Getting tough with the dragon? The comparative correlates of foreign policy attitudes toward China in the United States and UK

A large body of research suggests mass publics are capable of thinking coherently about international relations. We extend this body of research to show that domain relevant postures – in our case, more abstract beliefs about foreign policy – are related to how tough of a line representative samples of US and UK respondents want their governments to take toward China. More specifically, we utilize a unique comparative survey of American and British foreign policy attitudes to show broad support for toughness toward China. Beliefs about the use of the military and attitudes regarding globalization help explain preferences for tough economic and military policies toward China. In the two countries, the relationship between general foreign policy outlooks and the positions International Relations of the Asia-Pacific Vol. 17 No. 2 # The Author 2016. Published by Oxford University Press in association with the Japan Association of International Relations This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. International Relations of the Asia-Pacific Volume 17, (2017) 265–299 doi: 10.1093/irap/lcw008 Advance Access published on 20 June 2016

citizens take is robust to the addition of a general mediator that controls for the general affect those surveyed have toward China.Finally, the strength of the relationship between these abstract postures and specific preferences for a China policy are different across the countries.

Introduction
Over the past three decades, political scientists gathered a steady stream of evidence to show that citizens in multiple countries have well-structured foreign policy attitudes (e.g.Hurwitz and Peffley, 1987 [US]; Hurwitz et al., 1993 [Costa Rica]; Chittick et al., 1995 [US]; Richman et al., 1997 [US]; Bjereld and Ekengren, 1999 [Sweden]; Munton and Keating, 2001 [Canada]; Jenkins-Smith et al., 2004 [UK]; No€ el et al., 2004 [Canada]; Reifler et al., 2011 [UK]).Moreover, research shows that citizen attitudes respond to changing international and domestic circumstances in understandable and reasonable ways (Shapiro and Page, 1988;Peffley and Hurwitz, 1992;Wlezien, 1995;Gelpi, 2010;Kertzer, 2013; though see Baum and Groeling, 2010 for a more nuanced and pessimistic argument).In this article, we build on this existing research to show that these structured abstract foreign policy beliefs (or 'postures') help explain American and British policy preferences toward China.Specifically, we utilize hierarchical constraint models (e.g.Hurwitz and Peffley, 1987) to show that specific attitudes about China policy flow from foreign policy postures.We find that both security and economic postures are relevant in explaining support for toughness toward China.
There are important differences in the salience of China as an issue when the US and UK are compared to one another. 1In the United 1 By way of illustration, we utilized the Factiva database and conducted searches for articles with the terms 'China' and either 'Military' or 'Economy' that appeared in The New York Times (NYT), The Telegraph, The Guardian, and The Daily Mail over the course of the two weeks prior to the surveys entering the field (16 January-31 January, 2013).On the military aspect, coverage in the NYT is much different from what appears in the three large circulation (but ideologically distinct) British papers.In the former, multiple articles focus on the potential for military confrontations between the United States and China as a result of China's territorial dispute with Japan over islands in the East China Sea.Other coverage focuses on President Obama's interest in building up US military strength in Asia to counter the growth of the Chinese military and the potential for China to engage in cyber-attacks on US government and businesses.In the British papers, there was coverage of the Chinese-Japanese territorial dispute and the potential for a US-China military confrontation, but little mention of British military interests or involvement.
States, issues involving engagement with China -from economic and trade competition to the possibilities of military tension -frequently appear in the news.In the UK, less news coverage is devoted to China overall and in particular, there is less news reporting dedicated to the prospect of economic or military conflict between China and Great Britain.This article exploits this difference by comparing the influence of foreign policy postures on China attitudes across these contexts.We find foreign policy beliefs of Americans to be more closely linked to the level of toughness they demand of their Government's China policies.However, this is not to say that foreign policy postures are irrelevant for British attitudes toward China; postures are significant predictors of China preferences in the UK as well.We also find that the relevance of the postures is robust to adding a 'likability heuristic' or feeling thermometer toward China to the model -that is these broad beliefs have a direct effect on preferences for getting tough in two policy domains even after controlling for how these beliefs influence affect toward China.
American attitudes toward China feature important shifts over time.Americans view China as a primary threat in the 1950s and 1960s, but views soften somewhat in the 1970s (Kusnitz, 1984).However, there is important variation underneath these more gradual shifts.One example comes from the Truman Administration; Republicans were much more critical of Sino-American policies, likely attributable to the view that the Democratic President was responsible for 'losing China' [and Korea] to the Communists (Wittkopf, 1990).At other points in time, public opinion appears to have played an important role in the policymaking calculus of American leaders.Foyle (1997) details how President Eisenhower and Secretary of State John Foster Dulles turn against unilateral action in the defense of Taiwan from mainland Communist China in the early 1950s because of a perception of a divide in domestic public opinion.Similarly, Rusk and Papp (1990) notes Kennedy's caution in his engagement with the China issue and attributes this reticence to a fear of negative public reaction toward proposed White House policies.
While Eisenhower and Kennedy appear to allow concerns over public opinion to constrain action vis-a-vis China, Steele (1966) sees a public open to a more flexible and engaged policy during thaws in the Cold War.By the 1970s, the initiatives of Nixon and Carter help to improve both relations between the two countries and Americans' views of China, but attitudes sour considerably after the violent suppression of student protestors in Tiananmen Square.Ultimately, Holsti (2004) describes the American public's reactions to China and the policies they want their state to pursue as 'events driven' and 'rational'.Citizens cool when China engages in policies that threaten the United States, but warm when elites make overtures toward working with one another.Page and Bouton (2006) observe that by the early 2000s Americans overwhelmingly favor maintaining diplomatic relations with China and believe that the United States has a vital interest in engagement with China.In particular, the authors note a shift in American priorities toward matters of trade and economic commitments.Page and Xie (2010) show that Americans see the benefits China can provide in terms of the availability of cheap imports, but they also believe that the Chinese do not play fair when trading with the United States.On the military front, a slim majority of Americans want to work hard to contain Chinese military power, but only a minority are willing to see the United States mobilize ground forces against China's large military if the latter invades Taiwan (Page and Bouton, 2006;Page and Xie, 2010).

