-
Assessing Popular Support for the Communist Party of China
In the absence of multiparty electoral competition, it is challenging to assess popular support for the Communist Party of China. Existing methods rely on unidimensional measures which are often reduced to a simple dichotomy of high or low support, therefore missing a measure for ambivalence. This study argues that party support in China has two dimensions. The specific dimension is trust in the Party, while the diffuse one is endorsement of one-party rule. The study therefore identifies three attitudes towards the Party—support, ambivalence and dissent— and concludes that popular support for the ruling party is weaker than the Party itself has claimed it to be.
The Summit for Democracy, held in Washington, DC on 9–10 December 2021, declared an ideological war against autocracy. Inevitably, the attack triggered a counterattack. Before the summit convened, the Communist Party of China (CPC), the unnamed but unmistakable primary target, launched a pre-emptive ideological defence. It reiterated that there are multiple equally legitimate models of democracy and that each nation has the right to adopt the model that fits its unique conditions. Going beyond the defensive, the CPC declared that its “whole-process democracy” is superior to the electoral democracy advocated by the US-led democratic alliance.1 Notably, the Party started to use favourable survey results to back up its claim of “enjoying the wholehearted support of the people”.2 For instance, a spokesperson from the Ministry of Foreign Affairs cited a Canadian research team’s finding that over 90 [End Page 51] per cent of the Chinese people trust the central government.3 But does the Party indeed enjoy overwhelming popular support in China?
The question is crucial because it determines whether the US-led alliance for democracy is targeting an autocratic regime specifically or is aiming inadvertently at the vast majority of Chinese people. The issue is hard to articulate because China is a non-democratic single-party state.4 The CPC rejects Western multiparty systems and allows no multiparty electoral competition. In the absence of a popular vote as a direct and valid indicator, researchers have to employ survey data on trust in the central government, support for the existing political system and endorsement of authoritarian values.5 However, existing measurements are invariably unidimensional, even when they use multiple indicators. Moreover, current assessments often end up with a simple high–low dichotomy that blots out the gradation of support. As a result, current studies yield implausible estimates. For instance, a research team reported that over 80 per cent of Chinese people trust the central government and support the regime despite the COVID-19 pandemic.6
This article proposes a two-dimensional measurement of popular support for the CPC. It proceeds as follows. First, it argues that support for the CPC has two related but distinctive dimensions. The specific dimension is trust in the Party, and the diffuse dimension is endorsement of one-party rule. Second, it shows that people can be grouped according to three typical attitudes towards the Party, i.e. support, ambivalence and dissent. Thirdly, it validates the measurement results by examining if they conform to theoretical expectations and empirical findings regarding sources of support for a political party. The study concludes that popular support for the CPC is considerably weaker than the Party claims. [End Page 52]
A TYPOLOGY OF PARTY SUPPORT
The substance of popular support for a political party is multidimensional and can include the embrace of ideology, endorsement of policy goals, satisfaction with performance, approval of leadership candidates and appreciation of institutional values.7 Nevertheless, the manifestation of support is explicit and unitary in countries with free and fair multiparty elections. People vote for parties they support, thereby voting against those they do not support, and abstain from voting if they feel ambivalent about all competing parties.8 Researchers can confidently treat the popular vote as a valid and reliable indicator of party support unless an election is independently verified as fraudulent. Thanks to its straightforward nature, measuring popular support for competing political parties is not a salient issue on the research agenda of party politics in established democracies.
However, the accepted measure of party support does not apply to authoritarian China. The Leninist ruling party claims to “represent the people’s interests”9 but refuses to accept the test of the popular vote. It allows eight “democratic parties” to exist but turns them into political ornaments.10 The CPC convenes the National People’s Congress (NPC) every five years to elect the national leadership, but elections of deputies to people’s congresses are selections with predetermined results. The election of the state chairman at the NPC is nothing but a political show because the post has already been filled in the preceding Party Congress.11 Even elections of deputies to local people’s congresses are under tight control. Only the ruling party and its affiliated mass organisations can nominate candidates, and “independent candidates” are almost [End Page 53] always eliminated in primaries.12 While elections to villagers’ committees are genuinely competitive, what happens there is “competition without parties”.13
Due to the absence of multiparty electoral competition, scholars cannot use popular votes as an indicator of popular support for the CPC. Instead, they have to resort to utilising survey data as proxies. Some researchers regard trust in the central government or national leadership as a proxy of support for the CPC.14 Other scholars treat acceptance of authoritarian values and pride in the political system as proxies.15
There are two problems to existing survey-based measurements of popular support for the CPC. First, they tend to rely on a unidimensional scale. Some researchers use a single ordinal scale measure, assuming that trust or support can be divided into four or five categories that can be meaningfully rank-ordered.16 Others use multiple indicators to construct unidimensional scales of support for the one-party regime and support for the existing political system.17 However, popular support for the CPC in China has two related but distinctive dimensions. In a multiparty system, support for a ruling party presupposes accepting the multiparty system. In China, however, trusting the CPC is one thing, while accepting one-party rule is another. Moreover, the two dimensions of party support overlap but are non-compensatory, i.e. the strength in one dimension does not always offset a possible weakness in the other.18 People may trust the CPC without favouring one-party rule. For instance, people may trust the CPC because they have confidence in the current party leadership, but they may not favour one-party rule because they worry about the corruptive effect of the power [End Page 54] monopoly. Meanwhile, people may distrust the CPC but favour one-party rule. For instance, ambitious individuals may distrust the CPC but want to establish their own versions of one-party rule. Like revolutionary leaders in the past, they may seek a dynastic change by occupying the throne rather than replacing a monarchy with democracy.
