Abstract

“Rising nationalism” has been a major meme in commentary on the development of China's material power since the early 1990s. Analysts often claim that rising nationalism, especially among China's youth, is an important force compelling the Chinese leadership to take a tougher stand on a range of foreign policy issues, particularly maritime disputes in East Asia. The rising nationalism meme is one element in the “newly assertive China” narrative that generalizes from China's coercive diplomacy in these disputes to claim that a dissatisfied China is challenging a U.S.-dominated liberal international order writ large. But is this meme accurate? Generally, research on Chinese nationalism has lacked a baseline against which to measure changing levels of nationalism across time. The data from the Beijing Area Study survey of Beijing residents from 1998 to 2015 suggest that the rising popular nationalism meme is empirically inaccurate. This finding implies that there are other factors that may be more important in explaining China's coercive diplomacy on maritime issues, such as elite opinion, the personal preferences of top leaders, security dilemma dynamics, organizational interests, or some combination thereof.

Introduction

“Rising nationalism” has been a major meme in commentary on the development of China's material power since the early 1990s. Indeed, “rising” is one of the most common qualifiers used by analysts to modify the term “Chinese nationalism.” It is a common meme in media from around the world, as a search of LexisNexis will attest.1 It appears in think tank analyses,2 in statements from the U.S. Department of State, in U.S. Congress members’ commentary, and in other government analysis.3 Some of the earliest references to the meme came after Chinese leaders launched the Patriotic Education Campaign (PEC) in the post-Tiananmen period, in an effort to inoculate China's youth against an alleged Western ideological “peaceful evolution” strategy.4 If one searches “rising Chinese nationalism” in Google Scholar, one finds that much of the scholarship on Chinese politics and foreign policy has used the term. In short, among many, if not most, non-Chinese and many Chinese observers, it is largely taken for granted that nationalism in China has been rising since at least the early 1990s. Given the persistence of this empirical claim, one might wonder just how high or extreme nationalism in China can get.

Although little of this commentary defines the term “rising,” by usage it seems to mean anything from more intensely felt, to more extreme, to more pervasive within society, to more salient for leaders, to more influential on policy. For the purposes of this article, I focus on rising nationalism in reference to the intensity or degree of nationalism expressed by ordinary citizens. I do not focus on the question of whether Chinese leaders are more sensitive to nationalism (indeed, they could be more sensitive to nationalism even as it declines in intensity or degree). Nor do I examine in detail the question of whether China's leaders have been more intensively promoting nationalism over time. In general, the discourse about rising nationalism, particularly in the media and among pundits, tends to refer more to the intensity of popular nationalism and less to its influence on policy.

The rising nationalism meme is one element in the narrative of a “newly assertive China” that generalizes from China's coercive diplomacy in maritime space to claims that a dissatisfied China is challenging a U.S.-dominated liberal international order writ large.5 Rising nationalism, in this narrative, constrains the Chinese leadership and helps explain its more militantly realpolitik or noncooperative behaviors, as well as China's more proactive challenges to international “rules of the game.”6 In addition, rising popular nationalism has led many U.S. government officials to worry that the Chinese leadership will engage in diversionary conflict when China's economic growth slows.7 Thus the rising nationalism meme is an important element of an even larger emerging narrative about a revisionist China.

Using an original time-series survey dataset from Beijing and a range of indicators of nationalism extending back to 1998, I conduct five different tests of the claim about China's rising nationalism. The analysis shows that there has not been a continuously rising level of nationalism among the survey respondents. Indeed, most indicators show a decline in levels of nationalism since around 2009. Moreover, contrary to the conventional wisdom, the data do not show that China's youth express higher levels of nationalism than older generations. Indeed, it is China's older generations that are more nationalistic than its youth. These findings—with due regard for caveats about representativeness— suggest that rising popular nationalism may not be a critically important variable constraining Chinese foreign policy.

The findings from the survey have a number of implications. First, they suggest that China's coercive diplomacy in maritime disputes over the last few years is unlikely to be a driven by rising popular nationalism. A list of more likely explanations includes the search for energy, fishing rights, the organizational interests of the military and of maritime law enforcement, leaders’ preferences, elite nationalism, Chinese reactions to actions by other maritime claimants, an emerging U.S.-China security dilemma, and growing naval capabilities. Second, the findings suggest that rising popular nationalism is unlikely to spur Chinese leaders to engage in diversionary conflict in the face of an economic slowdown.8 On a number of measures, levels of Chinese nationalism have stagnated or dropped since around 2009, even as annual economic growth rates have declined somewhat. Finally, the findings underscore that analysis of Chinese foreign policy has to be more consciously self-critical of the assumptions, conventional wisdoms, and memes that are circulated in the media, among pundits, and in academia. They also confirm that these elements are being rapidly reproduced in an era of digital and social media. Given the implications for stability in U.S.-China relations of the emergence of a “revisionist China” meme in the United States and of a “containing China” meme in China, it is especially important that scholars examine emerging conventional wisdoms and viral memes carefully and with a variety of data and methods.

The article starts with a review of the debate in the academic literature over rising Chinese nationalism and suggests how the heretofore absence of time-series data has limited progress in analysts’ understanding of Chinese nationalism. The next section introduces the Beijing Area Study survey, a dataset that provides time-series data on a range of questions that tap into different components of nationalism. I then present five tests for rising nationalism, including the results of questions that are standard in the survey literature on nationalism, questions that tap into ethnocentrist views, and questions that measure hostility toward other countries. I also control for age, given that one element of the rising nationalism meme is the belief that the most extreme expressions come from Chinese youth as a whole. I find that on balance among the survey respondents Chinese nationalism has not been rising continuously over the last decade and a half, and in some measures has declined. Moreover, it is clear that younger respondents are less nationalistic than older ones.

Debates over Rising Nationalism in China

How do we know if Chinese nationalism is rising? We know in part because the media tells us so. As media system dependency theory in sociology and media agenda-setting literature in political science would suggest, the media is one of the main sources of information about international politics for nonspecialists,9 Yet, often, reporting or analysis that mentions “rising Chinese nationalism” does not provide much, if any, systematic evidence for this claim.10 And when it does provide evidence, it is mostly anecdotal, based on non-random interviews with a small number of Chinese students, scholars, or officials; the non-random selection of trending books or videos; the non-random selection of anonymous posts on the internet; or, most often, on reports about (relatively rare) street demonstrations by Chinese youth. Sometimes the claim is made that rising nationalism exists because it is assumed, though not shown, that official government policies such as the Patriotic Education Campaign, launched in the early 1990s, are having their intended effect. Many of these analyses select on the dependent variable; that is, they look only at evidence that confirms the rising nationalism meme. For each of these observations, it is likely that one could find counterexamples.11 As far as I am aware, no scholar, think tank analyst, or journalist has tried to determine the relative importance or weight of one set of anecdotes over another.

This is not to say that the extant literature on Chinese nationalism is of poor quality. Most scholars working on nationalism in China fully understand the complex content and evolution of modern Chinese nationalism. One only need think of Vanessa Fong's thoughtful anthropological work that uncovered a youth nationalism expressed as a filial, though critical, loyalty to the cause of China's modernization;12 or William Callahan's careful unpacking of the historical evolution and social construction of memories of humiliation;13 or Tang Wenfang and Benjamin Darr's innovative comparison of levels of nationalism in China with those in other countries;14 or Brian Rathbun's rigorous differentiation between patriotism and nationalism;15 or Peter Gries’ creative analysis of letters in the wake of the 1999 U.S. bombing of China's embassy in Belgrade to gauge the genuinely emotional content of youth nationalism;16 or James Leibold's careful discourse analysis of a revived Han chauvinism among some Chinese youth.17 Then there are James Reilly and Jessica Chen Weiss, who are leading the way in understanding the scope conditions under which nationalism may affect foreign policy.18 They focus mainly on the minority of citizens who participate in nationalist demonstrations, rather than on broad trends in popular nationalism per se. But as good as this and other work is at uncovering the subtleties of Chinese nationalism, much of it either accepts the rising nationalism meme or does not study the phenomenon across time.

I am not the first to ask whether Chinese nationalism is, in fact, rising. There are a number of skeptics of the rising nationalism meme. Zhu Tianbiao, for example, argues that the goal of Chinese nationalism is national independence and territorial sovereignty. This has been a long-standing goal for different Chinese governments and leaders across time, and there is little evidence of much variation in this goal.19 Jia Qingguo suggests that memes about rising nationalism or an out-of-control nationalism are ahistorical and empirically problematic, and that they divert attention away from the external or international drivers of Chinese foreign policy.20 Using surveys and interviews of participants in the demonstrations outside the U.S. embassy in Beijing in 1999, Yu Zhiyuan and Zhao Dingxin look at the demonstrators’ reasons for participation in this iconic expression of contemporary youth nationalism. They find that nationalism and anti-U.S. sentiment were not nearly as strong as other motivations such as anger at a discrete event, peer pressure, and for many an exciting chance to participate in political theater.21 Allen Carlson's skepticism is bluntly empirical: “[D]espite (or, perhaps, because of) the pervasiveness of the resulting conventional wisdom about a surge in Chinese nationalism, to date the literature lacks any rigorous empirical testing of such an observation.”22 Most recently, Li Liqing has drawn on structured interview data with Chinese students to suggest that quotidian consciousness about, and commitment to, the nation is quite weak.23 For the most part, critical voices such as these have been ignored beyond the academic world. My starting point, then, is the scholarly skepticism embodied in this literature.

