Is the poor quality of Chinese civic awareness preventing democracy in China? A case study of Zeguo township.
(1) Common Chinese people receive low education and have little cultural knowledge, hence are unable to grasp the essence of political affairs and are also unable to manage public affairs. (2) Common Chinese people lack the spirit of toleration and compromise, and cannot objectively process different opinions. Divergent opinions on a same issue often escalate into disputes, failing to form a public choice of rationality. (Lai 2001)
These conditions are said to prevent China from achieving democratization without descending into chaos (Cai 2005).
In recent years, this outlook has undergone subtle adjustments, switching from "population quality discourse" to "Chinese democratization model discourse." The contention now is that democracy lacks a universal standard and instead is determined by the conditions of the region. By this notion, China cannot implement "Westernized" democracy (Fang 2009). Other democratization theorists express similar concerns. As one study has put it, "Some contend that ordinary citizens cannot deal with complex policy issues, others that their deliberations will be distorted by gender or class inequalities, and yet others that they will be ineluctably polarizing" (Fishkin et al. 2010, 435).
Can ordinary Chinese citizens engage in democratic activities? Are they ready to put the public interest first in dealing with public affairs? On the grassroots level, democratic activities in China have already been in motion. For instance, mass elections of village leaders are now required under Chinese election law, and in certain districts citizens are given the opportunity to criticize public policy. Many towns in Wenling County of Zhejiang Province practice what is called "democratic heart-to-heart discussion" (kentan), where citizens gather to discuss local public issues. This very special case has already drawn widespread interest and scrutiny.
Within Wenling County, different towns have adopted various strategies of democratic deliberation. A fairly common strategy involves the discussion of the public budget by a sample of citizens. But in Zeguo Township, the incorporation of the deliberative polling method stands apart from the rest. James Fishkin et al. have already drawn important conclusions concerning its process and effectiveness (2010,439-446). In 2005 Fishkin and his team helped the local officials of Zeguo conduct democratic deliberation, which they called "deliberative polling." Through a series of questionnaires, he surveyed local villagers' attitudes on different subjects, assembling a valuable data set. A number of scholars have used this data set in their empirical analyses (Fishkin et al. 2010). As Fishkin and colleagues (2010) concluded, the process of deliberative polling could significantly change the participants' attitudes and result in legitimate consensus. Hence, deliberative democracy is effective.
Regarding the perception that China cannot develop democracy because of "poor citizen quality," many Chinese scholars have refuted this notion in theory. They argue that there is no strong relationship between citizen quality and effective democracy. Nevertheless, few if any studies have applied real-world cases and statistics to the empirical study of this matter. The main reason is the insufficiency of cases or data concerning democratization among Chinese citizens. In our opinion, the data set from Zeguo Township could stand further investigation based on Fishkin's pioneering studies. In particular, his data set recorded the personal information of all participants, which could be adopted as the scale for measuring citizen quality. This information makes it possible to examine citizen quality and its influence on the effectiveness of democratic deliberation. We believe it is a new perspective in the study of deliberative democracy. We also hold that this study provides further empirical evidence to understand the relationship between personal characteristics and effective democracy.
Our study, which builds on Fishkin's questionnaire design and its data set, determines if villagers' individual characteristics such as education, age, career, and gender affect deliberative polling proceedings. This study also considers how villagers form their rational choices during deliberative polling activities. The findings are a step toward explicating the Chinese people's suitability for democratization based on citizen quality.
Overview of Deliberative Polling in Zeguo Township
In Wenling County of Zhejiang Province a deliberative model known as kentan is currently in practice. For our study we have selected the 2005 township public works plan for analysis. (1) At the beginning of 2005 the Zeguo Township government finalized the public works plan for the year to include thirty projects, with a budget of roughly 136 million yuan, including 40 million yuan of available assets. Two questions arose: Which project was most important to the citizens of the township? How will an important project costing 40 million yuan be selected? After many years of previous experience in public deliberation, the township government and Communist Party committee gathered a random sample of residents in the town for a kentan conference.
Four influences determined the township's public funds allocation. First, after hearing the suggestions from representatives--including those from the township's people's congress and the Chinese People's Political Consultative Conference--concerning public works projects for the year 2005, the township government put forth a list of thirty items. Among them were road and bridge maintenance, public cleanliness, and city greening. Second, a panel of twelve experts was invited to determine the feasibility of the projects as well as their approximate budgetary requirements. Next, 275 townspeople were randomly selected as participants for the kentan conference. Then, based on the participants' suggestions, twelve public works projects were presented for review by the township's people's congress. On April 30,2005, during the fifth session of the Fourteenth Congress of Zeguo Township, the representatives voted 82-7 with one abstention to pass the plan--a legally binding resolution. The congress then printed in booklets the records of the kentan process and congressional proceedings and distributed them to civilian participants for feedback.
In this process known as "deliberative polling," (2) selection of participants followed a particular procedure. First, a random sample of 275 participants from the township's total population was chosen to participate in the deliberative process. In order to qualify for participant selection, subjects had to be registered residents of Zeguo Township and over 18 years of age. Stratified sampling was used, dividing the whole sample of qualifying citizens into 97 subsamples. Four representatives were selected for each village with over 1,000 qualified subjects, and two from villages with less than 1,000 people. Each residential committee produced two participants appointed to each village (residential committee), and then simple random sampling was used to select each subsample. Second, on March 30,2005, ten days prior to the kentan conference, participants were given official introductory materials about the thirty projects from experts in consultation with James Fishkin and Baogang He, collaborators on this project. The participants were asked to rate each of the thirty projects on a 10-point scale. (The scale ranged from 0 (extremely unimportant) to 10 (extremely important), with 5 being neither important nor unimportant.)
