimian, Reference Gramer and Allen-Ebrahimian2017). In 2018, acting Secretary of State John Sullivan also released human rights reports stating that ‘China continues to spread the worst features of its authoritarian system, including restrictions on activists, civil society, freedom of expression, and the use of arbitrary surveillance’ (Sullivan, Reference Sullivan2018). A 1997 Gallup poll found that 69% of respondents say that China poorly observes its citizens' human rights (Newport, Reference Newport1997). The figure was about the same in 1997 and 1999 (Tien and Nathan, Reference Tien and Nathan2001: 126). Tien and Nathan found that ‘39 and 67% of respondents believe that human rights issues should prioritize when the United States makes foreign policy toward China’(2001: 127). China's dire human rights record sharply contrasts with a Chinese-speaking democracy, Taiwan, raising the saliency of the China–Taiwan tension in the eyes of the US public.

Cyber Attack from China. Trump accused China of hacking the emails of Democratic officials to intervene in the 2016 presidential election (Reuters, 2017). President Trump's national security team proposed excluding China from building a super-fast 5 G wireless network to counter the threat of Chinese spying (Holland and Schroeder, Reference Holland and Schroeder2018). Based on the original four-point metric in the questionnaire, I re-compute a binary variable where 0 means ‘No’ (‘Not a Problem’ and ‘Not Too Serious’) and 1 means ‘Yes’ (‘Somewhat Serious’ and ‘Very Serious’). Because Taiwan is also a victim of China's cyberattack, Americans may find Taiwan is a valuable ally in fighting against in China's cyberattacks. Thus, the China–Taiwan tension stands out as an issue in their minds. Summary statistics of the variables are presented in Appendix I.

6. Empirical results

This paper examines how the level and nature of China's threat shape Americans' evaluatation of the China–Taiwan tension. Due to the dichotomous nature of the dependent variable, whether the China–Taiwan tension is a serious problem for the USA, I utilize logit models for the statistical analysis. I use robust standard errors at the individual level. I report the results with odds ratios (OR) for a more straightforward interpretation. The full results for logit models are in Table 1.

Table 1. Logit models for the perception of the China–Taiwan tension

Note: Coefficient (coef.) and odds ratio (OR) are presented,robust standard errors in bracketsand constants suppressed.

*P < 0.05, **P < 0.01, ***P < 0.001.

In models 1 and 2 of Table 1, I present the results about the effect of the level and nature of China's threat on whether Americans see the China–Taiwan tension as a serious problem for the USA when the demographic and political variables are under control. The positive coefficients increase the OR of taking the China–Taiwan tension seriously, while the negative ones decrease it. With a positive coefficient at the significance level of 0.001, a one-unit increase in the perceived level of China's threat is associated with a 1.897-unit rise in the log odds of the China–Taiwan tension being considered serious in model 1; that is, the chance Americans think of the China–Taiwan tension a serious issue increases by 189.7%.

In model 2, I examine the effect of the nature of the perceived threat on US public opinion about the China–Taiwan tension. Given that the Nature of China's threat is a categorical variable, I set Economic Strength as the baseline for respective comparisons with Military Strength and Both. Based on a positive coefficient at the 0.05 level, the OR for Military Strength is 1.529, which translates into an increase of 152.9% for the China–Taiwan tension considered serious for the USA. Despite having a positive sign, Both are not significantly different in Economic Strength.

In model 3, I integrate the explanatory variables, level, and nature of China's threat in the model after controlling for demographic and political variables. The results remain robust. Level of China's Threat and Military Strength remain positive and significant, and their log odds respectively translate into a 180.6% and 151.4% increase in the gravity of the China–Taiwan tension in the eyes of Americans. Note that I make the nature of Economic Strength the benchmark to evaluate the effect of Military Strength on the salience of the China–Taiwan tension.