Testing the Rising Nationalism Claim: The Beijing Area Study

Testing the empirical validity of the rising nationalism meme has three requirements: conceptually valid, consistent, and observable or measurable indicators of nationalism; a metric with which to observe whether nationalism is rising or not; and a starting point, or baseline, against which to observe nationalism over time.24 The Beijing Area Study (BAS), a more or less regularly administered random sample survey of opinion in the Beijing municipality, meets these requirements for testing the rising nationalism claim. From the late 1990s onward, in up to eleven waves, the study has included questions that are standard in the survey literature on nationalism.25

In terms of measuring or observing nationalism, the scholarly literature typically considers patriotism and nationalism to be different, but related, cognitive and emotional concepts. Patriotism typically means love of, pride in, and support for the nation-state as an end in itself, but without necessarily any stringent emotional or ethical commitment to the political, economic, or social status quo (hence, patriots are also critics). Nationalism is a relational concept. Although it, too, is constituted by love of or pride in the nation, it also includes more explicit comparisons between highly valued in-group traits and the devaluation of out-groups. Thus, nationalism includes a more emotional and normative commitment to the political/economic/social status quo or putative status quo (in more extreme forms, blind support for the nation-state), as well as the denigration of the traits of out-groups (citizens of other nations, ethnonational groups, or both).26 When measuring nationalism, then, it makes sense—as much of the nationalism scholarship does—to include questions that tap into three related concepts: pride in and valuation of the nation;27 blind or naïve support for the nation-state; and the in-group's valuation of relevant out-groups.28

The BAS includes questions that can tap into these three different components of nationalism. First, since 2002 it has asked respondents whether they strongly agreed, somewhat agreed, somewhat disagreed, or strongly disagreed with the following standard statements used in many other nationalism surveys: (1) even if I could choose any other country in the world, I would prefer to be a citizen of China than any other country; (2) in general, China is a better country than most others; and (3) everyone should support their government even when it is wrong.29 Although distinct, these components are nonetheless correlated.30

Second, since 2000 the BAS has asked questions about perceived ethnonational identity differences between Chinese people, on the one hand, and Japanese and Americans, on the other. Specifically, the BAS asks respondents to place Chinese people and members of an out-group (Japanese people and American people) on a 7-point scale anchored by antonyms. The difference between where respondents placed Chinese people and where they placed members of the out-group is a measure of ethnonational identity difference.31 Identity difference perceptions are proving to be important explanations of policy preferences.32

Finally, since 1998 the BAS has measured the level of respondents’ amity toward various countries on a 100-degree feeling thermometer, where 0 represents the most extreme cold and hostile feelings and 100 represents the most extreme warm and friendly feelings. The list of countries has included potential adversaries and competitors such as Japan and the United States.

In addition to these measures, one can use the BAS to look at changes in dispersion around the means to see if portions of the population have moved in more extreme directions across time. Is the proportion of those with extreme anti-Japanese or anti-American sentiments growing, as one might expect from a rising nationalism meme? Are youth, for instance, more nationalistic than older generations, thus increasing the overall level of nationalism as they grow older?

To be sure, there are downsides to the BAS data. First, the measures of nationalism that the BAS uses, though standard in the survey literature on nationalism, do not easily map on to some of the categories of nationalism found in the China literature (e.g., affirmative, aggressive, assertive, pragmatic, anti-Western, and liberal).33 Therefore it is hard to use the BAS to measure changes in, say, the proportion of “aggressive” nationalists in China. Some of the measures in the BAS could perhaps fit some of these categories—high perceptions of identity difference are clearly associated with more realpolitik preferences (as might describe “assertive” nationalism). But for the most part, I do not address whether the categories/types used in the China-specific literature show continuity or change in their distribution across time.34

Second, the sample is from those living in the Beijing municipality. Beijing does not necessarily represent the values and preferences of residents of other parts of China. Its inhabitants are wealthier and better educated than those in most other places in China. That said, as a check on representativeness I did compare the responses in the 2007 BAS to the three standard nationalism statements mentioned above with the responses to identical questions in a nationwide survey conducted in early 2008 by Tang Wenfang from Iowa State University; the proportions were almost identical.35

Related, the BAS sample may not be entirely representative of Beijing's population as a whole either. In general, the gender distribution in the BAS sample has been reasonably close to that of Beijing as a whole. The reported monthly per capita income in the BAS sample was also relatively close to official Beijing figures through to the mid-2000s, but had fallen considerably behind by 2010.36 The main difference is in levels of education: participants in the BAS sample tend to be consistently better educated than the rest of the Beijing population (based on comparisons with the censuses of 2000, 2005, and 2010).

Third, in 2007 the BAS changed from randomly sampling from household registration lists within Beijing's central districts to randomly sampling residents using GPS sampling of the wider Beijing metropolitan area, including some rural areas. Despite this change in sampling methods, in general the samples are relatively comparable in terms of basic demographic characteristics across the 2007 change. To make better comparisons across the 2007 change, unless otherwise indicated, I used data only for those respondents who had Beijing urban household registrations from 2007 on. In general, the samples are relatively comparable in terms of basic demographic characteristics across the 2007 change.37

Despite these problems, there are some advantages to a Beijing-focused sample. First, Beijing matters. Public expressions of popular nationalism on foreign policy issues, such as demonstrations, are most likely to worry the leadership if they occur in Beijing. Beijing is the center of politics and propaganda. If there were likely to be a hotbed of nationalism, it might be Beijing. Second, the change in sampling from the 2007 BAS on allows for some exploration of non-Beijing opinion, because one can compare those with Beijing residence permits to those who immigrated to Beijing from other parts of China. I should note that the trend lines in nationalism (though not the absolute levels of nationalism) from 2007 on, which are summarized in figures 1, 2, and 3, are similar for those with non-Beijing household registration. This similarity mitigates, to some extent, the problem of the representativeness of the Beijing sample. Third, to reiterate, as a preliminary test of the rising nationalism hypothesis, the BAS is the only available dataset as far as I am aware that provides a baseline and comprehensive, multiple indicator time-series data on nationalist sentiment.

Figure 1.

“I Would Prefer to Be a Citizen of China”

Figure 1.

“I Would Prefer to Be a Citizen of China”

Figure 2.

“China Is a Better Country Than Most”

Figure 2.

“China Is a Better Country Than Most”

Figure 3.

“You Should Support Your Country/Government Even When It Is Wrong”

Figure 3.

“You Should Support Your Country/Government Even When It Is Wrong”

What, then, do the BAS data show? In short, the conventional wisdom may overestimate the evidence for rising Chinese nationalism. In four tests for rising nationalism, I find evidence that in some cases levels of nationalism in Beijing today are lower than they were in the early 2000s. In other cases, the levels of nationalism may be somewhat higher today than they were in the early 2000s, but are lower than they were in the mid-to-late 2000s. Moreover, as a fifth test suggests, contrary to the conventional wisdom, Chinese youth are less nationalistic than older generations, and by some measures less nationalistic today compared to the early 2000s.

Five Tests of the Rising Nationalism Claim

Scholarship on nationalism uses a range of indicators to observe change in level or intensity. I follow this eclecticism, with due concern for construct validity, and develop five tests that tap into different elements of nationalism. These tests also try to capture the range of meanings of “rising popular nationalism” that appear in the academic and pundit commentary.

TEST 1: PRIDE IN NATION AND BLIND SUPPORT FOR THE STATE

The first test of rising nationalism looks at change in the distribution of agreement with three nationalism statements that are often used in scholarly surveys of nationalism in other countries. They tap into pride in citizenship, into beliefs in the superiority of China over other countries, and into more nativist or blind support for the state and country. Because each statement captures different dimensions of nationalism, I have not built a composite nationalism scale from the three measures. As I noted earlier, however, these three measures are strongly associated with one another.

The longitudinal data for these measures suggest that there has been no sharp rise in nationalism in China since 2002, the first year the BAS used these measures.38 As figure 1 indicates, prior to the Beijing Olympic Games in 2008, strong agreement with the statement about preferring to be a citizen of China than of any other country waivered around a stable mean, and the trend line was essentially flat. Strong agreement spikes in 2009, possibly in response to the effects of the 2008 Olympics. Thereafter it drops considerably to levels lower than before 2008. By 2015 the plurality response is no longer “strongly agree” but rather “somewhat agree.”39

Agreement with the statement that China is better than most other countries is somewhat less extreme, with higher portions choosing the “somewhat” answer (figure 2). As with the first measure of nationalism, agreement with this statement prior to the Olympic Games varied across time. There was a peak in strong agreement with the statement in 2009, but by 2013 opinion shifts toward the “somewhat agree” category as well.40

As for support for the nativist statement about supporting the country/government even when it is wrong, this is the weakest of the nationalism measures (figure 3). In three of the six years in which this statement was posed, a plurality disagreed with the statement. Even in 2009, with the Olympic Games effect apparently still at work, only a small majority of the sample agreed with the statement, about the same as in 2004.41

It is easier to visualize the absence of a secular rise in Chinese nationalism by looking at the change from one survey wave to the next in the percentage of respondents who “strongly agree” with each of the above statements (figure 4). If there were a constant rise in the most extreme forms of nationalism, the lines representing each of the three statements would be above 0 and basically flat; and if the rise were exponential, the lines would be above 0 and sloping up from left to right. Neither is the case.42

Figure 4.

Change from Survey-to-Survey in the Percentage of Respondents Who “Strongly Agree” with Nationalism Questions

Figure 4.

Change from Survey-to-Survey in the Percentage of Respondents Who “Strongly Agree” with Nationalism Questions

In short, the trends in these standard nationalism statements do not fit the meme of constantly rising or intensifying Chinese nationalism over the period covered by the data.

TEST 2: IDENTITY DIFFERENCES WITH JAPANESE AND AMERICANS

The second test of rising Chinese nationalism looks at change in what I call “identity difference scores” across time. The data used in this test are based on Osgood semantic differential procedures.43 As noted earlier, respondents were asked where they would place members of their in-group and those of the out-group on a 7-point scale anchored by antonyms. The particular antonyms in the BAS come from common words that Chinese people use when talking about themselves and others in terms of basic group traits (e.g., peacefulness, modesty, civilization, and sincerity). The scores on these traits were averaged for Self and for Other. The average score for Self across these traits was then subtracted from the average score for Other across these traits. This difference is the identity difference score. The identity difference score taps into the degree to which Chinese people believe that they are different in their essence from Japanese and American people. It captures a deep cognitive structure that influences individuals’ policy preferences toward out-groups. Social identity theory and a growing literature in the social neurosciences on in-group–out-group differentiation suggest that, under certain conditions, larger perceptions of difference predict to more hardline policy preferences toward out-groups. Some scholars use this type of identity difference measure as a proxy for ethnocentrism.44

Based on these identity difference scores, the data do not confirm a rapidly or even constantly rising level of Chinese nationalism. As figures 5 and 6 show, the quadratic trend line appears to show that perceptions of identity difference (mean value) peaked in 2009 and declined somewhat after that (though the 95 percent confidence interval in the Japan cases suggest possibly more stasis). There is virtually no difference if one looks at the trends in median values.45

Figure 5.