Next, participant discussion took place. On the morning of April 19, 2005, the participants were randomly divided into sixteen discussion groups, led by trained moderators, to talk about the issues before joining all participants in a larger discussion, repeated in the afternoon. Finally, after the discussions, participants filled out the same questionnaire they were given earlier. The answers to the two sets of questionnaires were collected afterward for data analysis, and each project was ranked according to the participants' importance ratings.
Fishkin and Colleagues' Research and Contributions
Many questions arose from this case study. Do the deliberations affect participant preferences? Does discussion help participants reach consensus? Does it promote public spiritedness? Fishkin et al. employed a series of statistical studies to reach their conclusions (2010, 439-446). First, after the deliberative process, the results for twelve of the thirty projects displayed a statistically significant change at the 0.1 level (two-tailed). Second, participants showed an increase in concern for the common good by choosing projects that were more beneficial for the entire township. Third, the deliberative process showed that participants came away with increased civic knowledge.
Based on the Fishkin study's conclusions, the deliberative polling in Zeguo Township was successful because its goal was to promote "publicly spirited" decision making of the participants' own choosing. We hope to build on this experiment. Our analysis indicates that personal data reflected the population's quality. Furthermore, we examine the effects of individual quality on the success of deliberative polling. Statistical analysis is employed to analyze participant personal data and deliberative polling. (3)
Individual Qualities and the Effectiveness of Deliberative Polling
We used data from the two sets of questionnaires answered by participants to statistically search for the effect of individual qualities on the effectiveness of deliberative polling. We rely on the preference changes of participants to indicate the effectiveness of democratic deliberation. A participant's preference for a project is represented by rating it on a scale from 0 to 10. The higher the rating, the stronger the preference. In addition, we assume this rating holds a ratio level of measurement; in other words, a rating of 0 means a null preference, and the preference differential between any two neighboring ratings is the same. If the "ex-ante versus post-hoc deliberation" changes in preference are consistent among participants with different individual characteristics--or, in other words, if participants with different personal characteristics show similar patterns in preference change--we would infer that the effectiveness of deliberative polling is unrelated to individual participant characteristics, and that any population has the capability of conducting deliberative polling. This is the primary hypothesis to be tested in this study.
Methodology and Data
We collected the original questionnaire sheets from the Zeguo Township government. Although we have all the data sets from three different deliberation cases in 2005, 2006, and 2008, our study is based on the questionnaire from 2005. (4) The data from 2005 were selected because of their analytical clarity, as the two sets of questionnaires provided sufficient comparison, whereas the questionnaire sets from 2008 are not of the same quality. The questionnaire design and kentan organization from 2005 were very methodically assembled, and we believe that the data for 2005, 2006, and 2008 can be regarded as similar enough for the purpose of our hypothesis.
Out of the 275 participants in the 2005 kentan conference, 267 answered the first questionnaire and 242 answered the second questionnaire; 202 were sufficiently completed to qualify for our use. Table 1 presents participants' individual characteristics, showing that males made up a majority of the participants, outnumbering women two to one. The 18- to 35-year-old demographic represented only 16 percent. Most participants were married, and more than 77 percent had less than middle school education. Career distribution indicated that agricultural workers and factory workers represented two-thirds of the participants.
The Econometric Model and Estimation Result
Usually we adopt linear regression Y = [alpha] + [beta]X + [epsilon] to estimate the value of parameter [beta], so as to explore the relationship between X and Y. In this study, Y indicates the preference (rating) change of participants across two questionnaires. The primary variable we examine is the direction of the preference change--that is, how the varying individual characteristics cause the preference to increase, decrease, or remain the same. We do not aim to study the numerical level of influence from individual characteristics to preference changes, so we do not directly adopt the ratings changes (across two questionnaires) as the explanatory variable. Therefore, this study proposes that each participant's rating of the thirty projects may involve one of three possible situations, whose occurring possibilities are used as explanatory variables.
Situation 1 refers to the "no significant change" situation. (5) The subject's utility under this situation is [U.sub.0i], associated with the probability at [P.sub.0i].
Situation 2 refers to the "significant decrease" situation. The subject's utility under this situation is [U.sub.2i], associated with the probability at [P.sub.1i].
Situation 3 refers to the "significant increase" situation. The subject's utility under this situation is [U.sub.2i], associated with the probability at [P.sub.2i].
These utility indicators [U.sub.0i] [U.sub.1i] and [U.sub.2i] are not directly observable and therefore assumed as the sum of a linear function of [X.sub.i] and the error variable [[epsilon].sub.i]. That is [U.sub.ij] = [[mu].sub.ij] + [[epsilon].sub.ij] = [[beta].sub.i] [X.sub.ij] + [[epsilon].sub.ij].