Britain and China
Measurements of British attitudes toward China are sparse and infrequent.Early work by Younger (1955) speculates that public opinion is not as relevant in the UK as it is in the United States in shaping elite actions when dealing with Communist China.With some limited exceptions (e.g.Hong Kong), the British public's attitudes toward China rarely appear as a topic of conversation in media or academic journals (Hoge, 1997).In contrast, British elites do see the rise of China as a 'driver of change' that will affect the UK in the coming decades.There are some conjectures that European public opinions toward China can harden if trade moves away from the import of low cost Chinese basic goods into the European Union and to flooding British and EU markets with high quality goods that rival those produced by domestic industries (Jacques, 2009).
British policymakers and their European counterparts also treat the rise of China as a military power with less suspicion than their American allies.Like the United States, the EU maintains an arms embargo on China, but opinions as to the appropriateness of maintaining bans on weapons sales is a topic more open to debate among European policymakers (Dai, 2009).The rise of China simply is less salient in Europe than it is in the United States, and elites are less likely at this point in time to view China as an imminent threat.However, this may change if more attention is paid to the country by elites, if trade wars heat up, or human rights practices by China catch the attention of activists in UK and across Europe.
Compared with the British, Americans hear more about China's status as a growing superpower in the media and the potential challenges this poses to US dominance (Peng, 2004;Stone and Xiao, 2007).Studies also find that the BBC's coverage of China is robust but does not focus on the impact of China's rise on the UK (Seib and Powers, 2010).This difference and the lack of worry by policymakers suggest that British attitudes concerning their country's China policy will not be as well formulated and that predictors of the public's attitudes in the UK likely will be weak and more difficult to establish.

Sources of specific attitudes toward China
Our primary interest in this article is to understand better American and British attitudes toward China.By attitudes we do not just mean how those residing in these two western states view China generally, but rather the toughness of the policies they want their country to pursue in two policy realms most relevant to bilateral and multilateral relations -trade and military engagement.We demonstrate below that UK and US citizens possess abstract foreign policy attitudes that retain the same structure in both states.A direct linkage between abstract policy views and demands for toughness toward China in two different policy spheres is suggestive of higher order thinking both about foreign policy, in general, and the stance people believe their country should take toward China, specifically.
Suggestive evidence of an even higher level of thinking on the part of citizens occurs if the postures have different relationships to each of the citizen demands for tough military and trade policies.As there are multiple substantive factors, partisan differences, and standard control variables that can be related to attitudes toward China, we first elaborate and justify the independent variables to appear in the multivariate models below.

Our key variables of theoretical interest -foreign policy postures
A cottage industry of research shows that American attitudes about foreign policy have a coherent multi-dimensional structure.The exact number and content of dimensions is a function of the available data and the modelling strategy.Nonetheless, a dominant theme in the work is that peoples' beliefs about the strength and proper use of their country's military (militarism) is separate from attitudes toward the appropriateness of employing instruments of soft power such as the dispersal of foreign aid (liberal internationalism) (e.g.Wittkopf, 1986Wittkopf, , 1990).2Although less well studied, these dimensions appear to hold approximately among the UK public as well (e.g.Reifler et al., 2011).The most relevant forerunner for our present analysis is the seminal article by Hurwitz and Peffley (1987) on belief systems and foreign affairs.Dimensions extracted via factor analysis reveal postures or a middle ground between an individual's core beliefs and their positions on specific policy issues.We emulate many features of Hurwitz and Peffley's analysis -we include variables that tap the three postures of isolationism, militarism, patriotism (or ethnocentrism as Hurwitz and Peffley call it), and include in the models controls for economic evaluations and partisan identification.Within this hierarchical framework, we make a number of specific predictions about sources of support for toughness toward China.
We first turn to isolationism.Isolationists want to avoid unnecessary entanglements beyond state borders.This should lead those high in this sentiment to shy away from wishing their government to pursue policies other states may view as provocative.Consequently, we expect: H1: Higher levels of isolationism present among respondents from the United States and Great Britain co-varies with a reduced propensity to support the government pursuing tough military and economic policies toward China.
An isolationism dimension taps respondents' attitudes for how they want their state to interact with other countries across a variety of issues.In contrast, a militarism dimension captures citizens' beliefs that their country's military should be strong and ready for action if the country's interest requires its use -either to keep it secure from attack or project its values and power (Page and Bouton, 2006).A citizenry that thinks more deeply about engagement with China is prone to utilize different dimensions when responding to different aspects of their state's China policy.If this higher level of awareness holds, empirical estimations should establish that: H2: Respondents' positions on the militarism dimension are more closely tied to their preferences for a tough state military policy toward China than they are for a tough economic policy toward China.Hurwitz and Peffley (1987) treat preferences for free trade as a specific issue, but we contend that attitudes toward globalization are properly thought of as a midlevel posture.As is the case with the militarism posture, we consider evidence of higher ordered citizen thinking about their state's China policy to occur if: H3: Respondent support for globalization has a significant and negative association with their preferences for a tough state economic policy toward China.The magnitude of this linkage is larger than the relationship between globalization and preferences for a tough military policy.
The three latent variables, whose structure and measurement we describe below, differ in that isolationism is expected to co-vary significantly with preferences for both a strong economic and military policy.If citizens think deeply about how their country should react to the multifaceted rise of China, preferences for globalization should be more closely linked to attitudes about the level of toughness of their country's economic policy while preferences for militarism should have a stronger association with demands for a tough military policy toward China.