Perceptions of Chinese Identity Difference with Japanese People

Figure 5.

Perceptions of Chinese Identity Difference with Japanese People

Figure 6.

Perceptions of Chinese Identity Difference with American People

Figure 6.

Perceptions of Chinese Identity Difference with American People

TEST 3: ARE EXTREME VIEWS BECOMING MORE COMMON?

A third test focuses on change in the distribution of these various measures of nationalism. It is possible to see little change in the mean of some measures while nonetheless observing a shift in its distribution. In this case, I focus on change in the distribution of the identity difference scores. Is there a continuous increase in the proportion of those who perceived Japanese and Americans in extremely negative terms, as one might expect with a continuously rising nationalism? The data do not support a linear increase in extremist views. Figures 7 and 8 show that the proportion of the sample who perceived the maximum negative difference between Chinese people and Japanese and American people (e.g., Chinese were ranked at the positive end of the scale and Japanese and Americans at the negative end) increased over the 2000s, but began to flatten out after 2007 (in the case of perceptions of Japanese people) or decline (in the case of perceptions of American people).

Figure 7.

Percentage of Respondents Perceiving Maximum Negative Difference with Japanese People

Figure 7.

Percentage of Respondents Perceiving Maximum Negative Difference with Japanese People

Figure 8.

Percentage of Respondents Perceiving Maximum Negative Difference with American People

Figure 8.

Percentage of Respondents Perceiving Maximum Negative Difference with American People

TEST 4: AMITY AND HOSTILITY TOWARD OTHER MAJOR POWERS

A fourth test focuses on the degree of Chinese amity toward other major potential adversary states. Although levels of amity are not usually used as a proxy for nationalism, they provide another window into the degree to which salient foreign countries (as opposed to peoples) may be denigrated or viewed with hostility.46 Thus it is useful to examine amity levels across time as another indicator of trends in nationalism.47

As figures 9 and 10 show, Chinese amity toward Japan and the United States declined from 1998 to around 2009, when it appears that the trend line bottomed out. In fact, there was a slight increase in amity toward the United States from 2009 to 2015, though this change falls within the 95 percent confidence lines. The quadratic trend lines would suggest a slowing down, if not halting, in the decline of amity toward China's major perceived political-military challengers.

Figure 9.

Trends in Amity toward Japan

Figure 9.

Trends in Amity toward Japan

Figure 10.

Trends in Amity toward the United States

Figure 10.

Trends in Amity toward the United States

TEST 5: NATIONALISM AND AGE

A fifth test focuses on trends in nationalism among Chinese youth. Regardless of whether Chinese nationalism is instrumentally produced from above, spontaneously arises from below, or both, the Chinese Communist Party (CCP) has explicitly tried to promote nationalism among Chinese youth.48 As a result, analysts from both inside and outside China often claim that rising nationalism is mainly a youth phenomenon, as references to a rising “angry youth” (fenqing) generation suggest.

The data show, however, that younger respondents are no more nationalist than older generations. For instance, in reexamining all the BAS statements related to nationalism, I controlled for respondents who were the targets of the Patriotic Education Campaign begun in 1993/94. For convenience, I call them members of the patriotic education generation (PEG). I define the PEG as including those who were fifteen years old in 1993 or younger. As figures 1113 show, from 2007 to 2015 the patriotic education generations were consistently less likely to strongly agree with the three standard nationalism statements compared to older respondents.49

Figure 11.

“I Would Prefer to Be a Citizen of China” and Generation (BAS 2007–15)

Figure 11.

“I Would Prefer to Be a Citizen of China” and Generation (BAS 2007–15)

Figure 12.

“China Is a Better Country Than Most” and Generation (BAS 2007–15)

Figure 12.

“China Is a Better Country Than Most” and Generation (BAS 2007–15)

Figure 13.

“You Should Support Your Country Even When It Is Wrong” and Generation (BAS 2007–15)

Figure 13.

“You Should Support Your Country Even When It Is Wrong” and Generation (BAS 2007–15)

This difference between older respondents and the PEG extends to identity difference scales as well. Table 1 shows that simple t-tests using the BAS 2007 to 2015 data indicate that the PEG holds perceptions of significantly smaller difference compared to older respondents. Put differently, the PEG tends not to treat Japanese and Americans as the Other as much as older generations do.

Table 1.

Identity Difference and Generation

Identity Difference (Japanese)Older GenerationPatriotic Education Generationt*p
2015 3.15 (N= 958) 1.83 (N= 366) 9.77 0.00 
2013 2.25 (N = 1,109) 1.34 (N= 301) 5.44 0.00 
2009 3.09 (N= 407) 2.87 (N= 173) 1.23 0.11 
2007 2.9 (N= 484) 2.64 (N= 206) 1.26 0.1 
Identity Difference (Japanese)Older GenerationPatriotic Education Generationt*p
2015 3.15 (N= 958) 1.83 (N= 366) 9.77 0.00 
2013 2.25 (N = 1,109) 1.34 (N= 301) 5.44 0.00 
2009 3.09 (N= 407) 2.87 (N= 173) 1.23 0.11 
2007 2.9 (N= 484) 2.64 (N= 206) 1.26 0.1 
Identity Difference (Americans)Older GenerationPatriotic Education Generationt*p
2015 2.56 (N= 895) 1.72 (N = 364) 6.71 0.00 
2013 1.78 (N = 1,105) 0.83 (N= 301) 5.89 0.00 
2009 3.15 (N= 405) 2.81 (N= 173) 1.83 0.03 
2007 2.36 (N= 484) 2.12 (N = 205) 1.26 0.10 
Identity Difference (Americans)Older GenerationPatriotic Education Generationt*p
2015 2.56 (N= 895) 1.72 (N = 364) 6.71 0.00 
2013 1.78 (N = 1,105) 0.83 (N= 301) 5.89 0.00 
2009 3.15 (N= 405) 2.81 (N= 173) 1.83 0.03 
2007 2.36 (N= 484) 2.12 (N = 205) 1.26 0.10 
*

two-tailed test

In sum, these findings do not support the argument that China's youth are more nationalistic than their elders.

The PEG is a fairly blunt category, however. There has been much talk about variation within the PEG across those Chinese born in the 1980s (called, confusingly, the post-80s generation) and those born in the 1990s (the post-90s generation). In addition, much of the analysis suggests that these two generations are more nationalist than were previous ones.50

The BAS data do not support the notion that there is a major difference between the 80s and 90s generations. Both generations are significantly less nationalistic than older generations, and the 90s generation appears to be less nationalistic than the 80s generation on most measures. For example, if one looks at the three standard nationalism statements, both the 80s generation and the 90s generation were significantly less likely to strongly agree with these statements than was the older generation (figures 1416).51 There is virtually no difference between the two younger generations regarding support for the first two statements. In the third statement, which captures a more nativist or unquestioning nationalism, the 90s generation is much less likely than the 80s generation to “strongly agree” and more likely to “somewhat disagree.” This difference suggests that the 90s generation has a somewhat more skeptical view of government than the 80s generation, but both are much more skeptical than is the older generation.

Figure 14.

“I Prefer to Be a Chinese Citizen” and Generation (BAS 2015)

Figure 14.

“I Prefer to Be a Chinese Citizen” and Generation (BAS 2015)

Figure 15.

“China Is Better than Most Countries” and Generation (BAS 2015)

Figure 15.

“China Is Better than Most Countries” and Generation (BAS 2015)

Figure 16.

“Support Your Country Even When It Is Wrong” and Generation (BAS 2015)

Figure 16.

“Support Your Country Even When It Is Wrong” and Generation (BAS 2015)

As for measures of identify difference with Japanese and Americans, table 2 shows the differences among the three generations in the BAS 2015 data. The average identity difference scores for 80s and 90s generations are statistically significantly lower than the score for older generations of respondents.

Table 2 also suggests that as respondents get younger, their perceptions of identity difference decline. The 90s generation generally holds lower perceptions of identity difference compared to the 80s generation.52 This difference is statistically significant regarding perceptions of Americans,53 but of only borderline significance with regard to perceptions of difference with Japanese.54

Table 2.

Generation and Identity Difference (BAS 2015)

 Identity Difference Score (Japanese)t (compared to older generation)p*
Older generations 3.19 (N = 1,396)     
80s generation 2.15 (N= 533) 9.37 0.00 
90s generation 1.83 (N= 209) 8.64 0.00 
 Identity Difference Score (Japanese)t (compared to older generation)p*
Older generations 3.19 (N = 1,396)     
80s generation 2.15 (N= 533) 9.37 0.00 
90s generation 1.83 (N= 209) 8.64 0.00 
 Identity Difference Score (Americans)t (compared to older generation)p*
Older generations 2.64 (N = 1,319)     
80s generation 1.85 (N= 520) 7.47 0.00 
90s generation 1.34 (N= 212) 8.86 0.00 
 Identity Difference Score (Americans)t (compared to older generation)p*
Older generations 2.64 (N = 1,319)     
80s generation 1.85 (N= 520) 7.47 0.00 
90s generation 1.34 (N= 212) 8.86 0.00 
*

two-tailed test

Perhaps age matters most at the extremes. Some analysts suggest, for instance, that China's so-called angry youth holds the most extreme nationalist views—nativist, anti-Western, and anti-Japanese. If, however, one looks at those who maximally denigrate the identity traits of Japanese and Americans (e.g., those who score a 6 on the identity difference scale), the mean age in 2015 is about ten years older than the age of those who hold less extreme views.55 Both the 80s and 90s generations are less likely to be in this extremist group compared to the older generation. As figures 17 and 18 show, there is a general linear relationship across generations. Those with extreme perceptions of identity difference toward Japanese and Americans are about twice as likely to be from the older generation than from the 90s generation. The generational differences are statistically significant with regard to perceptions of differences with Japanese,56 but only borderline significant when it comes to Americans.57 There is no statistically significant difference between the 80s and 90s generations themselves.58

Figure 17.

Generation and Extreme Identity Difference with Japanese (BAS 2015)

Figure 17.