Assuming participants make changes in the preferences (scores) for the sake of utility maximization, how the participants change their preferences depends on the relative size of [U.sub.ij]:
Set [U.sup.*.sub.ij] = [U.sub.ij] - [U.sub.i0] = [v.sub.ij] + [[epsilon].sup.*.sub.ij] = [[beta]'.sub.i] [X.sub.ij] + [[epsilon]'.sub.ij] then:
With situation 1, i.e., [Y.sub.i] = 0, [U.sub.0i] [greater than or equal to] [U.sub.1i], [U.sub.0i] [greater than or equal to] [U.sub.2i]
With situation 2, i.e., [Y.sub.i] = 1, [U.sub.1i] [greater than or equal to] [U.sub.0i] [U.sub.1i] [greater than or equal to] [U.sub.2i]
With situation 3, i.e., [Y.sub.i] = 2, [U.sub.2i] [greater than or equal to] [U.sub.0i], [U.sub.2i] [greater than or equal to] [U.sub.1i]
The probability of [Y.sub.i] = 1 is:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
Assuming the above probabilities follow a logistic distribution:
[P.sub.ji] = exp([B.sub.ji][X.sub.i])/[1 + [j.summation over (j=1)exp]([B.sub.ji][X.sub.i])](j=[sub.1,2]) (1)
[P.sub.0i] = 1/[1 + [j.summation over (j=1)]exp([B.sub.ji][X.sub.i])](j=[sub.1,2]) (2)
We derive the value of [P.sub.ji] to [P.sub.0i]. Impose log transformation on the values:
Ln ([P.sub.ji]/[P.sub.0i]) = [[beta].sub.ji][X.sub.i] + [[epsilon].sub.ji](j=[sub.1,2]) (3)
This study applies a multinomial logit model to test the relationship between individual characteristics (quality) and the effectiveness of democratic deliberation. [X.sub.i] indicates the explanatory variables of preference changes, [[beta].sub.ji] indicates the corresponding coefficients, and [[epsilon].sub.ji] indicates the error term.
We set up five models: preference changes on all projects, preference changes on road and bridge construction projects, preference changes on environment-related projects, preference changes on urban regulation projects, and preference changes on city greening projects.
Explaining the Variables
The variables documented in this article may be divided into four categories. One is the influence of participant demographics on preference outcomes. Another is participant civic consciousness, including understanding (or trust) in government activities, knowledge about civic participation, and feelings about deliberative methods. Here we want to investigate the relationship between participants' preferences and their civic consciousness. A third variable is the effects of external elements, which primarily focuses on observing how subgroup moderators' points of view and guidance techniques affect the participants' preferences regarding the public works projects, as measured in four perspectives. Fourth are variables related to the participants' perceptions of government behavior (or confidence in government). See Table 2 for further explanation of the measurements.
Regarding the values of the dependent variable, we set them at if there is no preference change, 1 if there is a significant decrease, and 2 if there is a significant increase. Detailed descriptions of preference changes are provided in Table 3.
Discussion of Results
Budgetary Estimates for All Projects and Discussions
Table 4 presents the estimates for all projects proposed. In regard to personal attributes, alterations in the negative preference changes of male participants were greater than the rate of no change. Those holding a high school diploma experienced a higher rate of positive change compared to the rate of no change. Teachers, civil servants, entrepreneurs, and merchants had relatively low rates of positive change. Overall, certain personal attributes may have affected the way participant preferences changed; males, entrepreneurs, and merchants were more likely to exhibit change. However, the correlations were not clear and did not display a clear trend. From our estimates, we inferred that educational level, career, and other civic attributes have limited and ambiguous effects on preference changes. Moreover, these attributes convey the false notion that a higher level of education as well as a more successful career signify higher levels of effectiveness in deliberative actions.
From our estimates, we infer that shifts in internal consciousness (opinions on the participation method and on the right of speech of common citizens), perceptions of government behavior, and preferences are strongly linked; changes in participants' knowledge and urge for civic participation were influential. From another perspective, if the deliberative process can increase the civic consciousness in participants, the effectiveness of deliberative polling would also increase. In order to effectively increase participant civic consciousness through the deliberative process, multiple occasions of deliberative polling as well as the close monitoring of results are essential.
We also found that the effects of moderators on the preferences of participants are not significant.
Statistic Results in Project Categories
Tables 5 through 8 represent results from the categories of road and bridge maintenance, public cleanliness, environmental projects, and city greening. The results from each category collectively yielded similar results and were also similar to the results for all the projects combined. In other words, even though certain individual attributes were influential in changing participant preferences, there were no measurable patterns. Furthermore, participants' civic consciousness and participatory inclinations directly correlated with the changes in their preferences. Findings from our previous section indicate deliberative polling's effectiveness in raising civic consciousness and participatory inclinations; that is, in order for deliberative polling to be effective, creating a nuanced and efficient deliberative method is very important.
Deliberative polling in Zeguo Township indicates that the personal attributes of participants are not influential factors in determining the effectiveness of the process, which doesn't mean that any type of person can function equally well in a deliberative process (really a different question). Rather, when a group conducts civic discussion, the personal attributes of participants do not affect the effectiveness of the endeavor.
A look at the structure of Zeguo Township's education levels reveals that 43.56 percent of the population completed primary education or less, 33.66 percent reached middle school, only 10.4 percent reached higher education, and 20.8 percent graduated from high school. Thus, about 80 percent of the participants failed to acquire a high school education; this group can be considered to comprise "low-education citizens." These numbers are consistent with the national average. In 2005 the national higher education ascension rate was 5.56 percent: 12.43 percent for high school and 17.99 percent beyond high school (Zhongguo tongji nianjian 2006). How then are these so-called poor-quality citizens able to conduct deliberative polling and voting? This case study offers several explanations.