The second problem with survey-based measurements of popular support for the CPC is that they assume that people have clear-cut political attitudes. As a result, existing assessments often reduce complex attitudes to a dichotomy, therefore missing out mixed attitudes. For example, studies of trust in the central government often merge “strong trust and moderate trust” into “trust”, although national surveys typically use four-point ordinal scales.19 However, dichotomisation risks oversimplification and even distortion. As Wang and You point out, “strong trust” and “moderate trust” are qualitatively different and should not be confused with each other. People may neither trust nor distrust the CPC.20 Instead, they may have doubts.21 Assessments of system support also tend to end up with a dichotomy. For instance, Chen uses four ordinal measures to construct a continuous summation scale of regime support,22 but she reduces the scale to a dichotomy of high support vs. low support to simplify multivariate analysis. In reality, however, people may neither support nor reject the existing political system. Instead, they may be indifferent, apathetic, or cynical.
Relying on unidimensional measures and implicitly dichotomising complex attitudes risk missing ambivalence, an essential variant of political attitudes towards the CPC. Citizens in electoral democracies may express their ambivalence towards political parties by abstaining from voting. The absence of multiparty elections in the PRC does not mean that political attitudes towards the CPC and the one-party rule are distinctly either black or white. As Shue points out: “Most people, most of the time, I would suggest, are quite appropriately ambivalent about the legitimacy of the system in which they find themselves…The very experience of domination most often marries objection with acceptance. It is bivalent, and so people are ambivalent”.23
This study constructs a two-dimensional typology of party support. Applying a well-established conceptual distinction, it treats trust in the CPC as specific support and endorsement of one-party rule as diffuse support.24 Following the example of [End Page 55] earlier studies,25 the research attempts to consider both specific and diffuse support. Moreover, the research recognises that people typically have three attitudes towards the trustworthiness of the CPC and the legitimacy of one-party rule. On the one hand, they may trust, doubt or distrust the Party. On the other hand, they may endorse, feel neutral or reject one-party rule. Since most national surveys do not include a neutral category in their ordinal measures of attitude, this study adopts two refined dichotomies. While it argues that people may trust or doubt/distrust the party, it also argues that people may endorse or feel neutral/reject one-party rule. Cross-tabulating the two dimensions results in a typology of four attitudes towards the CPC (Table 1).
Types of Attitudes towards the Communist Party of China (CPC)
The two dimensions constitute two coherent positions. Trusting the CPC and endorsing one-party rule constitute support. On the contrary, distrusting the CPC and rejecting one-party rule constitute dissent. Straddling in-between are two variants of ambivalence— one is trusting the CPC without approving of one-party rule to the corresponding degree and the other is distrusting the CPC without disapproving of one-party rule to the corresponding level.
MEASURING PARTY SUPPORT
The study also draws on the third wave of the Asian Barometer Survey in 2011 to examine how many people are supportive, ambivalent or dissenting regarding the CPC. The survey covers a probability sample of 3,473 respondents drawn from 94 counties and urban districts in mainland China (for more information about the survey, see Asian Barometer Survey, 2011). It measures trust in the CPC with a four-point ordinal scale (Table 2).26 [End Page 56]
Trust in the CPC
The survey also measures the endorsement of one-party rule. In the section concerning the ideal form of government, the survey asks respondents which one of the following two propositions they agree with: (i) “many political parties representing various interests compete with each other”; (ii) “one political party represents the interests of all people”. The follow-up question is: “Do you agree or strongly agree with the statement?”. Since the two propositions are antitheses, agreeing with one is equivalent to disagreeing with the other. With “one political party represents the interests of all people” as the indicator statement, the four responses constitute an ordinal measure of acceptance of one-party representation as the ideal form of government, which is treated as a proxy for endorsement of one-party rule. Table 3 summarises the results.
Acceptance of One-party Representation
Respondents who are more confident in the CPC tend to accept the principle of one-party representation. However, the two ordinal measures do not constitute a reliable summation scale (Cronbach’s alpha = 0.34), indicating that the two dimensions are related but distinctive. The conventional method is to treat such moderately correlated measures as separate variables.27 However, this conventional treatment misses the fact that trust in the CPC and acceptance of one-party rule are complementary dimensions of party support. [End Page 57]
This study adopts a two-step measurement scheme, aiming for a classification that optimally fits the data and conforms with theoretical expectations.28 First, the study employs k-means clustering to identify respondents across the four attitudes identified in the typological analysis.29 It treats trust in the CPC as one feature vector and endorsement of one-party rule as another. In addition, it treats the two four-point ordinal measures as crude representations of two latent continuous scales ranging from one extreme through a middle point to another. In other words, the study recognises that some people neither trust nor distrust the CPC. Instead, they may have doubts. Meanwhile, the study recognises that some people neither approve nor disapprove of the principle of one-party rule. Instead, they may be neutral. Furthermore, the study recognises that some people do not have coherent positions on the two issues. They may trust the CPC without endorsing one-party rule and vice versa.
Based on these considerations, the study sets the number of clusters at four, although all three heuristics, i.e. the elbow method,30 the silhouette value31 and the gap statistic,32 suggest that the respondents form three clusters. Table 4 summarises the coordinates of the centroids of four clusters.33
Coordinates of Centroids of Four Clusters
Second, the study draws on domain knowledge to simplify the clustering results, adopting the established practice of incorporating background knowledge with [End Page 58] algorithms.34 Respondents in cluster 2 trust the CPC without endorsing one-party rule to the corresponding degree, while respondents in cluster 3 have doubts about the CPC without disapproving of one-party rule to the corresponding extent. The two groups are merged into a group labelled ambivalent. Table 5 summarises the coordinates of centroids of the three finalised clusters.
Centroids of the Three Representative Clusters
Of the 3,473 respondents, 46.2 per cent are supporters, averaging 4.00 on the four-point scale of trust in the CPC and 3.34 on the four-point scale of endorsement of one-party rule. The results contrast sharply with the conventional reading that 85 per cent of respondents trust the CPC and 77.5 per cent accept the principle of letting one political party represent the interests of all people. The implication is that conventional measures may overestimate popular support for the CPC.
Secondly, 42.5 per cent of respondents are ambivalent, averaging 2.95 on the four-point scale of trust in the CPC and 2.78 on the four-point scale of endorsement of one-party rule. The ambivalent group is highly mixed. Some respondents express lukewarm trust in the CPC and lukewarm endorsement of one-party rule; some trust the CPC but reject one-party rule; others distrust the CPC but endorse one-party rule. Conventional analysis misses this critical group entirely.