Generation and Extreme Identity Difference with Japanese (BAS 2015)

Figure 18.

Generation and Extreme Identity Difference with Americans (BAS 2015)

Figure 18.

Generation and Extreme Identity Difference with Americans (BAS 2015)

The amity data tell a similar story. In 2007 there was no major difference between the PEG and older generations regarding amity toward the United States, though the PEG held cooler views toward Japan. Since then, however, the PEG has demonstrated significantly warmer feelings toward both countries compared to older respondents (see table 3). Indeed, the PEG generation appears to be the source of the slowing or halting of the decline in levels of amity toward Japan in the last several years.

Table 3

Generation and Amity (BAS 2015)

Amity toward JapanOlder GenerationPatriotic Education Generationtp*
2015 26.88 (N = 1,013) 40.10 (N = 408) −9.16 0.00 
2013 29.42 (N = 1,131) 38.50 (N=311) −6.33 0.00 
2009 31.75 (N= 398) 36.48 (N=168) −2.32 0.01 
2007 36.53 (N= 494) 31.87 (N=206) 2.10 0.02 
Amity toward JapanOlder GenerationPatriotic Education Generationtp*
2015 26.88 (N = 1,013) 40.10 (N = 408) −9.16 0.00 
2013 29.42 (N = 1,131) 38.50 (N=311) −6.33 0.00 
2009 31.75 (N= 398) 36.48 (N=168) −2.32 0.01 
2007 36.53 (N= 494) 31.87 (N=206) 2.10 0.02 
Amity toward the United StatesOlder RespondentsPatriotic Education Generationtp*
2015 43.35 (N= 981) 55.71 (N=408) −8.69 0.00 
2013 42.05 (N = 1,130) 53.5 (N=311) −8.08 0.00 
2009 39.74 (N= 404) 47.03 (N=176) −3.39 0.00 
2007 50.78 (N= 490) 50.53 (N=205) 0.12 0.45 
Amity toward the United StatesOlder RespondentsPatriotic Education Generationtp*
2015 43.35 (N= 981) 55.71 (N=408) −8.69 0.00 
2013 42.05 (N = 1,130) 53.5 (N=311) −8.08 0.00 
2009 39.74 (N= 404) 47.03 (N=176) −3.39 0.00 
2007 50.78 (N= 490) 50.53 (N=205) 0.12 0.45 
*

two-tailed test

Nor are there generational differences between the 80s and 90s respondents concerning amity toward Japan and the United States. As table 4 shows, compared to the older generation, both the 80s and 90s generations have much warmer views of Japan and the United States, with no significant difference between the two cohorts’ mean amity levels.

Table 4.

Amity and Generation (BAS 2015)

 Amity toward Japan (degrees)t (compared to older generation)p*
Older generation 27.61 (N = 1,543)     
80s generation 36.40 (N= 604) −7.53 0.00 
90s generation 37.88 (N= 233) −6.09 0.00 
 Amity toward Japan (degrees)t (compared to older generation)p*
Older generation 27.61 (N = 1,543)     
80s generation 36.40 (N= 604) −7.53 0.00 
90s generation 37.88 (N= 233) −6.09 0.00 
 Amity toward the United States (degrees)t (compared to older generation)p*
Older generation 43.69 (N = 1,478)     
80s generation 53.71 (N= 601) −8.55 0.00 
90s generation 56.76 (N= 232) −7.66 0.00 
 Amity toward the United States (degrees)t (compared to older generation)p*
Older generation 43.69 (N = 1,478)     
80s generation 53.71 (N= 601) −8.55 0.00 
90s generation 56.76 (N= 232) −7.66 0.00 
*

two-tailed test

In sum, even when one divides the patriotic education campaign generation into two commonly identified cohorts, Chinese youth do not appear to be more hostile to Japan and the United States than do their elders.59 These findings about age hint at a more general explanation for the trends in Chinese nationalism, at least among Beijing's population. Fundamentally, Beijing's population has become more educated. And over time, the 80s and now the 90s generations—the generations that have been most exposed to the economic and cultural effects of globalization—have come to make up a greater share of the overall population.60

Objections to the Argument

This study has provided five tests of the conventional wisdom about rising nationalism in China. The article is not an explanation for variation in levels of nationalism in China or for the meme of rising popular nationalism in the commentary about China. Rather it tests a very common empirical claim about rising popular nationalism on a sample of the Beijing population using a baseline and metrics provided by the BAS survey. The article finds that, contrary to the conventional wisdom, there seems to have been relative stability in many measures of Chinese nationalism since the early 2000s, though with a temporary spike probably in response to the 2008 Olympic Games in Beijing. Since 2009 (and in some cases since 2007), there seems to have been a decline in most of the nationalism measures, in some cases to levels below that in the early 2000s. Moreover, the article finds, again contrary to the conventional wisdom, that Chinese youth are less nationalistic, not more nationalistic, than older respondents across all measures of nationalism. Those born after the late 1970s constitute the least nationalistic demographic.

There are two potential rejoinders to these findings. First, even if nationalism in China is not rising as rapidly or as consistently as the conventional wisdom would suggest, it is possible that Chinese citizens rank higher on nationalism scales compared with citizens of many other countries. Tang and Darr note that Chinese nationalism ranks in the top-ten countries on various measures.61 The difference with other top-ten nationalist countries, however, is relatively small. Tang and Darr's study compares China in 2008 with various countries from the 2003 International Social Survey Program (ISSP) national identity survey. If one compares the most recent ISSP survey on national identity (2013) with the full BAS 2013 dataset, however, Chinese respondents are actually less likely to strongly prefer being a citizen of China than Japanese respondents are to prefer being citizens of Japan (48 percent to 62 percent). Both Chinese and Japanese respondents are about as likely to strongly agree that their country is better than most others (46 percent versus 49 percent). Further comparisons between China and other countries with regard to the content and intensity of nationalism would be an important step in testing the conventional wisdom about trends in the intensity of nationalism in China.

The second rejoinder is that even if popular nationalism in China is not currently rising, the rising nationalism meme really refers to the intensification of the CCP's anti-foreign propaganda efforts to, in part, divert people's attention away from the regime's performance and blame foreigners for the country's social, political, and economic ills. The emerging narrative in much of the U.S. commentary is that President Xi Jinping in particular—in contrast to allegedly more moderate previous leaders such as Hu Jintao and even Jiang Zemin—is presiding over an unprecedented revival of anti-foreign discourse in China. Some in the U.S. media have claimed that under Xi, there has been a surge in official rhetoric about “hostile forces” or “foreign hostile forces.”62 In this regard, perhaps popular nationalism in China may begin to rise again as the effects of this propaganda campaign become felt.

Although one cannot rule out a rise in Chinese nationalist sentiments in the future—the BAS shows considerable variation across time in some of the indicators—the reality is much more complicated than the narrative. There is no doubt that the regime often accuses “foreign hostile forces” or, more commonly, “hostile forces” of ideological subversion.63 But compared to previous regimes, there has been no surge in the Xi regime's use of the term, at least as of this writing. If one looks across time at the frequency of articles referencing “hostile forces” in the People's Daily (the primary media organ of the CCP and hence one of the most important ideological signaling devices for the regime64), it is clear that the term was much more prevalent during domestic crises under previous leaders. As figure 19 shows, use of the term spiked after the Tiananmen massacre in 1989, when the regime tried to inoculate the population against an alleged Western “peaceful evolution” strategy of ideological subversion. It peaked again after 1998, when the Falungong sect was deemed to pose a serious challenge to the CCP's monopoly on political organization. At the time, the Jiang Zemin regime railed against U.S. efforts to “Westernize” and “divide” China. Then, in the mid-2000s, when inter-ethnic violence surged in Xinjiang and Tibet, usage of the meme again spiked in the People's Daily. The pattern seems fairly clear: when the regime worries about major challenges to its legitimacy, it tries to redirect blame to (mostly foreign) “hostile forces.”65 This is not a new phenomenon, and the Communist Party propaganda system under Xi has not used this meme nearly to the extent of previous Chinese leaders, at least not when it comes to the population as a whole.66 Since 2011, however, there has been a surge in the use of the term in the Liberation Army Daily, the main ideological signaling device to the Chinese military.67 In contrast to previous leaderships, where the messaging to society at large (through the People's Daily) and to the military (through the Liberation Army Daily) was in lockstep, under Xi the “hostile forces” meme appears to be aimed more at the military. The meme seems designed to warn against alleged foreign efforts to subvert CCP leadership of the military by spreading ideas about national rather than Party leadership of the People's Liberation Army. This observation is consistent with Xi's own analysis of the collapse of the Soviet Communist Party—namely, that a key reason for the collapse was the Soviet military's abandonment of the Party.68 The sharp divergence in the prominence of the meme in the People's Daily compared to the Liberation Army Daily under Xi suggests that, in relative terms, he is more worried about “foreign” ideological subversion in the military than in society writ large.69

Figure 19.

Average Monthly Number of Articles in the People's Daily and the Liberation Army Daily That Use the Term “Hostile Forces”

Figure 19.

Average Monthly Number of Articles in the People's Daily and the Liberation Army Daily That Use the Term “Hostile Forces”

Instead of a focus on anti-foreign “hostile forces,” Xi's primary ideological message is the “great revival of the Chinese nation” (see figure 20). I cannot discuss this concept in detail here, but suffice it to say, it could be read by ordinary Chinese people as either a vaguely benign pride in the material accomplishments of the nation or as a more vindictive wiping clean of a century of humiliation at the hands of foreign imperialists. This ideological campaign is too recent for the BAS data to reflect any clear effects of this message on levels of nationalism or identity difference. Measuring the effects of this message, however, should be a focus of future research.

Figure 20.

Average Monthly Number of Articles in the People's Daily That Use the Term “Great Revival of the Chinese Nation”

Figure 20.