Partisanship
We have divergent cross-country expectations about the role partisanship should play as a 'perceptual screen' (cf.Lewis-Beck et al., 2008) when it comes to shaping citizen preferences toward their country's international relations with China.In contemporary America, it seems that on nearly all issues there are noticeable differences in the policy opinions of Republicans and Democrats (Jacobson, 2010).Consequently, we expect this also to apply to the issue of China and we will observe differences between Republicans and Democrats on preferences for toughness, even after controlling for other variables.
This expectation requires further explanation as there is a burgeoning debate in the literature concerning the role of partisanship in influencing American attitudes toward China.Some argue that partisanship (and ideology) primarily has an indirect effect on foreign policy preferences by affecting perceptions of threat and more abstract foreign policy goals (e.g.Page and Bouton, 2006 for a general argument, and Page and Xie, 2010 for a more China specific argument).Peter Gries (2011Gries ( , 2014a,b) ,b) critiques this approach arguing that partisanship and ideology play an important role in shaping Americans' attitudes far more directly.In the case of China, the more direct effects of partisanship are rooted in a greater antipathy toward communism.Our expectation is that for the United States, partisanship will reach statistical significance.Of course, a significant partisanship variable in the models we present does not necessarily validate or repudiate one or the other of these competing theoretical accounts.However, given the ubiquity of partisanship in explanations of American political opinions and behavior, we tend to favor an interpretation where partisanship has at least some direct influence on the content of both general foreign policy preferences and China attitudes more specifically.
In UK, research suggests that there are partisan differences in foreign policy beliefs -those more supportive of the use of force have a higher affect toward the Conservative Party and its leader, while those who fit the mold of 'cooperative internationalists' feel warmly toward the Liberal Democrats (Reifler et al., 2011).In practice, however, British parties have a habit of not differentiating themselves in a clear and meaningful way on foreign policy.Even on the salient matters of military interventions in Libya and Afghanistan, the similarity in party positions limits the ability to observe elite influence (Reifler et al., 2014).Consequently, we expect that partisan attachments will have little impact on British preferences for toughness toward China.In summary: H4: The long term attachments of Americans to a political party, as represented by their responses to a standard partisan identification question, will significantly influence the degree of toughness they seek from their country's China policies-Democratic respondents will favor a softer approach while Republicans will demand toughness.In Great Britain, respondents' partisan affiliations will have little direct effect on attitudes toward the approach the UK government should take toward China.

Economic perceptions
A fundamental variable motivating political behavior and voting is citizens' perceptions of the state of the economy (e.g.van der Brug et al., 2007;Duch and Stevenson, 2008).Page and Bouton (2006) note that Americans pay attention to their own economic interests when thinking about trade and globalization, and this leads to our hypothesis that: H5: Economic evaluations affect support for toughness on China.Those who have more negative evaluations of the economy or household finances will support greater toughness toward China.

Ethnocentrism/patriotism
Classic work in social psychology holds that foreign countries are socially acceptable targets for the projection and displacement of anxieties (Silverstein, 1989).Hurwitz and Peffley (1987) label the belief in national superiority 'ethnocentrism' and place it as a core value that predicts positions on postures and only indirectly influences specific foreign policy issues.Given a wider range of indicators available to us, we examine the link between blind patriotism (cf.Schatz et al., 1999;Huddy and Khatib, 2007) and preferences for tougher China policies.Like ethnocentrism, the measure of patriotism we describe below taps agreement with a belief in the superiority of one's country.However, our measure also incorporates beliefs about the appropriateness of blindly following the government's policies.We expect that: H6: There is a direct and positive relationship between respondents' agreement that citizens should blindly support US/UK policies and those who want tougher policies toward China.