Thirdly, 11.3 per cent of respondents express dissent. They distrust the CPC, averaging 2.28 on the four-point scale. Meanwhile, they also reject one-party rule, averaging 1.88 on the four-point scale. Although they are a relatively small minority in the population, dissenters deserve special attention because they may play a pivotal [End Page 59] role in bringing about significant changes by functioning as a “committed minority” at particular historical junctures.35
SOURCES OF PARTY SUPPORT
The study seeks to validate the two-dimensional measurement of party support by examining whether the identified attitudes have expected correlations with factors that are theoretically expected to affect support for a political party. Popular support for political parties derives from three primary sources. First, citizens may embrace a party’s ideology, i.e. their understanding of how a society should be governed.36 Second, they may approve of a party’s performance on issues of high salience such as developing the economy, maintaining effective governance or fighting terrorism.37 Third, party support may also have cultural sources, e.g. values regarding political authorities.38
Researchers have applied these arguments to China. Some scholars treat party membership as an indicator of ideological persuasion. However, they reach different conclusions. Two analysts observe that Party members are more likely than the non-Party masses to trust the central government,39 but one research team observes that while Party members in their suburban sample have stronger trust in the central government than do non-members, Party membership has no effect on trust in the rural sample.40 Applying the exposure-acceptance model of political persuasion,41 Kennedy finds that respondents who are more exposed to government-run news media tend to have stronger confidence in the national leadership,42 while other scholars [End Page 60] observe that internet use tends to weaken trust in the Party-state.43 Nonetheless, researchers agree on the instrumental source of Party support. On the positive side, scholars find that people who are more satisfied with the country’s economy and governance have stronger support for the Party-state.44 In addition to economic growth, equalising social policies, such as education reform or poverty relief campaigns, enhance popular trust in the central government.45 Furthermore, satisfaction with the health system is positively associated with support for the political system.46 On the negative side, perception of corruption weakens trust in the Party-state,47 and air pollution lowers regime support.48 In addition, people with stronger traditional cultural values tend to have stronger trust in the central government.49
The study joins the exploration of possible sources of party support. It treats the ordinal measure of party support as the dependent variable. Predictors are (i) party membership; (ii) assessment of the national economy; (iii) evaluation of governance; (iv) perception of corruption among government officials; (v) patriarchal values; (vi) exposure to official media; (vii) internet use. Four demographic variables are controlled, including gender, age, education and household registration. The ordinal regression model passes the test of parallel regression assumption (chi-square =18.05; df = 11; p = 0.08). Table 6 summarises the results. [End Page 61]
Predicting Sources of Support for the CPC
The results conform with the theoretical expectations and existing findings, validating the two-dimensional measurement. First, Party membership has a positive effect on Party support since Party members are more likely to have stronger support for the Party than non-members. Moreover, Party members are more likely to be ambivalent than dissenting and more likely to be supportive than ambivalent. Although it is statistically significant, the effect of Party membership is however weaker than the Leninist party expects from its rank and file. The CPC demands total loyalty from its members.50 The study performs ideal-typical cases to further explore the effect size51 and finds that the effect of Party membership is underwhelming. Holding other variables at their means, being a Party member decreases the probability of dissenting by only 0.020 (declining from 0.096 to 0.076), decreases the probability of being [End Page 62] ambivalent by 0.044 (declining from 0.483 to 0.439), and increases the probability of being a supporter by only 0.064 (growing from 0.421 to 0.485) (Figure 1). Technically speaking, holding other predictors at their sample means, one standard deviation of increase in Party membership decreases the probability of dissenting by 0.008, decreases the probability of being ambivalent by 0.011 and increases the probability of supporting by 0.019. The observation is consistent with the finding that many people join the Party for privileged access to career opportunities instead of ideological persuasion.52 The weak Party loyalty may also derive from ordinary Party members’ lack of voice.53 Despite the central leadership’s repeated commitments to building intra-Party democracy, the CPC remains a strictly hierarchical political organisation in which “most ordinary Party members are not insiders”.54 Under Leninist democratic centralism, ordinary party members have practically no opportunity to exert any meaningful impact on selecting even low-ranking party leaders, let alone the central leadership.55 All ordinary citizens are practically disenfranchised, but the lack of voice may be particularly demoralising for Party members who are constantly being told that they are special and more advanced than ordinary people. The finding is even more interesting in light of a recent observation that Party members, on average, hold substantially more modern and progressive views than the public does on gender equality, political pluralism and openness to international exchange.56 Being more enlightened and having stronger internal efficacy, Party members are more likely to feel frustrated, if not alienated, by their lack of voice.
Second, the assessment of the national economy has a much larger impact on Party support. The survey asks respondents to assess, on a five-point scale, the current condition of the national economy, compare it with how it was five years ago and project how it will be in five years (see Appendix for details).
The study uses responses to three questions to construct a simple summation index (Cronbach’s alpha = 0.75). Confirming the observations of existing studies,57 the more positive the assessment of the national economy is, the stronger the support for the Party becomes (Figure 2). Holding other variables at their means, an individual who feels more positive about the national economy is more likely to be a supporter and less likely to be ambivalent or a dissenter. However, the effect of assessment of [End Page 63]
Effect of Party Membership on Party Support
Source: Asian Barometer Survey 2011.
Effect of the Assessment of the National Economy on Party Support
Source: Asian Barometer Survey 2011.
[End Page 64]
the national economy on Party support is consistently monotonous in terms of the probability of being ambivalent. As one’s assessment becomes more positive, the probability of being ambivalent first increases and then decreases. As the assessment of the national economy becomes increasingly positive, the probability of being supportive increases steadily and remarkably. Meanwhile, the likelihood of being ambivalent or dissenting declines correspondingly, indicating that individuals with similar assessments of the national economy are equally likely to be ambivalent or dissenting. Holding all other characteristics at their means, compared with a mostly negative sentiment about the national economy, a mostly positive one decreases the probability of dissenting by 0.414 (declining from 0.463 to 0.049), decreases the probability of being ambivalent by 0.100 (declining from 0.456 to 0.356), and increases the probability of being supportive by 0.514 (growing from 0.081 to 0.595). Technically speaking, holding other predictors at their sample means, one standard deviation of increase of satisfaction with the national economy decreases the probability of dissenting by 0.036, decreases the probability of being ambivalent by 0.057, and increases the probability of supporting by 0.093. The implication is that the effect size of satisfaction with the national economy on Party support is over four times larger than that of Party membership. The observation corroborates the consensus that satisfaction with the country’s economic development substantially strengthens trust in the Party-state.