The Deliberative Polling Process Facilitates Reaching Consensus
After comparing the results on prioritized issues from the two questionnaires, the influence of deliberative polling on consensus formation could be determined. During the completion of the second questionnaire, if the first few items on the list of projects increase in assignment of importance (that is, the answers concentrate on the first few items on the list in selecting more important projects), there is no doubt that deliberative polling causes participant preferences on the projects to approach consolidation, aiding in consensus formation.
The first half of Table 9 depicts the five most important projects and their levels of importance to the participants. From the first questionnaire to the second questionnaire, wastewater treatment and Wenchang Road construction already held the positions as the two most important projects in the participants' ranked lists. But even within all selected projects, these two saw a dramatic proportional increase in acknowledged importance--an increase of almost 20 percent from 44.52 percent to 64.24 percent. Also, the accumulated weight of the five most important projects rose from 63.01 percent to 77.48 percent; thus, through deliberative polling, participant preferences for the most important projects moved in the direction of more uniformity, with concentration on particular projects.
The presence of subjectivity in the assignment of importance to projects may actually nullify the judgment displayed by participants. To verify this, we eliminated the ordering of most to least important for further analysis shown in the second half of Table 9. The projects to take the first two places remained the water treatment project and construction of Wenchang Road, and their combined weight rose from 49.48 percent to 62.58 percent. These results are basically consistent with the results from the first half of Table 9, revealing the increase in participant understanding and development of a stronger preference after the deliberative process.
Based on these analyses, we determined that deliberative polling could change participant preference and aid in the formation of consensus on certain issues. Additionally, both in the selection of the most important projects and ranking of the most important projects by importance, participants shared the same kind of understanding for public projects. In both questionnaires, water treatment, Wenchang Road construction, and urban regulation were all projects that were considered most important, revealing that citizens care most about the projected development of the area. Table 10 examines the elements of influence in the participants' decisionmaking process. We learned that environmental protection and economic development were the two most prominent categories, as they gained more prominence in the second round of questionnaires (significantly positive at both the 1 percent and 5 percent levels). This clearly indicates that participants are far more concerned for the long term--the publicly beneficial effects of projects on the environment as well as economic and urban development--rather than superficial, unnecessary, or narrowly targeted actions. The participants are logical in their decisions, with complete competence embracing a perspective for the betterment of the community.
The Informative Nature of the Deliberative Process
Table 11 shows that in the questionnaires, participants included "don't know" while evaluating the level of importance of the projects. Through deliberative polling, the "ignorance level" declined with regard to all the projects. Results were significantly positive on both the 1 percent and 5 percent levels. For instance, the ignorance level for all projects fell 8.2 percent, from 37.66 percent to 29.46 percent, and the ignorance level regarding environmental projects fell even more drastically, from 28.55 percent to 14.85 percent, a drop of 13.70 percent. This clearly shows that deliberative polling increases participants' understanding of public projects, since the process is, at its core, an open exchange and communication of information. This proves just how important it is for government policymakers to establish regular open dialogue with citizens.
The Improvement of Civic Awareness and Public Spiritedness
Figure 1 displays participants' opinions concerning the statement, "Most people don't know about public projects beneficial to the community, and it's best to let experts and public servants make decisions." Even though by the time the second questionnaire was answered, over 46 percent of the participants agreed with the statement, with a rise in the "strongly agree" category, the percentage who chose to disagree rose also, from 9.9 percent to 21.3 percent. This represents a relatively significant rise as the number of "no opinion" answers fell. Even though participants still retained doubt in their own ability to contribute toward public policy decisions, deliberative polling caused a change in their views. We believe that support for the aforementioned statement stems from the lack of understanding of public affairs. If the government makes relevant information accessible to the public, it would strengthen the civic knowledge level in citizens, raise the drive for civic participation, and at the same time lessen reliance on experts for decision making.
Figure 2 depicts participants' opinions on the statement, "Ordinary citizens like me have no say in the policymaking process of the government." Even though no change occurred in opinions by the time the second questionnaire was answered, "no opinion" dropped almost 10 percent, while "disagree" and "strongly disagree" rose by 12.8 percent. Deliberative polling clearly contributed to citizen formation of civic consciousness.
We also discovered that deliberative polling increases citizens' understanding of the government, which in turn leads to development of additional trust.
Table 12 displays changes in the level of understanding of the government and subsequent changes from the first questionnaire to the second. In other words, it measures changes in trust in the government. We found that participants held an overall trusting outlook toward the government. They believed that the government would regard suggestions from the questionnaire with sincerity, as well as execute decisions made during the deliberative polling process. The level of trust significantly increased after the deliberative polling, and Table 12 shows that both categories experienced positive change significant at the 5 percent level. These findings fully support the notion that citizen participation in public policy making is directly beneficial for increasing trust in the government and garners public support for it.
Additionally, even the participants with low educational levels saw growth in their participatory awareness and abilities. In this activity, eighty-eight participants had elementary or lower educational levels. With the help of the moderator, they were able to fill in the questionnaire. During the deliberative process, with moderator encouragement, they made their points known. As recorded by He Baogang and Youxing Lou (2008) during four kentan conferences in 2006 in Bianyu village of Zeguo, several illiterate women participated eagerly, demonstrating that those with a low educational level are still capable of participating in public affairs; moreover, recurring political participation can continuously improve illiterate participants' ability to be a part of the deliberative process.