Average Monthly Number of Articles in the People's Daily That Use the Term “Great Revival of the Chinese Nation”

The above discussion is not meant to suggest that nationalism is irrelevant to Chinese foreign policy. Although we still do not have a well-documented understanding of the mechanisms by which top Chinese leaders respond to nationalist public opinion, there is no reason to believe that Chinese leaders, and the elites who select them, are not themselves nationalistic to varying degrees.70 Even if popular nationalism is not rising, the presence of extremist or xenophobic, even racist, sentiments among elites may constrain Chinese foreign policy, particularly in relations with Japan. The claim that hard-core nationalists among the political elites may influence foreign policy, however, is based on a model of decisionmaking very different from the sweeping twenty-five-year conventional wisdom about the effects of rising popular nationalism. Moreover, the predictions about foreign policy should be different as well. Elite-based nationalism, for example, might predict that, ceteris paribus, top leaders should be more sensitive to nationalist sentiments around succession periods, when elite political struggle is likely to peak. In addition, to the extent that foreign governments consider elite-based nationalism expressed through high-profile voices in the media to be authoritative, this could contribute to security competition with other countries. On the other hand, a model of elite-based nationalism might predict that elite criticisms of leaders’ foreign policies can be contained, because the relatively small group of elites who influence succession politics can be more easily bought off through economic and political side payments.

Conclusion

What does it matter if popular nationalism in China is not rising the way it is often described to be by many in the media, pundit, and academic worlds? Does it matter if the narrative is empirically problematic, badly sourced, and yet continuously reproduced? The importance lies in the fact that the meme of rising Chinese nationalism is part of a broader meme about a less flexible, more assertive, and indeed more revisionist China. Not all of these descriptors of Chinese foreign policy are empirically inaccurate.71 However, the notion that rising nationalism is pushing China's leaders to take revisionist positions on everything from maritime disputes to trade and investment issues to the nature of the so-called liberal international order implies that their preferences are different from those of the Chinese population, and that they pay costs for ignoring them. These claims have not been systematically tested. Would Xi Jinping really prefer a more moderate diplomacy toward maritime disputes in the South China Sea if it were not for “rising” popular nationalism? There is no evidence that he has more moderate foreign policy preferences on territorial issues than the public at large. What costs would a powerful authoritarian leader such as Xi pay for ignoring public opinion when he faces no popular elections, when he commands a vast internal security apparatus as well as the military, and when he has powerful propaganda tools to guide opinion or, if necessary, suppress it?72 Because popular nationalism in China does not appear to have been rising or surging in the last few years, it is unlikely to be a plausible explanation for any recent sharp changes in levels of coercive diplomacy in China's maritime disputes.73

This article does not represent the final word on rising Chinese nationalism.74 When different sources and analytical methods lead to different conclusions about a political phenomenon, the next challenge is to determine why, rather than to claim victory for one approach over another. The findings here are only a starting point for the more systematic analysis of Chinese nationalism across time. To begin, the whole question of the impact of rising nationalism in China may be irrelevant if there is ultimately little constraint from public opinion on the views or options of top decisionmakers. As noted, this issue is still very much unsettled. Therefore, scholars need to develop logical models and mechanisms by which authoritarian leaders would want to take nationalist public opinion into account in foreign policy decisions. Some mechanisms might be indirect (e.g., political elites’ use of popular nationalist sentiments as a political tool in elite struggle). Some might be a function of elite opinion itself, whereby elite nationalism, not popular nationalism, is the source of pressure. Some might be normative, whereby China's top leaders believe that the CCP's role should be quasi paternalistic, reflecting public sentiment and wishes as the CCP interprets them. Some mechanisms might be entirely instrumental—namely, to head off nationalist-inspired anti-government social unrest or to use tied-hands bargaining strategies (claiming that public opinion constrains leader's options) to elicit concessions from other countries.

Testing these different mechanisms requires better data on nationalist sentiment in China. With regard to change or stasis in the content and intensity of popular nationalism, scholars of Chinese nationalism need more rigorously collected time-series data that measure nationwide change in various indicators of nationalism. This includes comparisons across findings both from surveys and from analysis of online opinion. The latter may be less representative, but nonetheless more immediately salient for political leaders. The former may be more representative and offer pools of more moderate opinion for elites to mobilize, should they so desire. In addition, the creative and inductive observation of alternative manifestations of popular nationalism (e.g., the content of popular culture experienced by ordinary people including television programs, movies, literature, and tourist sites) is critical—for instance, for observing how taken for granted or emotionally rooted these sentiments may be. This would be an important check on the validity of findings from surveys and from quantitative content analysis of online opinion.75 One might be able to take advantage of random exogenous shocks to tease out before and after “gut” impressions of the relative value of Self and Other (e.g., Chinese impressions of Japanese immediately before and after the 2011 Tohoku earthquake or perceptions of foreigners’ views of China before and after the Sichuan earthquake in 2008).

In addition, survey experiments could be useful for testing whether variations in levels of Chinese nationalism are related to variations in political behavior. Is nationalism or some other set of factors a better predictor of a willingness to participate in online debates or to go out on the streets to protest? And if it is, what rhetorical strategies could the CCP use to ameliorate or reduce the intensity of nationalist sentiments? How effective are efforts to “guide” nationalist opinion? The relationship between nationalism and political behavior is important to study because, without this connection, the political consequences of nationalist opinion are hard to observe. The connection speaks to the issue of how constrained or unconstrained the Chinese leadership might be by nationalist opinion.76

As for elite opinion, it is obviously harder to conduct random surveys. Still, non-random elite surveys built from informal personal networks may provide similar insights into elite opinion, as the elite surveys conducted by U.S. polling organizations such as Pew and the Chicago Council on World Affairs demonstrate. Alternatively, given the Leninist nature of China's political system, wherein a small number of CCP elites select the leadership, analysts might focus in particular on the content and levels of nationalism expressed by those few hundred Party members (especially members of the Party Central Committee) whose support Chinese leaders need to acquire and hold power.77 Doing so would create a more bounded and influential selection of elite opinion. For example, there may be sufficient paper trails of ex ante statements, media reports, and foreign policy-related actions by officials before they enter the Central Committee to construct a rough distribution of nationalist opinion at this elite level.

Analysts also need more systematic collection of information on the nationalist sentiments of China's highest leaders (e.g., members of the Politburo Standing Committee) prior to their movement into the institution, again through any paper trail created during their political ascent. There may be a selection process, of course, that leads to the homogenization of publicly expressed nationalist sentiments at the top. But even so, it is unlikely to be wholly instrumentally expressed throughout these leaders’ careers. For example, that Xi Jinping has chosen to use the language of genetics to express the alleged unique peacefulness of the Chinese people is perhaps revealing of the depth of his belief in China's exceptionalism.78

These research steps and procedures may or may not confirm the findings from the BAS surveys that Chinese nationalism has not been rising continuously over the last couple of decades. Nevertheless, they are necessary to ensure that the debate on this question is not prematurely closed because of either the conventional wisdom or viral memes.

Acknowledgments

For their helpful comments and criticisms, the author thanks Bjorn Jerden, Brian Rathbun, Wang Yuhua, Jessica Weiss, George Yin, and the anonymous reviewers.

Notes

1. 

Along with “rising,” a survey of hits across the years in LexisNexis for the word “China” appearing within ten words of “nationalism” finds similar but less commonly used quasi synonyms such as “muscular,” “aggressive,” “increasing,” “whipped up,” “heightened,” “resurgent,” “rising tide of,” “fanned,” “high,” “expanding,” “neo-,” “ebullient,” and “growing.”

2. 

See Clive Schofield and Ian Story, The South China Sea Dispute: Increasing Stakes and Rising Tensions (Washington, D.C.: Jamestown Foundation 2009), p. 3; Michael Schiffer and Gary Schmitt, “Keeping Tabs on China's Rise” (Muscatine, Iowa: Stanley Foundation, May 2007), www.stanleyfoundation.org/publications/other/SchifferSchimitt07.pdf; Francis Fukuyama “Reenvisioning Asia,” Foreign Affairs, Vol. 84, No. 1 (January/February 2005), http://www.cfr.org/asia-and-pacific/re-envisioning-asia/p7927; Yun Sun, “Chinese Public Opinion: Shaping China's Foreign Policy, or Shaped by It?” (Washington, D.C.: Brookings Institution, December 2011), http://www.brookings.edu/research/opinions/2011/12/13-china-public-opinion-sun; and Bruce Klingner, “Asia's Big Fear: Is America Emboldening China and North Korea?” (Washington, D.C.: Heritage Foundation, July 17, 2014), http://www.heritage.org/research/commentary/2014/7/asias-big-fear-is-america-emboldening-china-and-north-korea.

3. 

“Country Reports on Human Rights Practices: China” (Washington, D.C.: U.S. Department of State, February 23, 2000), http://www.state.gov/j/drl/rls/hrrpt/1999/284.htm. See also Secretary of State Hillary Clinton's remarks on a “new streak of assertive Chinese nationalism.” Clinton, “Inaugural Richard C. Holbrooke Lecture on a Broad Vision of U.S.-China Relations in the 21st Century” (Washington, D.C.: U.S. Department of State, 2011), http://www.state.gov/secretary/20092013clinton/rm/2011/01/154653.htm. See also Dianne Feinstein, “The Content of United States Engagement with China,” April 22, 1998, 105th Cong., 2nd sess., Congressional Record online, https://www.congress.gov/congressional-record/1998/04/22/senate-section/article/S3461-2.

4. 

See, for instance, Thomas L. Friedman, “China's Nationalist Tide,” New York Times, March 13, 1996.

5. 

For a preliminary effort to assess whether there is a hegemonic liberal world order that China seeks to challenge, see Alastair Iain Johnston, “China and International Order: Which China? Which Order?” paper presented at the conference “Negotiating the Future: Visions of Global Order,” German Institute of Global and Area Studies, Hamburg, Germany, December 3–4, 2015. See also Richard Fontaine and Mira Rapp-Hooper, “How China Sees World Order,” National Interest, April 20, 2016, http://nationalinterest.org/feature/how-china-sees-world-order-15846.

6. 