Third, assessment of governance is expected to affect support for the CPC because it is fused with the government.58 The survey includes three measures of governance assessment: (i) the government treats everybody fairly; (ii) the government treats the rich and the poor equally; (iii) the government ensures that everyone has enough food, clothing, shelter and other life necessities. The three four-point indicators constitute a reliable summation scale of governance satisfaction (Cronbach’s alpha = 0.76). Satisfaction with governance has a significant and symmetric effect on Party support (Figure 3). Instead of being ambivalent, individuals more satisfied with governance are more likely to be supportive and less likely to be ambivalent or dissenting. Holding all other variables at their sample means, compared with a totally dissatisfied person, an individual who feels fully satisfied has a higher probability of being a supporter (increasing from 0.273 to 0.549), a lower probability of being ambivalent (decreasing from 0.556 to 0.391) and a lower probability of being a dissenter (decreasing from 0.171 to 0.060). Technically speaking, holding other predictors at their sample means, one standard deviation of increase in satisfaction with governance decreases the probability of dissenting by 0.018, decreases the probability of being ambivalent by 0.025 and increases the probability of being supportive by 0.043. The observation corroborates the finding that satisfaction with governance increases popular trust in the central government.59 [End Page 65]
Fourth, individuals who perceive more severe political corruption tend to have weaker support for the CPC. The survey asks respondents to assess, respectively, the extent of corruption among central government officials and local government officials on a four-point scale: (1) almost nobody is corrupt; (2) not many are corrupt; (3) most officials are corrupt; (4) almost everyone is corrupt. The two indicators constitute a reliable summation scale (Cronbach’s alpha = 0.69). Corruption drains support for the ruling party (Figure 4). Perception of corruption has a significant and symmetric effect. Instead of being ambivalent, individuals who perceive more widespread corruption are less likely to be supportive and more likely to be dissenting. Holding all other variables at their sample means, compared with an individual who believes almost no government official is corrupt, an individual who suspects that almost all officials are corrupt has a lower probability of being a supporter (decreasing from 0.599 to 0.254), a higher probability of being ambivalent (increasing from 0.352 to 0.561) and a higher probability of being a dissenter (increasing from 0.049 to 0.185). Technically speaking, holding other predictors at their sample means, one standard deviation of increase in the perception of corruption increases the probability of dissenting by 0.034, increases the probability of being ambivalent by 0.033 and decreases the probability of being supportive by 0.067. This observation corroborates the finding that people who perceive more widespread corruption tend to have weaker confidence in the central leadership’s commitment and capacity to combat corruption.60
Fifth, adherence to patriarchal values is expected to affect Party support because patriarchalism favours hierarchical order and concentration of power.61 Among 3,473 respondents, 11.32 per cent strongly agree that “the government is like our parents and should make decisions for us”, 38.30 per cent agree, 36.22 per cent disagree and 14.17 per cent strongly disagree. Generally speaking, adherence to patriarchal values increases support for the ruling party. Holding all other variables at their sample means, compared with one who strongly rejects patriarchal values, an individual who firmly adheres to them has a higher probability of being a supporter (increasing from 0.369 to 0.500), a lower probability of being ambivalent (decreasing from 0.514 to 0.429) and a lower probability of being a dissenter (decreasing from 0.117 to 0.072). Technically speaking, holding other predictors at their sample means, one standard deviation of increase of adherence to patriarchal values decreases the probability of dissenting by 0.014, decreases the probability of being ambivalent by 0.019 and increases the probability of being supportive by 0.033. This observation corroborates the consensus that people with more robust traditional cultural values tend to have higher trust in the central government and central Party leadership.62
Sixth, the study shows that media exposure significantly affects Party support. On the one hand, a respondent who watches or listens to political news from government-controlled media is more likely to be a supporter and less likely to be [End Page 66]
Effect of Assessment of Governance on Party Support
Source: Asian Barometer Survey 2011.
Effect of the Perception of Corruption on Party Support
Source: Asian Barometer Survey 2011.
[End Page 67]
ambivalent or a dissenter. Holding all other variables at their sample means, compared with one who is least exposed to official media, an individual who is most exposed has a higher probability of being a supporter (increasing from 0.394 to 0.455), a lower probability of being ambivalent (decreasing from 0.500 to 0.460) and a lower probability of being a dissenter (decreasing from 0.106 to 0.085). Technically speaking, holding other predictors at their sample means, one standard deviation of increase of exposure to official media decreases the probability of dissenting by 0.007, decreases the probability of being ambivalent by 0.009 and increases the probability of supporting by 0.016. This finding corroborates the view that ideological indoctrination through mass media strengthens trust in the Party-state.63
Seventh, the study shows that internet use significantly affects Party support. Compared with an individual who does not use the internet, a respondent who accesses it several times a day is less likely to be a supporter and more likely to be ambivalent or a dissenter. Holding all other variables at their sample means, compared with one who does not use the internet, an individual who accesses it most frequently has a lower probability of being a supporter (decreasing from 0.451 to 0.382), a higher probability of being ambivalent (increasing from 0.463 to 0.507) and a higher probability of being a dissenter (increasing from 0.086 to 0.111). Technically speaking, holding other predictors at their sample means, one standard deviation of increase of internet use increases the probability of dissenting by 0.012, increases the probability of being ambivalent by 0.013 and decreases the probability of being supportive by 0.025. The finding supports the argument that access to uncensored political information may weaken trust in the authoritarian regime.64
There are three minor points worth mentioning. First, older people tend to have stronger support for the Party, consistent with findings in other countries where older people are politically more conservative.65 Second, school education has no effect on the support for the Party; this discovery is inconsistent with an earlier study’s finding that people with more years of education are more supportive of the Party-state.66 Given the importance of school education and the constantly changing school curriculum,67 the effect of school education on Party support merits further investigation. Third, respondents with a rural household registration tend to support the Party more strongly. The effect is marginal but nonetheless puzzling because the household [End Page 68] registration system discriminates against the rural population and favours urban residents.68
CONCLUSIONS
The study argues that support for the CPC has two distinctive dimensions. The specific dimension is trust in the Party and the diffuse dimension is endorsement of one-party rule. Furthermore, the study argues that attitudes regarding each dimension constitute a trichotomy instead of a dichotomy. The two-dimensional measurement shows that popular support for the CPC is significantly weaker than existing studies suggest. Measured by unidimensional measures, 85 per cent of 3,473 respondents in the Asian Barometer Survey in 2011 trust the CPC, and 77.5 per cent approve of the principle of letting one political party represent the interests of all people. However, measured by the two-dimensional typology, 46.2 per cent are supportive, 42.5 per cent ambivalent and 11.3 per cent dissenting. The two-dimensional measurement is valid in that its results conform with theoretical expectations and empirical findings. Stronger support for the CPC is associated with a positive assessment of the national economy, governance and corruption control, and vice versa. In addition, stronger Party support is associated with more exposure to ideological indoctrination, less access to uncensored information on the internet, stronger traditional cultural values, and vice versa. Popular support for the CPC is high by any standard but not as high as the Leninist party claims it to be.