There is a firmly held belief, prevalent at all levels of Chinese society--from the common man to high-level government officials--that poor population quality prevents democratization. This view is neither new nor exclusively held by Chinese people. As early as 1915 Frank Johnson Goodnow proclaimed, "The intelligence of the great mass of its people is not high, owing to the lack of school. The Chinese have never been accorded much participation in the work of government. The result is that the political capacity of the Chinese people is not large." (6) He argued that a monarchy was best suited to China (Goodnow 1915, 424). Regarding this point of view, the renowned Chinese philosopher Hu Shi retorted that democracy is a learning process; the democratic process in itself is the most effective way of democratizing the masses, in the same way that one must get in the water to learn to swim or own an instrument to learn to play (Hu 1935). If low population quality is used to explain resistance to realizing democratization, the population's political capacity will never be reached.
What is democracy? "Democracy is a form of government in which the people rule" (Sorensen 2010, 442). For Robert Dahl, a democracy is a system of government that meets some conditions, such as "a highly inclusive level of political participation in the selection of leaders and policies" (Dahl 1989). Accordingly, the deliberative polling that occurred in Zeguo Township constitutes a democratic experiment.
In the Zeguo deliberative polling experiment, around 80 percent of the participants had less than a high school educational level and, overall, were not well educated. But our findings show that those with a low educational level are completely capable of taking part in deliberative polling and formulating publicly beneficial decisions. Prior to the deliberative polling, participants were not overly concerned with public matters in Zeguo and had a very limited understanding of the topics discussed. But through discussion, their understanding of these matters improved significantly. It shows that the process itself was a learning process and that participants were able to acquire quickly a basic understanding of civic matters. Moreover, prior to the deliberations, participants were in agreement on the need to choose a publicly beneficial project, showing that they were in fact able to consider the public good. Through discussion and sharing of ideals, the participants formed an even stronger agreement and decided on what was publicly beneficial. Those who insist the Chinese people do not possess the spirit of serving their community and compromising their interests clearly are not supported by actual data. Through 2006, 2008, 2009, and 2010, Zeguo Township continued the deliberative polling activities and reached similarly successful outcomes.
In the 2005 deliberative polling, the projects in question were closely tied to the well-being of the community, and understanding the costs of the projects did not require extensive professional knowledge. If the topics of deliberative polling were less directly tied to the citizens' quality of life, if it were conducted on a much larger geographic scale, or if participation required greater understanding of more complicated elite knowledge, how much influence would deliberative polling have had? How much does the structure of the polling process affect its success? These questions require more extensive case studies and statistical analysis. Our study was able to unseat the belief that poor population quality prevents effective democratic procedures.
In China, a country lacking in democratic tradition, the success of deliberative polling in Zeguo not only explains the effectiveness of deliberative democracy on a provincial level. The village is the lowest level of government in China and governs people who are considered "low education citizens." Zeguo is one of 41,040 villages in China; it does not stand out from other villages, and the educational level of its occupants is comparable to the general educational level of other Chinese citizens. The success of deliberative polling in Zeguo shows that there is no reason that it would not succeed in other villages. This case study proves that even though most Chinese people are not well educated, they are still capable of democratic activities. The key is to begin the process and give the people a chance to participate in civic matters and educate them in this process. So-called citizen quality is not a valid reason for preventing the growth of democracy in China.
It has been over a decade since the city of Wenling began implementing the democratic kentan events. Though this case drew speculation and substantial support from higher levels within the government, it is not a widely practiced model. In China, the crux of the issue is not the civic spirit or quality of Chinese people but rather the state's willingness to push democracy forward of its own accord.
Zhenhua Su is associate professor of political economy in the College of Public Administration, Zhejiang University. His primary research fields include political economy, Chinese politics, and political institutions. His works have appeared in a number of Chinese academic journals, such as the Journal of Public Management, Chinese Journal of Labor Economics, and Comparative Economic and Social Systems. He can be reached at firstname.lastname@example.org. Junjie Le is associate professor of economics in the College of Public Administration, Zhejiang University. His primary research fields include labor economics, econometrics, and development economics. His works have appeared in many Japanese journals, including Asian Studies and Chinese Management Studies, as well as in Chinese journals such as Chinese Rural Economy. He is corresponding author and can be reached at lejunjiezju@ 126.com. Yongjing Zhang is assistant professor of economics at Midwestern State University in Texas. His primary research fields include political economy, public finance, and economic development. His works have appeared in Public Choice, European Journal of Political Economy, and Applied Economics Letters, among other journals. He can be reached at email@example.com. Jun Ma is assistant professor of economics at the University of Alabama, where he has taught since 2007. His recent publications have appeared in the Journal of Economic Dynamics and Control and Studies in Nonlinear Dynamics & Econometrics. He can be reached at firstname.lastname@example.org.
We gratefully acknowledge financial support from the China National Natural Science Fund (No. 70703028). The research was also supported by the Fundamental Research Funds for the Central Universities. We thank Professor Dali Yang for his encouragement and valuable suggestions that have greatly improved this article. We also thank the editor and two anonymous referees for providing numerous valuable suggestions that have substantially improved the article. Part of this work was completed when Zhenhua Su visited the Department of Political Science at the University of Chicago. We thank the department for its hospitality.
(1.) We appreciate the local officials in Zeguo Township for providing us the data set for this study.
(2.) Deliberative polling is not only used in China but in sixteen other countries as well.
(3.) The participants' personal data, such as gender, education, and occupation, were taken from the Deliberative Polling survey.
(4.) Data for 2006 and 2008 were collected by the local government following the basic methodology of the 2005 questionnaire designed by Fishkin et al. However, some important variables were dropped in the new questionnaire, which makes the data of 2006 and 2008 not as complete as the 2005 data.