The likeability heuristic
Recent work by Peter Gries (2014b, p. 43) asserts that 'gut feelings toward foreign countries serve as a vital mediator between ideological predispositions on the one hand, and specific foreign policy preferences on the other.'Our interpretation of this position is that the immediate antecedent to preferences for toughness vis-a-vis China is how warmly one feels toward China, and that other variables influence one's warmth toward China.If this approach is correct, then once a feeling thermometer is added to the model as a mediator, much of the direct relationships we observe between postures, partisanship, patriotism, and economic evaluations should weaken significantly.We depart slightly from Peter Gries (2014b) and examine whether mediation via a China feeling thermometer occurs for five key variables in our primary analysis.
H7: Respondents' support for a tough China policy in both the economic and military policy realms operates via general affect toward China.
The greater salience of China in the United States compared with Great Britain leads us to cross-country expectations concerning the rival hypotheses presented above: H8: The covariates described in this section will have better explanatory power in explaining citizen preferences for tough China policies in the United States than in Great Britain.
H9: Americans will utilize different postures to inform their foreign policy preferences toward China in the economic and military domains while the influence of the postures will be more even across the two policies among the less informed British respondents.
4 Data, variables, and methods

Data and dependent variables
In February 2013, 1,891 British and 1,858 American respondents completed the third wave of a YouGov online panel survey explicitly fielded for the purposes of better understanding the foreign policy attitudes of both publics. 3Two outcome variables come from this wave of the survey.These are (i) a question asking respondents 'Thinking about US/ UK policy toward China, do you think it is important to be tough with China on economic and trade issues?';and (ii) a question asking respondents 'Thinking about US/UK policy toward China, do you think it is important to be tough with China on military issues?'Available responses to these two questions falls along a four point ordinal scale ranging from 'No, not at all important' to 'Yes, very important', and the question is similar to one employed by the Pew Research Centre in a recent China focused survey. 4In the last set of multivariate analyses below, we employ a 101 point 'feeling thermometer' as a mediator.This probe provides respondents with a scale ranging from 0 to 100, with 0 labeled 'very coldly' and 100 labeled 'very warmly', and 3 YouGov's methodology for ensuring the samples are reasonably representative of target populations differs slightly across the two countries: In the United States, respondents to the YouGov survey are drawn by matched quota sampling from a non-probability sample of over 1 million volunteer panelists who take the surveys in exchange for prizes or cash.
The methodology for YouGov's procedures is described in Rivers (2007) and Ansolabehere and Schaffner (2014) report the high comparability of estimations utilizing this method to those obtained utilizing standard nationally representative RDD telephone interviewing.In the UK, YouGov builds a sample by drawing from its respondent panel of over 350,000 Britons who have signed up to take surveys in exchange for entries into prize draws or points on an account that eventually can be redeemed for a cash prize.The methodology for YouGov's UK procedures is reported by Sanders et al. (2007) to yield multivariate results that are highly comparable to estimates obtained utilizing responses from probability sampling with face-to-face interviewing.To obtain a sample that approximately is representative of Great Britain (Northern Ireland is excluded from the sampling frame), specific surveys are opened to a subset of the panel that is representative of the known population in terms of age, gender, class, and newspaper readership in Britain.In the multivariate analyses, post-sampling weights are employed to bring the sample further into line with the population's known characteristics as derived from either the national census or a large scale probability sample (e.g. the Current Population Survey in the United States or the National Readership Survey in the UK).Details on the observed and weighted characteristics of the US and UK samples are presented in the Online Appendix, but the big difference between the observed and weighted demographic characteristics of the panel lie with the underestimation of youth who participated in both samples.Attrition from this demographic mimics what is observed in probability based panel studies as is the underrepresentation of minorities and those of lower socio-economic status (see Watson andWooden, 2009 andLipps, 2009).
asks them to evaluate 'China' in general -not the Government, Chinese people, or in any other context.

Independent variables
Hypothesized covariates for the multivariate models derive from the interviewees' responses to questions asked on November 2011 and May/ June 2012 waves of the survey.To improve the reliability and validity of key independent variables of interest, we hypothesize multiple survey indicators to be reflective of latent factors of each concept.Below, we employ Confirmatory Factor Analysis (CFA) to ensure that the factor structure is valid; these analyses employ a robust weighted least squares estimator to enable us to derive the latent variables by appropriately treating responses to the agree-disagree survey questions as ordinal.A brief description of the indicators follows, but for the sake of brevity, we relegate full question wording of the items, the wave each item appears on the survey and response distributions to the online Supplementary Appendix.

Latent variables
i. Militarism Posture: A latent variable where higher scores motivate positive responses to questions asking whether respondents think military spending should be increased and agreement with a question as to whether the respondent believes that their country should be willing to use force in the face of expansionist aggression by another state.ii.Isolationism Posture: A latent variable where those scoring highly on the dimension are prone to agree with survey questions asking whether their country should avoid involvement with other countries, whether getting involved in other states risks the wellbeing of domestic citizens, and whether their country should simply mind its own business in its conduct of foreign policy.iii.Globalization Posture: A latent variable where those scoring highly have a propensity to believe increases in trade between countries has a positive outcome on the respondent and their families, domestic factory workers, the domestic economy writ large, domestic consumers, and domestic businesses.iv.Patriotism: Respondents high on the negative form of patriotism likely agree that people who do not back the state's policies should live elsewhere, people should support policies simply because they are the policies of the country, people should avoid saying bad things about the country, their state is always correct in matters of international affairs, and that the country has the moral high ground in policymaking.v. Economic Performance Perceptions: Respondents scoring high on this factor tend to agree to four survey questions asking whether their own and the country's economy performed better in the past year and would do so in the next year.