Conceptually, the research highlights the importance of distinguishing support for a political party from endorsing the underlying party system. In an electoral democracy, support for a political party usually entails accepting the prevailing party system. Therefore, it is generally legitimate to treat the popular vote as an indicator of party support. However, trusting the CPC does not entail accepting one-party rule. Consequently, it is critical to regard trust in the CPC and endorsement of one-party rule as related but distinctive dimensions of party support. Relying on a unidimensional scale tends to result in the overestimation of support.
Methodologically, the study reaffirms the value of integrating typological analysis with k-means clustering when measuring complex political attitudes. People often have mixed feelings about political issues. Moreover, political attitudes and values often have multiple dimensions. Consequently, it takes a multidimensional typological analysis to knit a useful conceptual net for capturing attitudinal patterns. In this study, it is critical to realise that there are three representative attitudes towards the CPC. Equally important, it requires sophisticated data analysis to identify hypothesised patterns. K-means clustering is a useful tool for identifying patterns of attitudes. However, algorithms cannot replace thinking and analysis. Domain knowledge is needed to develop a theoretically valid and computationally optimal classification. [End Page 69]
To conclude, this study has some policy implications. It suggests that the CPC can maintain popular support by performing well in developing the economy, maintaining good governance and reining in corruption. However, it also cautions that the Party’s rank and file may be demoralised by the lack of intra-Party democracy. The study cautions that the US-led alliance of democracy may have underestimated popular support for the authoritarian regime.
Li Lianjiang (lianli@ln.edu.hk) is a Professor in the Department of Government and International Affairs at Lingnan University of Hong Kong. He received his PhD in Political Science from the Ohio State University. His research interests include comparative politics, Chinese politics, quantitative methods of political analysis and social movements.
ACKNOWLEDGEMENTS
For helpful comments on earlier drafts, the author would like to thank Lance L.P. Gore, Frank Pieke, Gunter Schubert and Vivienne Shue. He also thanks Kai Yang for rendering research assistance. The research was supported by the Research Grants Council of Hong Kong (project no. CUHK14601820).
APPENDIX. Description of Variables
Min | Max | Mean | SD | |
---|---|---|---|---|
How much do you trust the Communist Party? (1=strong distrust; 4=strong trust) | 1 | 4 | 3.36 | 0.85 |
The ideal form of government is to let one political party represent the interests of all people (1=strongly disagree; 4=strongly agree) | 1 | 4 | 2.94 | 0.78 |
What is your opinion of our country’s overall economic situation? (1=very poor; 5=very good) | 1 | 5 | 3.90 | 0.84 |
Please compare our country’s current economic condition with that five years ago (1=much worse; 5=much better) | 1 | 5 | 4.31 | 0.79 |
What do you think our country’s economic condition will be in five years? (1=much worse; 5=much better) | 1 | 5 | 4.28 | 0.70 |
The government treats everybody fairly (1=strongly disagree; 4=strongly agree) | 1 | 4 | 2.78 | 0.63 |
The government treats the rich and the poor equally (1=strongly disagree; 4=strongly agree) | 1 | 4 | 2.67 | 0.70 |
Everybody has enough to eat, to wear, has a shelter and other basic life necessities (1=strongly disagree; 4=strongly agree) | 1 | 4 | 2.95 | 0.53 |
How widespread is corruption among central government officials? (1=almost none; 4=almost all) | 1 | 4 | 2.07 | 0.94 |
How widespread is corruption among local government officials? (1=almost none; 4=almost all) | 1 | 4 | 2.64 | 0.74 |
The government is like our parents and should make decisions for us (1=strongly disagree; 4=strongly agree) | 1 | 4 | 2.47 | 0.87 |
How often do you listen, watch or read political news? (1=never; 6=several times a day) | 1 | 6 | 2.95 | 1.54 |
How often do you use the internet? (1=never; 6=almost every day) | 1 | 6 | 2.34 | 2.02 |
Gender (0=female; 1=male) | 0 | 1 | 0.53 | 0.50 |
Age | 18 | 93 | 45.29 | 15.64 |
Educational level in years | 0 | 25 | 5.80 | 4.81 |
Communist Party member (0=non-member; 1=member) | 0 | 1 | 0.15 | 0.34 |
Notes: N = 3,473; missing responses are multiply imputed.
Row entries are minimum and maximum values, means and standard deviations.
Source: Asian Barometer Survey 2011.