(5.) Since the scores on the importance of the projects are ordinal (based on the rankings subject to personal judgment) rather than cardinal, two different scores (ex-ante and ex-post) do not mean that there is a change on scoring. Therefore, without proper statistical testing, we cannot conclude that two different scores show a meaningful difference in terms of statistics.
(6.) Frank Johnson Goodnow (1859-1939) was an American political scientist and the founding president of the American Political Science Association. In 1913 Goodnow was invited as a constitutional adviser to China. Meanwhile, the president of the Republic of China, Yuan Shih-kai, was planning to restore the monarchy and enthrone himself, sparking a national debate on whether China was more fitted for monarchy or democracy. Goodnow published an article arguing that a republican system was unfit for China, thereby creating public opinion in favor of the monarchy idea.
Cai, Dingjian. 2005. "In Defense of Democracy: Answers to Anti-Democracy Discourse." Peking University Law Journal, vol. 19, no. 3 (May/June), pp.257-279.
Dahl, R. A. 1989. Democracy and Its Critics. New Haven, CT: Yale University Press.
Fang Ning. 2009. "Chinese Democracy Is Better." Hongqi Wengao, vol. 146, no. 2 (January), pp. 4-9.
Fishkin, James S., Baogang He, Robert C. Lushkin, and Alice Siu. 2010. "Deliberative Democracy in an Unlikely Place: Deliberative Polling in China." British Journal of Political Science, vol. 40, no. 2 (January), pp. 435-148.
Goodnow, Frank J. 1915. Memorandum on Governmental Systems, Submitted by Dr. Frank J. Goodnow, Constitutional Adviser to the Chinese Government, to President Yuan Shih-Kai. Originally printed in the Far Eastern Review (August 1915), pp. 103-105. Reprinted in Charles Gregg, American Imperialism: Viewpoints of United States Foreign Policy, 1898-1941, reprint ed. (Arno Press, 1970), pp. 423-424.
He Baogang and Youxing Lou. 2008. "The Enhancement of Deliberative Democracy in Local China Democratic Deliberation on the 2008 Financial Budgeting in Zeguo Township." Studytimes, March 10, p. D5.
Hu, Shi. 1935. "The Politics from One Party to No Party." Independent Review, vol. 171, no. 40 (October), pp. 10-12.
Lai Hairong. 2001. "Dui Yige Liuxing Shuofa De Fenxi: Lun 'Guomin Suzhi' Yu 'Minzhu'" (Analyzing a Popular Expression: On "People's Quality" and "Democracy"). Dangdai Shijie Yu Shehui Zhuyi (Contemporary World and Socialism), vol. 75, no. 1 (January/February), pp. 34-38.
Sorensen, Georg. 2010. "Democracy and Democratization." In K. T. Leitcht and J. C. Jenkins, eds., Handbook of Politics: State and Society in Global Perspective. London: Springer, pp. 441-158.
Zhongguo tongji nianjian 2006 (China Statistical Yearbook 2006). Beijing: National Bureau of Statistics.
Table 1 Individual Characteristics of Participants Number Percentage Participants 202 100.00 Sex Male 127 62.87 Female 75 37.13 Age 18-35 32 15.84 36-50 108 53.47 50+ 62 30.69 Marital status Married 197 97.52 Single 5 2.48 Education Primary and below 88 43.56 Middle school 68 33.66 High school 21 10.40 Trade school 21 10.40 Missing 4 1.98 Career Agricultural/factory worker 133 65.84 Entrepreneur 27 13.37 Merchant 17 8.42 Educator/civil servant 13 6.44 Other 12 5.93 Source: 2005 Zeguo Township questionnaire, in James S. Fishkin, Baogang He, Robert C. Lushkin, and Alice Siu, "Deliberative Democracy in an Unlikely Place: Deliberative Polling in China." British Journal of Political Science, vol. 40, no. 2 (January 2010), pp. 435-448. Table 2 Definitions of the Explanatory Variable Category Unit Individual Sex 2 characteristic Age Age Education 4 level Career 3 Internal Change in 2 consciousness participatory cognition Opinions on 2 participation style Exogenous Moderator 2 influences influences 1 Moderator 2 influences 2 Moderator 2 influences 3 Moderator 2 influences 4 Other Perception 2 change on government behavior Category Definitions Individual Dummy variable: male=1 female=0 characteristic Years of age Dummy variable D1: elementary school and below=1, other=0 Dummy variable D2: middle school=1, other=0 Dummy variable D3: high school=1, other=0 Dummy variable D4: junior college and special secondary school=1, other=0 Reference variable is D1 Dummy variable D1: agricultural/factory worker=1, other=0 Dummy variable D2: entrepreneur/ merchant=1, other=0 Dummy variable D3: teacher/civil servant=1, other=0 Reference variable is D3 Internal Dummy variable: Opinions on the statement consciousness that citizens have no say in policymaking, if answers post-polling changed from agree/no opinion to disagree=1, other=0 Dummy variable: Believes small group discussion participation is extremely important=1, other=0 Exogenous Dummy variable: Believes the moderator influences provides equal opportunity to participate=1, other=0 Dummy variable: Believes moderator's opinion influenced that of the group=1, other=0 Dummy variable: Believes moderator attempted to guide the opinions of the group members=1, other=0 