One may suspect that Americans' assessment of the factors involved in US–China relations trickles down to their perception of the China–Taiwan tension. In model 4, I control the four most prominent factors in US–China relations, including Loss of US Jobs to China, US Trade Deficit with China, China's Policy on Human Rights, and Cyber Attack from China. For China's Threat and Military Strength Level, the coefficients remain positive and significant, where the ORs are 144.1% and 179.6%, respectively. Loss of US Jobs to China (204.5%), US Trade Deficit with China (213.8%), China's Policy on Human Rights (298.8%), and Cyber Attack from China (236.3%) carry positive signs, meaning that they significantly increase the chance of the China–Taiwan tension casting a shadow over Americans' minds. The results echo the previous suspicion that the US public opinion on the major factors involved in US–China relations influences how seriously Americans treat the China–Taiwan tension.

As for other controls, Age Groups systematically exerts a positive and significant effect on the China–Taiwan tension being taken seriously across all models. Older respondents in the USA are more cautious about the China–Taiwan tension than the younger ones because of their Cold War memories and awareness of US history in battling communism. In addition, Male and College Education both carry negative coefficients, in line with the prediction. Nevertheless, they are significant in some models but not all four. Moreover, Party Identification and Political Ideology coefficients are insignificant in any four models, while the negative signs fit squarely with the predictions. In sum, the effect of the grave China threat on the military front is proven robust with statistical significance across all four models. Age, widely publicized US disadvantages to China (Loss of US Jobs to China and US Trade Deficit with China), and China's aggressions (Cyber Attack from China and China's Policy on Human Rights) are also proven to be statistically significant factors in shaping Americans' perception of the China–Taiwan tension.

Next, I move on to visualizing the main findings in Figure 3. I mark the point estimates with a circle, diamond, or cross-references to represent the predicted probability of respondents considering the China–Taiwan tension a serious problem for the USA. The two bars stretching from the point estimates are the 95% confidence intervals. Based on logit regression models with robust standard errors, the left graph in Figure 3 shows the predicted probabilities of the China–Taiwan tension being a serious problem for the USA on the condition that China is a significant threat and a military threat to the USA in the eyes of Americans. I find that the point estimate of Military Threat is the highest among the three natures of China's threat across all three levels of China's threat. As the level of China's threat escalates, the point estimates for the three natures of China's threat systematically rise.

Figure 3. Predicted probabilities and the differentials for the China–Taiwan tension being a serious problem for the USA.

Note: Economic strength is set as the baseline for the right-hand graph.

For more intuitive comparisons of the effect of China's threat on Americans' evaluation of the China–Taiwan tension, I use Economic Threat as the baseline and present the differences in the predicted probabilities for the nature of China's threat across the three levels of China's threat. If the 95% confidence interval bar crosses the red dotted line of zero, the predicted probability is not statistically distinguishable from zero. As the right graph in Figure 3 shows, both point estimates are no different from those of Economic Strength across the three levels of China's threat, as the three bars of the 95% confidence intervals touch on the red dotted line. However, the point estimates of Military Strength are significantly different from those of Economic Strength regardless of China's threat level when the 95% confidence intervals bars do not cross the red dotted line. Judging from Figure 3, I conclude that the concerns over China's military draw more Americans' attention to the destructive consequences of the China–Taiwan tension to the USA than the concerns over China's economic power. Intuitively, concerns over military and economic threats should make more respondents worry about the China–Taiwan tensions. A possible explanation is that economic concerns may carry more weight in the pooling of military and economic concerns, such that the effect of the pooled concerns is not significantly different from the sheer concern over economic influence. An increasingly strong Chinese economy eventually fosters more excellent military capability. Economic strength may be detrimental to US interests. However, it is ultimately China's military capacity rather than China's economic power that induces Americans' concerns over the China–Taiwan tension.

7. Additional analysis

The statistical results may depend on the model choice, which is conditional on the nature of the dependent variable. To rule out the possibility of model dependency, I perform a robustness check by computing the dependent variable as an ordered variable on a four-point scale, from ‘Not a Problem’ (0), to ‘Not Too Serious’ (1), ‘Somewhat Serious’ (2), and ‘Very Serious’ (3). I then re-run the ordered logit models. Presented in Appendix II, the results corroborate the findings of the logit model that a greater level of China's threat and military-related China threat boost the percentage of Americans considering the China–Taiwan tension to be serious to the USA.