Footnotes
1. See for example, Nectar Gan and Steve George, “China Claims Its Authoritarian One-party System Is A Democracy—and One That Works Better than the US”, CNN, 8 December 2021, at <https://edition.cnn.com/2021/12/08/china/china-us-democracy-summit-mic-intl-hnk/index.html> [22 September 2022]. On the “whole-process democracy”, see Wang Zhongyuan’s article in this special issue: Wang Zhongyuan, “Remaking Bonds: Adaptive Party Linkage Building in Contemporary China”, China: An International Journal 21, no. 2 (May 2023): 114–35.
2. The Central Committee of the Communist Party of China, “Resolution of the CPC Central Committee on the Major Achievements and Historical Experience of the Party over the Past Century”, 16 November 2021, at <https://english.www.gov.cn/policies/latestreleases/202111/16/content_WS6193a935c6d0df57f98e50b0.html> [22 September 2022].
3. Ministry of Foreign Affairs of the People’s Republic of China (PRC), “Foreign Ministry Spokesperson Zhao Lijian’s Regular Press Conference on June 15, 2021”, at <https://www.fmprc.gov.cn/mfa_eng/xwfw_665399/s2510_665401/2511_665403/202106/t20210615_9170765.html> [20 September 2022]; see Cary Wu, “Did the Pandemic Shake Chinese Citizens’ Trust in Their Government? We Surveyed Nearly 20,000 People to Find Out”, The Washington Post, 5 May 2021, at <https://www.washingtonpost.com/politics/2021/05/05/did-pandemic-shake-chinese-citizens-trust-their-government/> [22 September 2022].
4. Tang Tsou, “Twentieth-Century Chinese Politics and Western Political Science”, PS: Political Science & Politics 20, no. 2 (1987): 327–33.
5. For example, John J. Kennedy, “Maintaining Popular Support for the Chinese Communist Party: The Influence of Education and the State-controlled Media”, Political Studies 47, no. 3 (2009): 523; Chen Dan, “Local Distrust and Regime Support: Sources and Effects of Political Trust in China”, Political Research Quarterly 70, no. 2 (2017): 321; Jennifer Pan and Xu Yiqing, “China’s Ideological Spectrum”, Journal of Politics 80, no. 1 (2018): 254–73.
6. Guang Lei, Margaret Roberts, Xu Yiqing and Zhao Jiannan, “Pandemic Sees Increase in Chinese Support for Regime, Decrease in Views towards the US”, China Data Lab, 21st Century China Center at the UC San Diego School of Global Policy and Strategy, at <http://chinadatalab.ucsd.edu/viz-blog/pandemic-sees-increase-in-chinese-support-for-regime-decrease-in-views-towards-us/> [22 September 2022].
7. See for example, Maurice Duverger, Political Parties: Their Organization and Activity in the Modern State (London: Methuen, 1954); Philip E. Converse, “The Nature of Belief Systems in Mass Publics”, in Ideology and Discontent, ed. David Apter (New York: Free Press, 1964), pp. 206–61; “A ‘Necessary Evil’ Even During Hard Times? Public Support for Political Parties in Portugal Before and After the Bailout (2008 and 2012)”, Party Politics 22, no. 6 (2016): 720; Korinna Lindemann and Lukas F. Stoetzer, “The Effect of Televised Candidate Debates on the Support for Political Parties”, Electoral Studies 69 (2021): 1022–43.
8. Angus Campbell, Philip E. Converse, Warren E. Miller and Donald E. Stokes, The American Voter (Chicago, IL: University of Chicago Press, 1980); Michael S. Lewis-Beck, Helmut Norpoth, William G. Jacoby and Herbert F. Weisberg, The American Voter Revisited (Ann Arbor, MI: University of Michigan Press, 2008); John Richard Petrocik, “Measuring Party Support: Leaners Are Not Independents”, Electoral Studies 28 (2009): 562–72.
9. Richard Gunther and Larry Diamond, “Species of Political Parties: A New Typology”, Party Politics 9, no. 2 (2003): 180.
10. For example, Zheng Yongnian, “The Party Domination of the State”, in The Chinese Communist Party as Organizational Emperor: Culture, Reproduction, and Transformation, ed. Zheng Yongnian (London: Routledge, 2010); Guo Sujian, “The Party–State Relationship in Post-Mao China”, China Report 37, no. 3 (2001): 301–15; Gunter Schubert, “One-Party Rule and the Question of Legitimacy in Contemporary China: Preliminary Thoughts on Setting Up A New Research Agenda”, Journal of Contemporary China 17, no. 54 (2008): 191–204.
11. For example, Joseph Fewsmith, “Balances, Norms and Institutions: Why Elite Politics in the CCP Have Not Institutionalized”, The China Quarterly 248, SI (2021): 265–82.
12. He Junzhi, “Independent Candidates in China’s Local People’s Congresses: A Typology”, Journal of Contemporary China 19, no. 64 (2010): 311–33; Yuan Zaijun, “Independent Candidates in China’s Local People’s Congress Elections”, Journal of Chinese Political Science 16, no. 4 (2011): 389–405; Sun Ying, “Independent Candidates in Mainland China: Origin, Development, and Implications for China’s Democratization”, Asian Survey 53, no. 2 (2013): 245–68.
13. Pierre F. Landry, Deborah Davis and Wang Shiru, “Elections in Rural China: Competition without Parties”, Comparative Political Studies 43, no. 3 (2010): 763–90.
14. For example, Kennedy, “Maintaining Popular Support for the Chinese Communist Party”, pp. 522–3; Lü Xiaobo, “Social Policy and Regime Legitimacy: The Effects of Education Reform in China”, American Political Science Review 108, no. 2 (2014): 431; Tang Wenfang, Populist Authoritarianism: Chinese Political Culture and Regime Sustainability (New York: Oxford University Press, 2016), p. 69.
15. Jennifer Pan and Xu Yiqing, “China’s Ideological Spectrum”, The Journal of Politics 80, no. 1 (2018): 254–73; Chen, “Local Distrust and Regime Support”, p. 321.
16. For example, Wang Zhengxu, “Before the Emergence of Critical Citizens: Economic Development and Political Trust in China”, International Review of Sociology 15, no. 1 (2005): 155–71.