Dummy variable: Believes few participants dominate the panel discussion=1, other=0 Other Dummy variable: the perception score after the deliberation conference is larger than the score a priori=1, other=0 Table 3 Description of Preference Change Preference Change (%) Negatively Project Insignificant Significant Average of road and other construction projects 49.74 28.57 Average of community and other city planning projects 26.98 37.04 Average of park and other landscaping projects 39.15 31.22 Average of sanitation projects 30.16 16.40 Average of all projects 56.08 20.11 Preference Change (%) Positively Project Significant Average of road and other construction projects 21.69 Average of community and other city planning projects 35.98 Average of park and other landscaping projects 29.63 Average of sanitation projects 53.44 Average of all projects 23.81 Source: 2005 Zeguo Township questionnaire, in James S. Fishkin, Baogang He, Robert C. Lushkin, and Alice Siu, "Deliberative Democracy in an Unlikely Place: Deliberative Polling in China." British Journal of Political Science, vol. 40, no. 2 (January 2010), pp. 435-448. Table 4 Determinants for Preference Changes: Average of All Projects Significant Decrease in Preference Marginal Variables Effect t-stat Constant -0.167 -0.695 Individual Gender dummy 0.129 1.877 ** characteristic Age -0.003 -0.922 Education dummy 2 0.028 0.462 Education dummy 3 0.100 1.508 Education dummy 4 0.086 0.588 Career dummy 1 -0.009 -0.336 Career dummy 2 0.034 -0.400 Internal Change in consciousness participatory -0.071 -0.618 cognition Opinions on participation 0.127 2.146 ** influences style Exogenous Moderator -0.036 -1.187 influences 1 Moderator -0.020 -0.563 influences 2 Moderator 0.111 1.249 influences 3 Moderator -0.052 -0.533 influences 4 Other Change in perception of 0.125 2.008 ** government behavior R-squared 0.192 Log likelihood -167.703 Sample size 189 Significant Increase in Preference Marginal Variables Effect t-stat Constant 0.087 0.167 Individual -0.018 0.317 characteristic 0.000 -0.344 0.015 0.313 0.180 1.908 * -0.121 -0.719 -0.125 -1.058 -0.315 -2.136 ** Internal consciousness 0.115 1.711 * 0.025 0.969 influences Exogenous -0.130 -2.168 ** -0.155 -1.148 0.021 0.531 0.013 -0.022 Other 0.009 0.610 R-squared Log likelihood Sample size Notes: Benchmark: no preference changes are marked if the changes are insignificant at the 10% significant level. *, **, and ***,respectively, represent statistical significance levels of 10%, 5%, and 1%. Table 5 Determinants for Preference Changes: Road and Bridge Maintenance Significant Decrease in Utility Marginal Effect Variables (dp/dx) t-stat Constant -0.432 -1.614 Individual Gender dummy 0.065 0.712 characteristic Age 0.003 0.516 Education dummy 2 0.153 1.862 * Education dummy 3 0.206 2.070 ** Education dummy 4 0.133 0.648 Career dummy 1 0.019 0.048 Career dummy 2 0.034 -0.233 Internal Change in participatory consciousness cognition -0.074 -0.383 Opinions on participation style 0.057 0.989 Exogenous Moderator influences 1 0.072 0.119 influences Moderator influences 2 -0.008 0.034 Moderator influences 3 0.136 1.145 Moderator influences 4 -0.008 0.054 Other Change in perception of government behavior 0.172 2.392 R-squared 0.169 Log likelihood -179.224 Sample size 189 Significant Increase in Utility Marginal Effect Variables (dp/dx) t-stat Constant 0.147 0.103 Individual Gender dummy -0.043 -0.399 characteristic Age -0.003 -0.595 Education dummy 2 -0.027 -0.250 Education dummy 3 0.071 1.240 Education dummy 4 -0.173 -1.010 Career dummy 1 -0.036 -0.250 Career dummy 2 -0.187 -1.289 Internal Change in participatory consciousness cognition 0.133 1.703 * Opinions on participation style 0.030 0.754 Exogenous Moderator influences 1 -0.185 -2.589 ** influences Moderator influences 2 0.033 0.287 Moderator influences 3 -0.043 -0.032 Moderator influences 4 0.040 0.428 Other Change in perception of government behavior 0.012 0.851 R-squared Log likelihood Sample size Notes: Benchmark: no preference changes are marked if the changes are insignificant at the 10% significant level. *, **, and ***, respectively, represent statistical significance levels of 10%, 5%, and 1%. Table 6 Determinants for Preference Changes: Public Cleanliness Significant Decrease in Utility Marginal Effect Variables (dp/dx) t-stat Constant -2.537 -0.002 Individual Gender dummy -0.047 -0.792 characteristic Age 0.001 0.405 Education dummy 2 -0.008 -0.308 Education dummy 3 -0.161 -1.338 Education dummy 4 -0.028 -0.260 Career dummy 1 2.495 0.002 Career dummy 2 2.465 0.002 Internal Change in participatory consciousness cognition -0.068 -0.142 Opinions on participation style -0.004 0.331 Exogenous Moderator influences 1 -0.095 -1.875 * influences Moderator influences 2 -0.014 -1.007 Moderator influences 3 -0.028 -0.226 Moderator influences 4 0.046 0.623 Other Change in perception of government behavior -0.051 -0.923 R-squared 0.143 Log likelihood -173.614 Sample size 189 Significant Increase in Utility