Methods
Conducting valid cross-cultural multivariate analyses to compare American and British attitudes toward China first requires determining whether the factor structure of the hypothesized latent covariates is similar enough to argue that their meaning is roughly equivalent across the two states (see Davidov et al., 2011).Multiple-group analysis in a CFA framework is the method used to establish this validity (Byrne, 2012).The second step is, for each country, to estimate simultaneously via ordered probit the two toughness variables on the factors, partisan identification and control variables.One of our key interests is the degree to which respondents bring different judgments to bear on getting tough with China on economic and military matters.To examine this question, we tested whether the co-variation between independent and dependent variables are statistically equivalent to one another across the two policy spaces.To more easily see when and where there is (and is not) a discernible difference in how independent variables are associated with the two dependent variables we use equality constraints.If an equality constraint is warranted (that is, the independent variable has the same effect on economic toughness as it does on military toughness), the table reports a single coefficient in a country.If the independent variable has an unequal effect on the economic and military dependent variables, the table reports two coefficients.By doing this, we can more clearly see which variables affect support for economic and military toughness differently, and where a variable is just generally associated with greater toughness toward China. 5To foreshadow our results, most of the time we can impose an equality constraint -that is the independent variable in question affects economic and military toughness equally (this is true on all variables for the UK sample, and all but three variables for the US sample).
The final step is to determine the degree to which affect toward China, as represented by a feeling thermometer, mediates the relationship between postures and policy.If citizen attitudes about the policies the US and UK governments should pursue toward China are mostly caused by general affect toward China, the addition of the feeling thermometer as a predictor should render the coefficients for the direct paths between the threat variables and substantive variables insignificant.

Evaluations of China
The February 2013 surveys ask British and American respondents to evaluate a number of countries on the 101 point feeling thermometer, and the average score those in the survey give to each country appears 5 More specifically, we are using seemingly unrelated ordered probit (SOP) with equality constraints.In other words, we are simultaneously running two ordered probit models and testing the equivalence of slopes across these two models.SOP models also allow the error terms of the two ordinal outcome variables to covary.Efficiency gains are obtained in simultaneous estimation (Kennedy, 2003).
in Figure 1.Citizens of both states feel cool toward China.Not surprisingly, they feel warmest toward their own country and generally feel more warm than cool toward Canada.The European countries in the feeling thermometer battery receive relatively high scores.At the opposite end of the spectrum lay two Arab states in the midst of turmoil, Egypt and Libya.Americans rate China lower than Egypt (t ¼ À2.60, P < 0.01) while the British are more generous toward the former than the latter but still cool (t ¼ 9.09, P < 0.001).In short, most peoples of both states are not Sinophiles.
Respondents from both countries also desire their politicians to pursue a 'tough' China policy.Table 1 provides distributions for the two main dependent variables of interest.Comparing across the two countries, the key difference is that Americans are prone to believe that tough China policies in the economic and military arenas are 'very' instead of 'somewhat' important.British responses cluster in the latter category, and they are slightly more likely than Americans to declare that tough China policies are 'not too important'.
Britons and Americans are slightly more supportive of tough economic and trade policies toward China than they are of placing importance on their country taking tough military stances against this state.However, the take away point is that support for tough policies pervades both policy realms.Although the response distributions are significantly different from one another in both countries, the polychoric correlations are 0.74 in the UK and 0.79 in the United States.

Factor structure
Before multivariate analyses to test the above hypotheses commence, it is imperative to obtain scores of the latent variables that are valid within countries and functionally equivalent across states.Table 2 presents the results of a CFA where the unstandardized factor loadings and thresholds on the ordinal responses to the survey questions are constrained to be equal across countries.As is evident from the Table, the standardized factor loadings are strong, and the fit of the model, while not exact ðv 2 WLSMV ¼ 1; 541:214; P < 0:001Þ, approximately matches the data structure (RMSEA ¼ 0.04) even when equality   constraints are placed on the loadings. 6This approximate fit suggests that the 'meanings' of the latent variables -the three postures, the measure of patriotism/ethnocentrism, and economic evaluations, are the same across the two states and valid cross-country comparisons are possible.
Differences in latent means tests suggest that Britons are more militaristic (0.241), believe that globalization has a positive impact (0.244), and, surprisingly, more patriotic (0.207) than Americans.Circa early 2013, British economic outlooks, however, are much more negative (À0.735). 7