See Robert D. Kaplan, “Eurasia's Coming Anarchy: The Risks of Chinese and Russian Weakness,” Foreign Affairs, Vol. 95, No. 2 (March/April 2016), https://www.foreignaffairs.com/articles/china/2016-02-15/eurasias-coming-anarchy; Jian Zhang, “China's Growing Assertiveness in the South China Sea: A Strategic Shift?” in Leszek Buszynski and Christopher Roberts, eds., The South China Sea and Australia's Regional Security Environment, National Security College Occasional Paper No. 5 (Acton: National Security College, Australia National University, September 2013), http://nsc.anu.edu.au/documents/occasional-5.pdf; Michael Yahuda, “China's New Assertiveness in the South China Sea,” Journal of Contemporary China, Vol. 22, No. 81 (2013), pp. 446–459; Jiang Zongqiang and Hu Xin, “China's Dilemma: Nationalism Could Hijack Policy Response,” Straits Times, December 8, 2015, http://www.straitstimes.com/opinion/chinas-dilemma-nationalism-could-hijack-policy-response; and Sasa Petricic, “Forget the ‘Farce’ Bluster, China Received the Tribunal Ruling It Dreaded,” CBC News, July 12, 2016, http://www.cbc.ca/news/world/south-china-sea-ruling-1.3675714.

7. 

Author conversations with senior U.S. military officers, August 2015 and April 2016.

8. 

New research suggests that China has engaged in controlled diversionary conflict in reaction to elite, rather than mass, dissatisfaction with the state of the economy. The regime, however, tries to minimize the destabilizing effect of these actions on relations with other states. See Erin Baggott, “Four Essays in Computational Approaches to Autocratic Politics,” Ph.D. dissertation, Harvard University, 2016.

9. 

For formative statements, see S.J. Ball-Rokeach and M.L. Defleur, “A Dependency Model of Mass Media Effects,” Communication Research, Vol. 3, No. 1 (January 1976), pp. 3–21; and Maxwell E. McCombs and Donald L. Shaw, “The Agenda Setting Function of Mass-Media, Public Opinion Quarterly, Vol. 36, No. 2 (Summer 1972), pp. 176l–187.

10. 

The following is based mainly on an analysis of the twenty-eight news media articles and opeds in LexisNexis from 1996 through to September 2016 that refer to “Chinese nationalism” as “rising.”

11. 

Much of the scholarly literature that refers uncritically to a “rising” popular nationalism also relies on unsystematically collected evidence. See Li Liqing, “China's Rising Nationalism and Its Forefront: Politically Apathetic Youth,” China Report, Vol. 51, No. 4 (November 2015), pp. 311–326.

12. 

Vanessa Fong, “Filial Nationalism among Chinese Teenagers with Global Identities,” American Ethnologist, Vol. 31, No. 4 (November 2004), pp. 631–648.

13. 

William A. Callahan, “History, Identity, and Security: Producing and Consuming Nationalism in China,” Critical Asian Studies, Vol. 38, No. 2 (2006), pp. 179–208.

14. 

Tang Wenfang and Benjamin Darr, “Chinese Nationalism and Its Political and Social Origins,” Journal of Contemporary China, Vol. 21, No. 77 (2012), pp. 811–826.

15. 

Brian C. Rathbun, “Chinese Attitudes toward Americans and Themselves: Is There a Relationship?” in Alastair Iain Johnston and Shen Mingming, eds., Perception and Misperception in American and Chinese Views of the Other (Washington, D.C.: Carnegie Endowment for International Peace, 2015), pp. 9–21.

16. 

Peter Hayes Gries, “Tears of Rage: Chinese Nationalism and the Belgrade Embassy Bombing,” China Journal, July 2001, pp. 25–43.

17. 

James Leibold, “More Than a Category: Han Supremacism on the Chinese Internet,” China Quarterly, September 2010, pp. 539–559.

18. 

See James Reilly, Strong Society, Smart State: The Rise of Public Opinion in China's Japan Policy (New York: Columbia University Press, 2012); Jessica Chen Weiss, Powerful Patriots: Nationalist Protest in China's Foreign Relations (Oxford: Oxford University Press, 2014); Hao Yufan and Lin Su, eds., Zhongguo waijiao juece: kaifang yu duoyuande shehui yinsu fenxi [Chinese foreign policy making: Societal forces in Chinese foreign policy making] (Beijing: Chinese Academy of Social Sciences, 2007); and Wang Jun “Wangluo minzuzhuyi yu Zhongguo waijiao” [Internet nationalism and China's diplomacy] (Beijing: Chinese Academy of Social Sciences, 2011).

19. 

Zhu Tianbiao, “Nationalism and Chinese Foreign Policy,” China Review, Vol. 1, No. 1 (Fall 2001), pp. 1–27.

20. 

Jia Qingguo, “Disrespect and Distrust: The External Origins of Contemporary Chinese Nationalism, Journal of Contemporary China, Vol. 14, No. 42 (February 2005), pp. 11–21.

21. 

Yu Zhiyuan and Zhao Dingxin, “Differential Participation and the Nature of a Movement: A Study of the 1999 Anti-U.S. Beijing Student Demonstrations,” Social Forces, Vol. 84, No. 3 (March 2006), pp. 1755–1777. Yu and Zhao's analysis also underscores why using the frequency of demonstrations against foreign targets (Japanese or American symbols, property, or both) may not be a valid measure of nationalism across time. The motivations for participation can vary. Even assuming, however, that a primary motive is nationalism, there is a potential selection problem: people with high levels of nationalism are more likely to participate in these kinds of demonstrations than those with lower levels of nationalism. Counting only the actions of those with high levels of nationalism ignores those who do not participate, possibly because of their lower levels of nationalism. This possibility biases estimates of levels of nationalism upward. In addition, demonstrations are often reactions to some exogenous event. Thus, overall societal levels of nationalism could be constant, or even declining, but the exogenous shock may be sufficient nonetheless to mobilize small numbers of those with high levels of nationalism. The frequency of demonstrations may not reflect rising nationalism as much as the changing frequency of exogenous events.

22. 

Allen Carlson, “A Flawed Perspective: The Limitations Inherent within the Study of Chinese Nationalism,” Nations and Nationalism, Vol. 15, No. 1 (January 2009), p. 25.

23. 

Li, “China's Rising Nationalism and Its Forefront.”

24. 

Carlson, “A Flawed Perspective,” p. 25.

25. 

The BAS is administered by the Research Center for Contemporary China at Peking University. It began in 1995 and was modeled on the Detroit Area Study. The survey is mainly focused on attitudes toward city and local governance performance. Over the years, however, it has included some questions about national government performance, consumer confidence, political reform, and foreign policy. Since 1998 I have collaborated with the designers of the BAS in fielding questions about foreign policy and international affairs. From 1995 to 2007, the BAS sampled Beijing residents from the main urban districts, using probability proportional to size sampling procedures. Since 2007 it has used GPS sampling so as to increase the sample size and to include urban and rural registered residents of Beijing and unregistered residents living in the larger municipal area. On BAS sampling, see Hao Hongsheng, “The Sampling Design and Implementation for the 1995 Beijing Area Study,” Peking University, 1996; and Pierre F. Landry and Shen Mingming, “Reaching Migrants in Survey Research: The Use of the Global Positioning System to Reduce Coverage Bias in China,” Political Analysis, Vol. 13, No. 1 (Winter 2005), pp. 1–22. The BAS data used in this study are freely available from me for academic research purposes. As far as I am aware, the BAS is the only available time-series survey of ordinary Chinese people's views of a range of foreign policy issues such as military spending, foreign aid, trade preferences, amity, and nationalism. The Japanese nongovernmental organization, Genron NPO, has gathered times-series data on Chinese views of Japan. Although the Genron NPO survey includes detailed and useful questions that address the constitutive elements of these views, the primary attitudinal question concerns favorability—a vague concept that is hard to interpret and that does not measure nationalist sentiments. For the Genron NPO studies, see http://www.genron-npo.net/en/opinion_polls/. The Pew Research Center has also asked favorability questions across several years. Its Global Trends and Attitudes survey of Chinese opinion of Japan includes two separate years (2006 and 2016) with questions about stereotypes of the Other (in this case, Japanese people). The stereotype questions, however, do not ask about stereotypes of Self, so it is not possible to generate perceptions of difference and therefore measures of ethnocentrism. See Bruce Stokes, “Hostile Neighbors: China vs. Japan” (Washington, D.C.: Pew Research Center, September 2016), http://www.pewglobal.org/2016/09/13/hostile-neighbors-china-vs-japan/.

26. 

On the distinction between patriotism and nationalism, see Theodore W. Adorno et al., The Authoritarian Personality (New York: Harper and Row, 1950); Seymour Feschbach, “Individual Aggression, National Attachment, and the Search for Peace: Psychological Perspectives,” Aggressive Behavior, Vol. 13, No. 5 (1987), pp. 315–325; and Daniel Druckman, “Nationalism, Patriotism, and Group Loyalty: A Social Psychological Perspective,” Mershon International Studies Review, Vol. 38, No. 1 (April 1994), pp. 43–68. On the distinction between patriotism and nationalism in the Chinese context, see Peter Hays Gries et al., “Patriotism, Nationalism, and China's U.S. Policy: Structures and Consequences of Chinese National Identity,” China Quarterly, March 2011, pp. 1–17.

27. 

As Tom W. Smith and Seokho Kim note, pride in nation is an element in both patriotism and nationalism. See Smith and Kim, “National Pride in Comparative Perspective: 1995/96 and 2003/04,” International Journal of Public Opinion Research, Vol. 18, No. 1 (Spring 2006), pp. 127–136.

28. 

On measures of nationalism that tap into these three concepts, see Thomas Blank and Peter Schmidt, “National Identity in a United Germany: An Empirical Test with Representative Data,” Political Psychology, Vol. 24, No. 2 (June 2003), pp. 289–312.

29. 

These types of questions are asked, for example, in all three waves (1995, 2003, and 2013) of the International Social Survey Programme (ISSP) surveys on national identity. For the 2013 survey, see http://www.issp.org/uploads/editor_uploads/files/2013_final.doc, p. 3. The ISSP is widely used in studies of nationalism. In some years, BAS replaced “government” with “country” in the third question. In 2003 we ran a survey experiment to see if there was a difference in responses depending on which term was used. There was no statistically significant difference in the distribution of responses using these two different wordings. The third nationalism question was not asked in 2013. In addition, in 2002 and 2003 the BAS offered a “neutral” option between the “agree” and “disagree” choices. This was removed beginning in 2004. Unless otherwise indicated, reported results exclude “don't knows” and “no answers.”

30. 