17. Chu Yun-han, “Sources of Regime Legitimacy and the Debate over the Chinese Model”, The China Review 13, no. 1 (2013): 1–42; Meir Alkon and Erik H. Wang, “Pollution Lowers Support for China’s Regime: Quasi-Experimental Evidence from Beijing”, The Journal of Politics 80, no. 1 (2018): 328; Xiang Jun and Jay D. Hmielowski, “Alternative Views and Eroding Support: The Conditional Indirect Effects of Foreign Media and Internet Use on Regime Support in China”, International Journal of Public Opinion Research 29, no. 3 (2017): 416.
18. Alexander Wuttke, Christian Schimpf and Harald Schoen, “When the Whole is Greater than the Sum of Its Parts: On the Conceptualization and Measurement of Populist Attitudes and Other Multidimensional Constructs”, American Political Science Review 114, no. 2 (2020): 356–74.
19. See for example, Wang, “Before the Emergence of Critical Citizens”.
20. Wang Zhengxu and You Yu, “The Arrival of Dissentient Citizens: Decline of Political Trust and Shifting Public Priorities in China”, International Review of Sociology 26, no. 1 (2016): 114–5.
21. Li Lianjiang, “Decoding Political Trust in China: A Machine Learning Analysis”, The China Quarterly 249 (2022): 1–20.
22. Chen, “Local Distrust and Regime Support”, p. 321.
23. Vivienne Shue, “Legitimacy Crisis in China?”, in State and Society in 21st-century China, ed. Peter Hays Gries and Stanley Rosen (New York: Routledge, 2004), p. 43.
24. David Easton, A Systems Analysis of Political Life (New York: Wiley, 1965); David Easton, “A Reassessment of the Concept of Political Support”, British Journal of Political Science 5, no. 4 (1975): 435–57.
25. Chen Jie, Popular Political Support in Urban China (Stanford, CA: Stanford University Press, 2004), pp. 23–6; Lu Jie and Bruce Dickson, “Revisiting the Eastonian Framework on Political Support: Assessing Different Measures of Regime Support in Mainland China”, Comparative Politics 52, no. 4 (2020): 671–701.
26. The survey employs a six-point scale, but only a very small minority of respondents express strong, very strong and extremely strong distrust. For the sake of simplicity, the study merges the three expressions of strong distrust into one category.
27. For example, Robert Harmel and Alexander C. Tan, “One-Party Rule or Multiparty Competition? Chinese Attitudes to Party System Alternatives”, Party Politics 18, no. 3 (2012): 341.
28. John V. Guttag, Introduction to Computation and Programming Using Python: With Application to Understanding Data, 3rd ed. (Cambridge, MA: MIT Press, 2021).
29. Douglas Steinley, “K‐means Clustering: A Half‐century Synthesis”, British Journal of Mathematical and Statistical Psychology 59, no. 1 (2006): 1–34.
30. Robert L. Thorndike, “Who Belongs in the Family?”, Psychometrika 18, no. 4 (1953): 267–76.
31. Peter J. Rousseeuw, “Silhouettes: A Graphical Aid to the Interpretation and Validation of Cluster Analysis”, Journal of Computational and Applied Mathematics 20 (1987): 53–65.
32. Robert Tibshirani, Guenther Walther and Trevor Hastie, “Estimating the Number of Clusters in a Data Set via the Gap Statistic”, Journal of the Royal Statistical Society: Series B (Statistical Methodology) 63, no. 2 (2001): 411–23.
33. The centroid of a cluster is the conceptual centre of mass but is not always a member of the cluster. See Guttag, Introduction to Computation and Programming Using Python.
34. For example, Kiri Wagstaff, Claire Cardie, Seth Rogers and Stefan Schrödl, “Constrained K-means Clustering with Background Knowledge”, International Conference on Machine Learning 1 (2001): 577– 84; Mikhail Bilenko, Sugato Basu and Raymond J. Mooney, “Integrating Constraints and Metric Learning in Semi-supervised Clustering”, Proceedings of the Twenty-first International Conference on Machine Learning (Banff, Canada, 2004): 81–8; John S. Ahlquist and Christian Breunig, “Model-based Clustering and Typologies in the Social Sciences”, Political Analysis 20, no. 1 (2012): 92–112; Guang, Roberts, Xu and Zhao, “Pandemic Sees Increase in Chinese Support for Regime, Decrease in Views towards the US”. Yue Guan, “Deconstructing Multidimensional Public Opinion Constructs: A Concept-driven Clustering Approach”, paper prepared for the 11th Annual Conference of the European Political Science Association, 24–5 June 2021.
35. Xie Jieri et al., “Social Consensus through the Influence of Committed Minorities”, Physical Review E 84, no. 1 (2011): 011130.
36. See for example, Philip E. Converse, “The Nature of Belief Systems in Mass Publics”, in Ideology and Discontent, ed. David E. Apter (New York: Free Press, 1964), pp. 206–61; Odelia Oshri, Omer Yair and Leonie Huddy, “The Importance of Attachment to an Ideological Group in Multiparty Systems: Evidence from Israel”, Party Politics 28, no. 6 (2021): 1164–75.
37. See for example, Ronald Rogowski, Rational Legitimacy (Princeton, NJ: Princeton University Press, 1974); S. Erdem Aytaç and Ali Çarkoğlu, “Terror Attacks, Issue Salience, and Party Competence: Diagnosing Shifting Vote Preferences in a Panel Study”, Party Politics 27, no. 4 (2019): 755–66.
38. For example, Joseph La Palombara and Myron Weiner, Political Parties and Political Development (Princeton, NJ: Princeton University Press, 1966); Lewis-Beck, Norpoth, Jacoby and Weisberg, The American Voter Revisited.
39. Kerry Ratigan and Leah Rabin, “Re-evaluating Political Trust: The Impact of Survey Nonresponse in Rural China”, The China Quarterly 243 (2020): 833.
40. Cui Ernan, Tao Ran, Travis J. Warner and Dali L. Yang, “How Do Land Takings Affect Political Trust in Rural China?”, Political Studies 63, no. 1, SI (2015): 102.