Direct relationship
For each country, the two toughness variables simultaneously are regressed on the factors, whose valid structure derives from Table 2, as well as partisanship and socio-demographic controls.In a first step, the estimations of the magnitudes of the coefficients for the latent and observed covariates can vary across the two outcome variables measuring the importance respondents place on their state pursuing tough economic and military policies toward China.One by one, equality constraints on coefficients and v 2 WLSMV difference tests determine whether the fit of the model declines when the assumption that the independent variables have an equal effect on each of the outcome variables is in force.Table 3 (Byrne, 2012).
7 Factor scores have a range from À2.257 to 2.304.American latent variable means are fixed to zero to allow for comparisons.
8 Interpret the standardized continuous independent covariates such that a one standard deviation unit change in the covariate associates with a slope change in the underlying latent response to the categorical toughness variables.For the binary independent variables, a move from 0 to 1 corresponds to a slope change in the underlying latent response to the categorical toughness variables.Coefficients obtained via the ordered probit WLSMV estimator in Mplus 7.3.In UK, constraining all coefficients to be equal across the two simultaneous ordered probit estimations results in a model that fits the data no worse than does an output where the magnitude of any of the coefficients vary.This suggests that the impact of each of the covariates on the degree of toughness respondents prefer in the economic and trade and military policy realms is equivalent.We discuss the implications of this finding in the section below.
Militarists in the UK want the Government to pursue tougher China policies, and, although support for globalization and trade co-varies with favoring softer China policies, the militarism latent variable is the stronger predictor.The most relevant substantive variable, however, is the respondents' positions on the general isolationism dimension.Those skeptical of foreign entanglements do not want the British Government to pursue an aggressive China policy.Finally, those scoring high on patriotism find a tougher policy more palatable.In terms of the control variables, adherents to the Church of England and Catholic faiths as well as those practicing non-Christian religions are more favorable to a tougher stance toward China than those who do not belong to a church.University educated respondents are less bellicose than those with less than a bachelor's degree and men want a less confrontational policy toward China both in the military and economic and trade arenas. 9Partisanship is not a significant predictor of the level of toughness the survey respondents' desire.Despite their significance in a number of instances, the explanatory power of the predictor variables is modest, and the R 2 of each of the latent variables underlying the ordinal toughness scales are just over 0.10.
The American story has important nuanced differences.The magnitudes of a majority of the coefficients are equivalent across the 9 The large gender gaps suggesting British women to be significantly more likely to want to follow a tough China policy are curious and a potential area for future research.On the one hand, much of the feminist literature argues that women are more pacifistic than men, but studies of foreign policy attitude qualify this assertion.In her study of Denmark, Togeby (1994) notes that gender gaps in foreign policy attitudes only emerge in the presence of left wing and feminist mobilization.In their comparative study of feminist identities, Hayes et al. (2000) note that British women were no more likely than men to support feminist issues and were significantly less likely than American women to be classified as feminists according to the responses they provided on the cross-national World Values Studies.
In short, a potential reason for the above finding that women in Great Britain actually are more bellicose toward China might be the antipathy toward feminism observed among British women, but this is speculative.
economic and trade and military toughness estimations.Hawkish beliefs have the strongest impact on the demand for toughness in both.However, the impact of believing globalization is a net positive has a far stronger impact on the desire for a less tough US economic policy toward China.Favoring globalization remains a significant predictor of a less aggressive military policy, but the strength of this latent variable on this estimation is muted.Somewhat surprisingly, those high on patriotism are slightly less inclined to see an aggressive economic policy toward China, but the impact of this latent predictor on military policy is insignificant.Finally, confidence in one's own and America's economic condition suggests a desire for tough policies toward China across multiple policy realms.10Unlike the results for the UK, older Americans are far more supportive than the young of taking a tough line toward China, and the impact is the same across policy domains.Members of all minority groups want to see the US pursue softer policies toward China, and unlike the UK, there appear to be partisan differences in policy demands between Republicans and Democrats.The former see tough policies as more important while the latter favor less aggressive stances.Protestants and Catholics also want tougher US policies toward China.
A key difference between the estimations across the two samples is the fact that the covariates included in the American estimations are able to explain much more variation in the desire Americans have for tough policies toward China, both militarily and economically.

Thermometer as a mediator
Given the similarity of the importance the British and American public's place on the toughness of their governments' China policies across disparate policy domains, it is a fair question to ask whether the responses people give to the two outcome variables of interest simply are reflective of general attitudes toward China.To partially test this, we regress the two threat variables onto a question asking respondents to evaluate China on a 0-100 thermometer scale and the above covariates.11To capture indirect effects, the feeling thermometer additionally is regressed on the covariates.
Table 4 reports the results of the estimations of the two toughness variables on the latent covariates in the presence of the feeling thermometer acting as a mediator.As one would expect, some of the effects of our variables are mediated by our measure of affect toward China.Nonetheless, our variables continue to exert direct effects on support for toughness toward China.
Results from the estimation in Table 4 show that only a small proportion of the effect of Militarism operates indirectly. 12In contrast, a much larger proportion of the effect of globalization is mediated by the China feeling thermometer.Interestingly, in the US estimation, there is not a significant direct effect from beliefs about globalization on desires for a tough military policy, but the direct effect is significant and quite strong when it comes down to preferences for a tough economic policy.
Another interesting result emerges concerning perceptions of the economy -the direct and indirect effects run in opposite directions.Even after accounting for standard demographic controls that capture household income, those confident about the economy feel warmer toward China, which depresses support for toughness.But positive economic evaluations maintain a significant and substantively large direct effect for support for getting tough with China.In other words, more positive economic outlooks appear to coincide with a more outward and aggressive looking foreign policy -but this is tempered by the more positive feelings toward China that also come with more upbeat economic assessments.
The direct and indirect effects of partisanship in the United States are not described in the table but warrant attention.The divide between Republicans and Democrats exists because the former believe it is important for the US government to pursue tougher economic and military policies toward China.Democratic partisanship does not directly affect the China policy positions -the mediation analysis suggests the significant coefficients for Democratic partisanship on display in Table 3 are an artifact of Democrats feeling warmer toward China as a whole.Likewise, the racial disparities in policy preferences we see in the United States in Table 3 are likely a function of the fact that minorities feel warmer toward China rather than any differences in their foreign policy preferences.The effect of age, however, is mostly direct.Older people in the United States place greater importance on the government pursuing tougher policies toward China.
In the UK most of the effect of gender on policies is direct -men want a more friendly approach toward China.Further, the demands for a tougher approach toward China among religious adherents are not filtered through the thermometer.The role of partisanship remains insignificant.