In the 2015 BAS data, for instance, there is a strong relationship between desire to be a citizen of China and nativist blind support for the state (X2=821.99 p=0.000, N=2381). There is also a strong association between valuation of China as a country and blind support for the state (X2=778.9 p=0.000, N=2357). Finally, there is a strong association between desire to be a citizen and valuation of China as a country (X2 =2800 p=0.000, N=2389). There is also a strong linear relationship between preferring to be a citizen of China and negative perceptions of Japanese (ANOVA F=65.90 p=0.000, N=2040) as well as between the positive valuation of China as a country and negative perceptions of Japanese (ANOVA F=88.01 p=0.000, N=2023). In this regard, like Karl Gustaffson and Richard K. Herrmann, Pierangelo Isernia, and Paolo Segatti, I believe that the distinction between patriotism and nationalism is blurred in practice because of the complex related sentiments that link perceptions of the in-group and perceptions of the out-group. See Gustaffson, “The Struggle over the Meaning of Chinese Patriotism in the 21st Century,” China: An International Journal, Vol. 14, No. 3 (Autumn 2016), p. 138; and Herrmann, Isernia, and Segatti, “Attachment to the Nation and International Relations: Dimensions of Identity and Their Relationship to War and Peace,” Political Psychology, Vol. 30, No. 5 (October 2009), p. 725. The BAS shows that there is a significant correlation between a positive pride in country and more exclusionary ethnocentric or identity difference beliefs.

31. 

The main antonyms used were “peaceful-warlike,” “civilized-uncivilized,” “moral-immoral,” “modest-arrogant,” and “sincere-hypocritical.” Not all surveys asked about all five dimensions. Although the relevant identity difference for an individual need not be a measure of nationalism, when the concept is used to measure perceptions of ethnonational difference, it captures the degree to which people believe that their national traits are better than those of other nations.

32. 

See Edward D. Mansfield and Diana C. Mutz, “Support for Free Trade: Self-Interest, Sociotropic Politics, and Out-Group Anxiety,” International Organization, Vol. 63, No. 3 (July 2009), pp. 425–457; and Donald R. Kinder and Cindy D. Kam, Us against Them: Ethnocentric Foundations of American Opinion (Chicago: University of Chicago Press, 2010). For a summary of the social neuroscience microfoundations for the link among identity difference, fear, and preferences for relative gains, see Han Dongfang et al., “Nationalism, Ethnocentrism, and Trade Preferences: Findings from Beijing,” People's University of China, 2015.

33. 

Michel Oksenberg, “China's Confident Nationalism,” in “America and the World,” special issue, Foreign Affairs, Vol. 65, No. 3 (1987), pp. 501 n. 523; Allen S. Whiting, “Chinese Nationalism and Foreign Policy after Deng,” China Quarterly, June 1995, pp. 295–316; and Erica Strecker Downs and Phillip C. Saunders, “Legitimacy and the Limits of Nationalism: China and the Diaoyu Islands,” International Security, Vol. 23, No. 3 (Winter 1998/99), pp. 114–146; and Zhao Suisheng, A Nation State by Construction: Dynamics of Modern Chinese Nationalism (Stanford, Calif.: Stanford University Press, 2006).

34. 

The categories of Chinese nationalism in the extant literature tend not to be well theorized, often do not meet the conditions of a typology (i.e., mutually exclusive and exhaustive), and frequently do not have clearly operationalized indicators that allow replication.

35. 

I thank Tang Wenfang for graciously sharing his 2008 data with me. Tang's study was also conducted by the Research Center for Contemporary China, and therefore used questions about pride in citizenship, valuation of China, and blind support for the government identical to those used in the BAS.

36. 

See the data provided by the Beijing Municipal Human Resources and Social Security Bureau, http://www.bjrbj.gov.cn/bmfw/ywml/201601/t20160112_55858.html.

37. 

In terms of gender balance, the percentage of females in the 1998–2004 BAS averaged 48.8 percent per wave. Across the 2007–15 waves, this figure was 50.7 percent. From 1998 to 2004, respondents had an average of 11.3 years of education per wave, whereas the average for 2007–13 was 11.9 years per wave. The average age in the 1998–2004 BAS was 43.2 years old per wave. The figure for 2007–13 was 44.5 years old. From 1998 to 2009, the BAS cut off the sampling at age 70, whereas from 2013 to 2015 the maximum age was 90. Thus, in the calculation of the average ages I excluded those older than 70 from the last two years. In the analysis below, I do not cut off the sample at 70 for 2013 and 2015. If one does cut it off, the percentage of the “strong agrees” for all the nationalism questions declines in 2013 and 2015, and my conclusions about declining trends in nationalism after 2009 are strengthened.

38. 

As noted earlier, in the 2002 and 2003 BAS these three nationalism questions included a neutral option between the two “agree” and two “disagree” options. From 2004 onward, this neutral option was removed. It is hard to infer how nationalistic these neutral respondents are and what effect excluding them has on nationalism trends. Thus, the percentages in figures 13 report only the “agree” and “disagree” numbers. One method of inference is to look at how neutral respondents in one nationalism question answered other nationalism questions and use these answers to interpolate their response to the first question. This method shows that in 2002 and 2003 a large majority (between 60 percent and 77 percent) of those who gave neutral responses on the “China is a better country” question and on the “support your government/country even when it is wrong” questions somewhat agreed or strongly agreed that “I prefer to be a Chinese citizen.” This means that if the neutral respondents on the first nationalism question were recoded on the basis of their answers to the second and third nationalism questions, it is likely that the levels of agreement with the nationalism questions in 2002 and 2003 would have been even higher than they are in figures 13. Put differently, this method of inference means that the slope of the trend line over the 2000s would likely show even lower levels of nationalism over time.

39. 

In the 2008 nationwide survey conducted by Tang Wenfang and the Research Center on Contemporary China at Peking University, 55 percent of respondents strongly agreed with this statement. In the 2007 BAS—the survey conducted closest in time to Tian's 2008 survey—the proportion was 56 percent.

40. 

In Tian Wenfang's 2008 national sample, the proportion strongly agreeing with this question was 36 percent; in the 2007 BAS, it was also 36 percent.

41. 

In Tang's national sample, the proportion strongly agreeing with this question was 12 percent, and the proportion in the 2007 BAS was 13 percent.

42. 

If one combines the “strongly agree” and “somewhat agree” responses, the variation across time is similar, except that the total percentage that agrees moves slightly higher in the 2015 survey compared with the 2013 survey. This increase, however, is accounted for by a large increase in “somewhat agree” responses over the previous survey, and thus arguably reflects a dilution of nationalism compared to 2013. I find, however, that this kind of aggregation is of limited use, as it assumes that the views represented by “strongly agree” and “somewhat agree” are closer together than, say, the views represented by “somewhat agree” and “somewhat disagree.” I also examined the quadratic trend lines in the percentage of respondents who strongly agree with the three nationalism statements. In all three, nationalism appears to have peaked in the late 2000s; and on two of the questions, the percentage in 2015 is actually lower than in 2002. In the case of the statement about China being better than most countries, where the percentage strongly agreeing is slightly higher in 2015 than in 2002, this change may reflect a realistic comparative judgment about improved standards of living across time.

43. 

See Charles E. Osgood, George J. Suci, and Percy H. Tannenbaum, The Measurement of Meaning (Urbana: University of Illinois Press, 1957).

44. 

Kinder and Kam, Us against Them; and Mansfield and Mutz, “Support for Free Trade.”

45. 

The data used in these figures are averages across all traits in each survey. The number of traits in the questionnaire expanded over time, so the trends for individual traits may look slightly different than the average. In addition, because of the change in survey sampling in 2007, I include in these figures only the data for urban Beijing residents, so as to ensure as much consistency as possible in the samples across time. Interestingly, James Reilly cites data from a joint Japan-China survey of attitudes toward Japanese that also asks questions about perceptions of identities along a peaceful-warlike dimension only. He reports that the percentage of respondents who agree that Japanese are peaceful increased from 2006 to 2010 (though he combines “strongly agree” and “somewhat agree” responses, so it is hard to tell whether most of this change is a function of shifts from “strongly agree” to “somewhat agree” or from “somewhat disagree” to “somewhat agree”). See Reilly, Strong Society, Smart State, pp. 197–198.

46. 

Amity tends to be more volatile than identity difference and is likely to be more affected by short-term shocks in the interstate relationship. Nevertheless, in the BAS there is a moderate negative correlation between levels of amity toward a country and perceptions of identity difference with the people of that country. In other words, feelings of amity decline as perceptions of difference increase. In the full 2015 BAS that includes those with and without Beijing residency permits, the Pearson r between perceptions of identity difference with Americans, on the one hand, and amity toward the United States, on the other, is − .47 (p=0.000), and for Japanese and Japan the Pearson r is −.58 (p=0.000).

47. 

Moreover, measures of amity capture one version of the “rising nationalism” meme, namely, that popular hostility toward countries such as the United States and Japan is rising.

48. 

The Patriotic Education Campaign was launched in the early 1990s in response to the student and worker demonstrations against Communist Party rule in 1989. The campaign was designed to inoculate Chinese youth against so-called Western values by playing up historical memories of China's past humiliation at the hands of imperialists. See Zhao Suisheng, “A State-led Nationalism: The Patriotic Education Campaign in Post-Tiananmen China,” Communist and Post-Communist Studies, Vol. 31, No. 3 (September 1998), pp. 287–302; Zheng Wang, Never Forget National Humiliation: Historical Memory in Chinese Politics and Foreign Relations (New York: Columbia University Press, 2012); and William A. Callahan, China: The Pessoptimist Nation (Oxford: Oxford University Press, 2010).

49. 

To obtain as representative a sample as possible, I used the full dataset for figures 1113, including respondents with urban, rural, Beijing, and non-Beijing residency permits. Doing so limits the years for the analysis to 2007–15. In ten of the eleven total comparisons between those who had experienced the PEC and older generations in this time period, the differences between the generations were statistically significant. I also looked at the same data for the pre-2007 period, and the findings are consistent with the later period.

50. 

Ying Jiang, Cyber-nationalism in China: Challenging Western Media Portrayals of Internet Censorship in China (Adelaide, Australia: University of Adelaide Press, 2012), http://www.oapen.org/search?identifier=560092; and Sandra Valeri Constantin, “The Post-80s Generation in Beijing: Collective Memory and Generational Identity,” International Journal of Area Studies, Vol. 8, No. 1 (2013), p. 22.