41. William McGuire, “Personality and Susceptibility to Social Influence”, in Handbook of Personality Theory and Research, ed. Edgar F. Borgatta and William W. Lambert (Chicago, IL: Rand McNally, 1968), pp. 1130–87; Barbara Geddes and John Zaller, “Sources of Popular Support for Authoritarian Regimes”, American Journal of Political Science 33, no. 2 (1989): 319–47.
42. Kennedy, “Maintaining Popular Support for the Chinese Communist Party”.
43. For example, Huang Haifeng and Yeh Yao-Yuan, “Information from Abroad: Foreign Media, Selective Exposure and Political Support in China”, British Journal of Political Science 49, no. 2 (2019): 611–36; Lei Ya-Wen, “The Political Consequences of the Rise of the Internet: Political Beliefs and Practices of Chinese Netizens”, Political Communication 28, no. 3 (2011): 291–322; Tang Min and Narisong Huhe, “Alternative Framing: The Effect of the Internet on Political Support in Authoritarian China”, International Political Science Review 35, no. 5 (2014): 559–76.
44. See for example, Wang, “Before the Emergence of Critical Citizens”; Yang Qing and Tang Wenfang, “Exploring the Sources of Institutional Trust in China: Culture, Mobilization, or Performance?”, Asian Politics & Policy 2, no. 3 (2010): 415–36; Bruce J. Dickson, Pierre F. Landry, Shen Mingming and Yan Jie, “Public Goods and Regime Support in Urban China”, The China Quarterly 228 (2016): 859–80.
45. For example, Lü, “Social Policy and Regime Legitimacy”; Cai Vera Zuo, Wang Zhongyuan and Zeng Qingjie, “From Poverty to Trust: Political Implications of the Anti-Poverty Campaign in China”, International Political Science Review 44, no. 2 (2021).
46. Jane Duckett and Neil Munro, “Authoritarian Regime Legitimacy and Health Care Provision: Survey Evidence from Contemporary China”, Journal of Health Politics, Policy and Law 47, no. 3 (2022): 375–409.
47. Zhu Jiangnan, Lu Jie and Shi Tianjian, “When Grapevine News Meets Mass Media: Different Information Sources and Popular Perceptions of Government Corruption in Mainland China”, Comparative Political Studies 46, no. 8 (2013): 920–46; Kang Siqin and Zhu Jiangnan, “Do People Trust the Government More? Unpacking the Distinct Impacts of Anti-corruption Policies on Political Trust”, Political Research Quarterly 74, no. 2 (2021): 434–49.
48. Alkon and Wang, “Pollution Lowers Support for China’s Regime”.
49. Shi Tianjian, “Cultural Values and Political Trust: A Comparison of the People’s Republic of China and Taiwan”, Comparative Politics 33, no. 4 (2001): 401–19; Zhai Yida, “Traditional Values and Political Trust in China”, Journal of Asian and African Studies 53, no. 3 (2018): 350–65.
50. Gregor Paul, “From Marx and Engels to Sino–Marxism Focusing on Communist and Confucian (rujia) Notions of Loyalty and Self-criticism”, Asian Studies 9, no. 1 (2021): 263–80.
51. J. Scott Long and Jeremy Freese, Regression Models for Categorical Dependent Variables in Stata, 3rd ed. (College Station, TX: Stata Press, 2014).
52. Bruce J. Dickson, “Who Wants to Be A Communist? Career Incentives and Mobilized Loyalty in China”, The China Quarterly 217 (March 2014): 42–68; Frank N. Pieke, “Party Spirit: Producing a Communist Civil Religion in Contemporary China”, Journal of the Royal Anthropological Institute 24, no. 4 (2018): 709–29.
53. See Matthew Rhodes-Purdy, “Procedures Matter: Strong Voice, Evaluations of Policy Performance, and Regime Support”, Political Studies 69, no. 2 (2021): 412–33.
54. Barry Naughton, “Inside and Outside: The Modernized Hierarchy that Runs China”, Journal of Comparative Economics 44, no. 2 (2016): 411.
55. See He Baogang and Stig Thøgersen, “Giving the People A Voice? Experiments with Consultative Authoritarian Institutions in China”, Journal of Contemporary China 19, no. 66 (2010): 690.
56. Ji Chengyuan and Jiang Junyan, “Enlightened One-Party Rule? Ideological Differences between Chinese Communist Party Members and the Mass Public”, Political Research Quarterly 73, no. 3 (2020): 651.
57. For example, Wang, “Before the Emergence of Critical Citizens”.
58. Lance L.P. Gore, “The Communist Party-dominated Governance Model of China: Legitimacy, Accountability, and Meritocracy”, Polity 51, no. 1 (2019): 161–94.
59. Yang and Tang, “Exploring the Sources of Institutional Trust”.
60. For example, Kang and Zhu, “Do People Trust the Government More?”.
61. Chu, “Sources of Regime Legitimacy and the Debate over the Chinese Model”.
62. Shi, “Cultural Values and Political Trust”; Zhai, “Traditional Values and Political Trust in China”.
63. Kennedy, “Maintaining Popular Support for the Chinese Communist Party”, p. 531.
64. For example, Huang and Yeh, “Information from Abroad”; Lei, “The Political Consequences of the Rise of the Internet”; Tang and Huhe, “Alternative Framing”; Xiang and Hmielowski, “Alternative Views and Eroding Support”.
65. Lewis-Beck, Norpoth, Jacoby and Weisberg, The American Voter Revisited.
66. Kennedy, “Maintaining Popular Support for the Chinese Communist Party”.
67. For example, Wang Qinghua, “Strengthening and Professionalizing Political Education in China’s Higher Education”, Journal of Contemporary China 22, no. 80 (2013): 332–50; Ye Wangbei, “The Return of ‘One Guideline, One Textbook’ Policy: Moral Education Textbook and Teacher Interaction in China”, Comparative Education Review 66, no. 1 (2022): 60–79.
68. Cheng Tiejun and Mark Selden, “The Origins and Social Consequences of China’s Hukou System”, The China Quarterly 139 (1993): 644–68.