Discussion
This article constitutes a unique opportunity to explore the specific foreign policy attitudes of Americans and Britons in direct, side-by-side, comparative way.The distributions of our key outcome variables, citizen preferences for tough state China policies in the economic and military spheres as well as a chosen mediator -general citizen affect for China -demonstrates a great deal of concern about China's rise as a global power.Britons and Americans are wont to favor tough policies, the latter more so.Those on both sides of the 'Special Relationship' view China coolly in comparison to their western allies.With this information in mind, we construct a set of rival hypotheses to better understand the root covariates of this skepticism.In short, we ask whether broad and abstract foreign policy postures, coupled with variables standard in models of political choice (economic evaluations and partisanship), co-vary with the specific attitudes British and American citizens have about the level of policy toughness their country should exhibit in their economic and military policies for China.A summary of our hypotheses and findings appears in Table 5.
In both states there is support for Hurwitz and Peffley's (1987) core argument that abstract foreign policy postures can explain variation in the more specific aspects of foreign policy citizens want their country to pursue.A key posture in the hierarchical model, ours and theirs, militarism, associates strongly with British and American demands for a tough response to the rise of China.Another, isolationism, makes Britons shy away from preferences for getting tough with China, likely because they view such actions as needlessly provocative.Support for

Hypothesis Finding
H1: Higher levels of isolationism present among respondents from the United States and Great Britain co-varies with a reduced propensity to support the government pursing tough military and economic policies toward China.

Supported in the UK
H2: Respondents' positions on the militarism dimension are more closely tied to their preferences for a tough state military policy toward China than they are for a tough economic policy.
Not supported.
H3: Respondent support for globalization has a significant and negative association with their preferences for a tough state economic policy toward China.The magnitude of this linkage is larger than the relationship between globalization and preferences for a tough military policy.
Fully supported in the US.Significant but weak relationship that is equal across policy domains in the UK H4: The long term attachments of Americans to a political party, as represented by their responses to a standard partisan identification question, will significantly influence the degree of toughness they seek from their country's China policies-Democratic respondents will favor a softer approach while Republicans will demand toughness.In Great Britain, respondents' partisan affiliations will have little direct effect on attitudes toward the approach the UK government should take toward China.
Fully supported in the UK.Mostly supported in the US. and the belief that the United States should not be tough toward China in the economic and trade policy realm. 13 Results from testing rival hypotheses and making cross-country comparisons also are indirectly suggestive of the level of sophistication citizens have when answering questions about specific policies.Postures can be informative shortcuts, but only if the postures are wellformulated and citizens can make the association between them and the specific policy under scrutiny.Results from Table 2 suggest citizens in both countries have clearly formulated abstract beliefs about foreign policy that are cross-culturally valid.That the postures in both states have a direct relationship with the toughness variables without a need for mediation is evidence of higher order thinking.But we would be remiss to not point out that the relationship appears stronger in the United States.Unlike Britons, American respondents also are more likely to, albeit in limited circumstances, utilize different postures for different policy aims.Although foreign policy often is remote from the day-to-day lives of citizens in both countries the coverage and consequences of China's rise is more apparent for Americans.Thus, it does not surprise that the postures and other independent variables do a better job at explaining American responses than they do in the UK and that, unlike their British counterparts, interviewees from the United States utilize different postures in coming to conclusions about economic and military policies toward China.