51. 

For maximum representativeness, I used the full BAS 2015 dataset. The size of the 1990s generation in the sample declines rapidly once one moves back in time to earlier waves.

52. 

I should also note that there is a significant general correlation between age and perceived identity difference with Japanese (r=.30 p=0.000 N=2138) and with Americans (r=.27 p=0.000 N=2051). That is, the older the respondent, the greater the perceived identity difference.

53. 

Intergenerational difference in identity difference with Americans, t= −2.96 p=0.00, two tailed.

54. 

Intergenerational difference in identity difference with Japanese, t=−1.67 p=0.1, two tailed.

55. 

47.35 years old versus 37.71 years old (t=9.98 p=0.00, two-tailed).

56. 

Comparing the older generation with the generation born in the 1980s, X2=8.93 p=0.003. Comparing the older generation with the generation born in the 1990s, X2=12.57 p=0.00.

57. 

Comparing the older generation with the generation born in the 1980s, X2=3.31 p=0.07. Comparing the older generation with the generation born in the 1990s, X2=3.07 p=0.08.

58. 

Here X2=0.29 p=0.59.

59. 

These findings about youth and nationalism are consistent with the findings in Tang and Darr, “Chinese Nationalism and Its Political and Social Origins,” p. 820; and in Robert Hoffmann and Jeremy Larner, “The Demography of Chinese Nationalism: A Field-Experimental Approach,” China Quarterly, March 2013, p. 198. Andrew Chubb also suggests that youth are less supportive of more militarized policies in the South China Sea. See Chubb, “Exploring China's ‘Maritime Consciousness': Public Opinion on the South and East China Sea Disputes” (Perth, Australia: Perth USAsia Center, 2014), p. 11.

60. 

In the 2000 census, 17.5 percent of residents in Beijing municipality had technical-level or university-level education or above. By 2005 that figure had increased to 25.5 percent; and by 2010, it had reached 33.5 percent. In the 2000 BAS, only 4.4 percent of the sample was born in the 1980s. By 2015 this figure had risen to 24 percent.

61. 

Tang and Darr, “Chinese Nationalism and Its Political and Social Origins,” p. 816.

62. 

Peter Ford, “China Targets ‘Hostile Foreign Forces’ in Crescendo of Accusations,” Christian Science Monitor, November 9, 2014, http://www.csmonitor.com/World/Asia-Pacific/2014/1109/China-targets-hostile-foreign-forces-in-crescendo-of-accusations; Edward Wong, “In New China, ‘Hostile’ West Is Still Derided,” New York Times, November 11, 2014, http://www.nytimes.com/2014/11/12/world/asia/china-turns-up-the-rhetoric-against-the-west.html?_r=0; and Edward Wong and Mia Li, “Q. and A.: Zhang Lifan on the New Chinese Nationalism,” New York Times, November 27, 2014, http://sinosphere.blogs.nytimes.com/2014/11/27/q-and-a-zhang-lifan-on-the-new-chinese-nationalism/. See also the reference to this claim in Fareed Zakaria, “China's Growing Clout,” fareedzakaria.com, November 14, 2014, https://fareedzakaria.com/2014/11/14/chinas-growing-clout/.

63. 

For example, in a forced public confession, a Chinese dissident criticized efforts by “hostile forces” to wage ideological war against China. See Chris Buckley, “Week of TV Trials in China Signals New Phase in Attack on Rights,” New York Times, August 5, 2016, http://www.nytimes.com/2016/08/06/world/asia/china-trial-activists-lawyers.html.

64. 

I would like to thank Alice Miller for pointing out the People's Daily's importance in this regard.

65. 

When I was studying at Peking University early in the Deng Xiaoping era, I witnessed firsthand the use of anti-foreign and anti-Western propaganda during the short-lived anti-spiritual pollution campaign in 1983–84. During the campaign, many so-called lifestyle practices (in those days, these included disco dancing, Western suits, and long hair on males, among other expressions of “spiritual pollution” were arbitrarily banned or suppressed far more rigorously than any “foreign” lifestyle practices have been under Xi.

66. 

Xi has, however, used ideological purity and unity as a way of trying to compel Party members to line up behind his leadership and, most crucially, to improve their standards of behavior and rectitude so as to shore up the quality and legitimacy of Party rule. Campaigns against ideological slackness inside the Party and the cultivation of nationalism outside the Party, however, are not the same. Some of the more simplistic descriptions of ideological tightening under Xi tend to blur the differences.

67. 

This suggests that, like much of Xi's ideological messaging, the renewed focus on military obedience to the Party began in the Hu Jintao era. Xi, however, has magnified its importance even further.

68. 

See the analysis by Chinese journalist Gao Yu of an internal speech allegedly given by Xi in December 2012, in Gao, “Nan er Xi Jinping” [Xi Jinping's “The Man”], Deutsche Welle (Chinese ed.), December 25, 2013, http://www.dw.com/zh/%E7%94%B7%E5%84%BF%E4%B9%A0%E8%BF%91%E5%B9%B3/a-16549520. Excerpts from Gao's analysis were translated as “Leaked Speech Shows Xi Jinping's Opposition to Reform,” China Digital Times, January 2013, http://chinadigitaltimes.net/2013/01/leaked-speech-shows-xi-jinpings-opposition-to-reform/.

69. 

Nor has there been a major spike under Xi in the frequency of articles using the terms “hostile forces” in hundreds of other media outlets found on the China Digital News Library newspaper database. Even in the often hard-line Global Times, the monthly average per year of such articles under Xi does not reach the highest level under the Hu Jintao regime. Finally, in a unique database of more than 40,000 social media posts from the Internet Propaganda Office of a city district in Jiangxi Province dating from 2012 to 2014, none used the term “hostile forces,” suggesting that there was no official campaign at the grassroots levels either. My thanks to Molly Roberts for searching this database for this term.

70. 

For example, a summary of the discussion in the Politburo Standing Committee after the U.S. bombing of China's embassy in Belgrade in 1999 shows that China's highest leaders had internalized the long-time narrative of imperialist humiliation. See Wang, Never Forget National Humiliation, pp. 171–173.

71. 

As I noted in earlier research on the “assertive China” meme, “assertive” is an accurate description of China's coercive diplomacy in maritime disputes, but it is not easily generalizable to a range of other issues where Chinese policy is not changing as much as many pundits assume. See Alastair Iain Johnston, “How New and Assertive Is China's New Assertiveness?” International Security, Vol. 37, No. 4 (Spring 2013), pp. 19–20, 31, 41.

72. 

Television is still the primary source of news in China, and it is less independent from the state than are online information systems. See Chubb, “Exploring China's ‘Maritime Consciousness,'” pp. 10, 20.

73. 

This was the implication of comments by a senior Chinese official involved in the day-to-day foreign policy decisionmaking process whom I interviewed in 2011. This official noted that the more extreme the opinion, the less likely the regime was to take it into account, and the more the regime would try to guide and mold it. Over the last five years, I have interviewed a number of Chinese scholars and officials about models and mechanisms through which public opinion might influence the top leaders in China. Many, though not all, agree that a critical mechanism is elite political struggle. That is, potential competitors at the top might use negative public opinion to highlight the weakness or failure of the leader. To deprive competitors of this political tool, leaders have an interest in taking positions close to those of more nationalistic or hard-line publics. The attention to public opinion, therefore, does not come from fear of popular demonstrations per se; rather it comes mainly from the costs of losing in elite political competition. This suggests, in addition, that sensitivity to public opinion may be cyclical, peaking around periods of political succession. In short, it is likely that Xi is more sensitive to elite than to popular opinion.

74. 

Indeed, as Reilly puts it, survey data alone are “not a sufficient indicator” for assessing the totality of public opinion, but they are “a necessary element.” See Reilly, Strong Society, Smart State, p. 29. For a helpful discussion on how much additional research and analysis is needed regarding the evolution, effect, and measurement of Chinese nationalism, see Anna Costa, “Focusing on Chinese Nationalism: An Inherently Flawed Perspective? A Reply to Allen Carlson,” Nations and Nationalism, Vol. 20, No. 1 (2014), pp. 93–112.

75. 

As one model of this inductive process, see Ted Hopf and Bentley B. Allan, eds., Making Identity Count: Building a National Identity Database (New York: Oxford University Press, 2016). For another example that examines an earlier period of Chinese nationalism, see Callahan, “History, Identity, and Security,” pp. 179–208.

76. 

For example, in a crisis with Japan over disputed islands, strong nationalists may be less likely to punish leaders who back down after threatening to use force if they instead threaten to use economic sanctions. They may, however, be more likely to punish leaders who back down in the face of U.S. deterrence threats. These are findings from a survey experiment that Hong Kong University professor Kai Quek and I conducted in China in 2015 on the effect of different rhetorical and behavioral strategies that Chinese leaders could employ to minimize loss of public support if they want to make concessions in territorial disputes with Japan. See Quek and Johnston, “Crisis Management in the Shadow of Audience Costs,” paper prepared for the annual meeting of the American Political Science Association, Philadelphia, Pennsylvania, September 1–4, 2016.

77. 

Since at least the Hu Jintao period, reports have circulated that the top leaders conduct straw polls of the Central Committee to gauge approval for changes in leadership lists developed by the Politburo Standing Committee. To the extent that the Central Committee serves as a rough selectorate for top positions, it may play an important role in pressuring or at least constraining the top leaders’ foreign policy choices on some issues.

78. 

“Xi: There Is No Gene for Invasion in Our Blood,” China Daily (U.S. ed.), May 16, 2015, http://usa.chinadaily.com.cn/china/2014-05/16/content_17511170.htm. The BAS data show that, like U.S. exceptionalist beliefs in the virtuousness of the United States as a “shining city on a hill,” Chinese exceptionalist beliefs about China's inherent peacefulness are linked to more hard-line foreign policy preferences. See Alastair Iain Johnston, “On Chinese Exceptionalism: The Paradox of a Peaceful People,” in Jennifer Rudolph and Michael Szonyi, eds., The China Questions (Cambridge, Mass.: Harvard University Press, forthcoming).