Conclusions
This article makes several primary contributions to the literature on public opinion on matters of foreign policy.First, it explores the important area of attitudes toward China in two key Western states.Second, the article moves beyond the American case to explore the similarities and differences in opinion formation on matters of foreign policy among the British and American publics in the sources of their attitudes toward China.Third, this article shows that foreign policy variables continue to play a powerful role in shaping China attitudes 13 One potential explanation is that uncritical patriots tend to be those who believe their country can do no wrong.If one is hesitant to criticize one's country, one might also be reticent to believe that their state's policies were too 'soft'.
even after controlling for simpler heuristics such as country feeling thermometers, partisanship, and economic perceptions.
Comparing public opinion across countries is difficult because it is easy to question the cross-cultural validity of the hypothesized constructs (cf.Davidov et al., 2011).We show that even though average scores across countries may differ, Americans and Britons conceptualize postures or dimensions of foreign policy attitudes (militarism, isolationism/internationalism, and globalization), patriotism, and economic evaluations in a similar manner.This finding helps reassure us that the variables take on a comparable meaning across states, and thus has positive implications for future research on opinion formation on matters of foreign policymaking in the United States and UK.Although the findings presented in Table 2 should be replicated, it does appear that key foreign policy postures are valid in both countries.It remains a question whether the cross-cultural validity we find the postures to have hold when comparing more dissimilar states such as Japan to the UK, France to Australia, and so forth.Seeing if this finding holds across a larger set of countries is paramount to the crosscultural study of public opinion on matters of foreign policy.
Our argument is that US-UK differences are rooted in the greater salience of China as a political issue in the United States.Different factors likely contribute to this greater salience-the greater coverage of the rise of China in the US media, the visible presence of Chinese products on the shelves of American stores as well as Britain's lack of military interests in the Asia-Pacific and concern with the European Union.This analysis still leaves fundamental questions unanswered, and our data alone cannot answer them.Is the greater media coverage about China in the United States the cause of greater China salience, or a consequence of greater China salience?Does the US's role as economic and military role as superpower make China a bigger issue because the United States could choose to challenge her, whereas more limited British resources simply keep this off the table for the UK?Of course we expect the opposite as well-issues that are higher salience in the UK should also result in more nuanced and separable evaluations among the British public across dimensions that we would find in America (e.g.reactions to policies of the European Union).
The linking of postures to specific attitudes and noting their points of variation across states using the issue of China in two policy realms is just the tip of what can be a wider cross-cultural project to link individuals' foreign policy postures to their response to specific and often remote foreign policy issues.In the United States, our results show that postures tied to elements of hard (militarism) and soft (globalization) power are key predictors of American beliefs concerning the importance of their country's tough policies toward China in the military and economic policy realms.The impact these postures have on specific policies is mostly direct and, contra the results suggested by previous work, quasiideological postures do not need to work through the simpler heuristic of general attitudes toward a country to motivate specific policy responses.For respondents from the United States, there is also some (but not much) differentiation across the two policy domains under observation.In the UK, where the relevance of China is lower, it is the broader posture of isolationism that plays the largest role, and the China feeling thermometer is more of a relevant mediator.More exploration should be done as to whether people's opinions toward specific aspects of their country's foreign policies toward another state are a function of specific or broad postures.A separate cross-country concern is the extent to which the different electoral systems (one candidate centered, one party centered) connect the public's China preferences to government policy -is there an electoral connection on the issue of China (see Aldrich et al., 2006)?
Lastly, returning to China, the fact that our multivariate models exhibit more explanatory power when the American sample is considered suggests that American policymakers have a bit less free space to maneuverer when it comes to setting policy.While it is unlikely that slight shifts in America's China policy will swing a presidential election, a move against the will of the American voters may lead them to question, ever so slightly, the competence of their leaders.In the UK, this is less likely to be the case but still possible.Ultimately, a larger research program will more firmly link citizens' positions on specific postures and the ensuing positions on specific policy beliefs to behavior and attitudes toward the politicians who make policy.This is a small piece in the puzzle of what hopefully will emerge into a larger research program.
Partisan Identification: Dichotomous variables capture whether, in the United States, the respondent is a Democrat or Republican, and, in the UK, whether the respondent is a supporter of Labor, the Conservatives, the Liberal Democrats, the Greens, UKIP, or one of the two nationalist parties (Scottish Nationalists or Plaid Cyrmu).ii.Other Controls: The estimations below consider or control for the impact of the following social-demographic variables on attitudes toward i. China: age, gender, university education or not, region (South in the United States, Wales and Scotland in the UK), religious denomination (Protestant/Church of England, Catholic, other Christian, other religion), income (via dichotomous high and middle income controls) and race (African-American, Hispanic, and Asian in the United States, nonwhite in the UK).

Table 1
Thinking about (a) UK and (b) US Policy toward China, do you think it is important to be tough with China on. . .
Note: Variables treated as ordinal in multivariate analyses.

Table 2
CFA model of independent variables presents the standardized coefficients from the final estimations. 86 Standardized loadings differ slightly across states because the co-variances between the latent variables and indicators are allowed to vary freely across countries.Notable differences include the fact that Isolationism has a small negative correlation with militarism (À0.09) in the United States but the two latent variables have a small positive association in the UK (0.19).There is a moderate negative correlation between militarism and economic wellbeing in the United States, but this relationship does not obtain in the UK.Full results for the CFA are available upon request.The Root Mean Square Error of Approximation is a commonly employed approximate fit statistic employed in Structural Equation Modeling, and scores of below 0.05 are judged to have very good near fit

Table 3
Covariates of economic and military toughness, UK and United States Simultaneous unrelated ordered probit estimation via the WLSMV estimator as implemented in Mplus 7.3.Latent predictors are indicated in boldface.While we are simultaneously estimating two ordered probit models for each country, we only report a single coefficient when slopes are equivalent across models.All UK predictors constrained to equality.US predictors constrained to equality for all variables except Globalization, Patriotism, and Male.UK n ¼ 1891, US n ¼ 1859.Difference Test for UK equality constraints (vs.unconstrained model): v 2 ¼ 24.286 (df ¼ 23) P ¼ 0.388.Difference Test for imposed US equality constraints (vs.unconstrained model): v 2 ¼ 14.087 (df ¼ 17) P ¼ 0.661 Standardized coefficients displayed.

Table 4
Direct and indirect effects of selected predictors on preferences for tough policy via a china feeling thermometer

Table 5
Summary of hypotheses and results