Marginal Effect Variables (dp/dx) t-stat Constant 1.650 -0.218 Individual 0.015 -0.222 characteristic 0.000 0.187 -0.027 -0.407 0.021 -0.797 0.007 -0.100 -1.502 1.416 -1.470 1.399 Internal consciousness 0.166 1.591 0.060 0.811 Exogenous -0.007 -1.021 influences -0.201 -1.782 0.029 0.088 0.004 0.326 Other -0.010 -0.581 R-squared Log likelihood Sample size Notes: Benchmark: no preference changes are marked if the changes are insignificant at the 10% significant level. *, **, and ***, respectively represent statistical significance levels of 10%, 5%, and 1%. Table 7 Determinants for Preference Changes: Urban Regulations Significant Decrease in Utility Marginal Effect Variables (dp/dx) t-stat Constant -0.106 0.053 Individual Gender dummy 0.149 1.776 * characteristic Age -0.001 -0.446 Education dummy 2 0.062 0.082 Education dummy 3 -0.091 0.399 Education dummy 4 0.212 0.436 Career dummy 1 0.087 0.075 Career dummy 2 0.178 -0.072 Internal Change in participatory consciousness cognition -0.122 -0.623 Opinions on participation style -0.061 -0.019 Exogenous Moderator influences 1 0.048 0.123 influences Moderator influences 2 0.127 1.479 Moderator influences 3 -0.011 -0.739 Moderator influences 4 -0.027 -0.358 Other Change in perception of government behavior 0.073 1.193 R-squared 0.194 Log likelihood -186.443 Sample size 189 Significant Increase in Utility Marginal Effect Variables (dp/dx) t-stat Constant 0.210 0.738 Individual -0.056 0.389 characteristic 0.000 -0.264 -0.099 -0.947 0.252 1.742 * -0.299 -1.474 -0.138 -0.760 -0.335 -1.749 * Internal consciousness 0.141 1.164 0.107 1.195 Exogenous -0.072 -0.753 influences 0.083 1.293 -0.104 -1.171 -0.011 -0.291 Other 0.026 0.870 R-squared Log likelihood Sample size Notes: Benchmark: no preference changes are marked if the changes are insignificant at the 10% significant level. *, **, and ***, respectively represent statistical significance levels of 10%, 5%, and 1%. Table 8 Determinants for Preference Changes: Environmental Projects Significant Decrease in Utility Marginal Variables Effect (dp/dx) t-stat Constant Individual Gender dummy -0.360 -1.005 characteristic Age 0.041 0.302 Education dummy 2 0.001 0.418 Education dummy 3 0.041 0.540 Education dummy 4 0.030 0.118 Career dummy 1 0.308 2.269 ** Career dummy 2 0.038 -0.367 Internal Change in participatory 0.165 -0.011 consciousness cognition Opinions on participation 0.035 0.494 style Exogenous Moderator influences 1 0.127 2.079 ** influences Moderator influences 2 0.061 0.696 Moderator influences 3 -0.039 -1.013 Moderator influences 4 0.012 0.867 Other Change in perception of 0.029 -0.258 government behavior R-squared 0.127 1.757 * Log likelihood 0.194 Sample size -186.783 189 Significant Increase in Utility Marginal Effect Variables (dp/dx) t-stat Constant 0.273 0.518 Individual -0.048 -0.483 characteristic 0.000 0.207 -0.002 0.211 -0.042 -0.262 -0.116 0.107 -0.221 -1.437 -0.394 -2.049 ** Internal consciousness -0.006 0.129 0.022 1.159 Exogenous -0.040 -0.211 influences -0.170 -1.482 0.195 1.856 * -0.129 -1.135 Other -0.002 0.695 R-squared Log likelihood Sample size Notes: Benchmark: no preference changes are marked if the changes are insignificant at the 10% significant level. *, **, and ***, respectively represent statistical Table 9 Project Priority Ranking Most Important Project First Opinion Poll Accounts for Sorting Project Content the Proportion 1 Sewage treatment project 28.08 2 Wenchang trunk road 16.44 3 Urban and rural planning and design 8.90 4 Muyu Hill park 5.48 5 Danya Hill park 4.11 Cumulative weight 63.01 Most Important Project Second Opinion Poll Accounts for Sorting Project Content the Proportion 1 Wenchang trunk road 39.07 2 Sewage treatment project 25.17 3 Urban and rural planning and design 5.96 4 The first construction phase of Dongcheng Road 3.97 5 The second construction phase of Dongcheng Road 3.31 Cumulative weight 77.48 Sum Total of the Most Important Project and the Second Important Project First Opinion Poll Accounts for Sorting Project Content the Proportion 1 Sewage treatment project 17.42 2 Wenchang trunk road 12.89 3 Urban and rural planning and design 8.01 4 Danya stopover station of environmental sanitation 6.97 5 The first construction phase of Muchang trunk road 4.18 Cumulative weight 49.48 Sum Total of the Most Important Project and the Second Important Project Second Opinion Poll Accounts for Sorting Project Content the Proportion 1 Wenchang trunk road 24.50 2 Sewage treatment project 20.20 3 Danya stopover station of environmental 4 sanitation 9.60 Urban and rural planning and design 4.97 5 The first construction phase of Muchang trunk road 3.31 Cumulative weight 62.58 Source: 2005 Zeguo Township questionnaire, in James S. Fishkin, Baogang He, Robert C. Lushkin, and Alice Siu, "Deliberative Democracy in an Unlikely Place: Deliberative Polling in China." British Journal of Political Science, vol. 40, no. 2 (January 2010), pp. 435-448. Table 10 Determinants in the Focus Project Selection First Opinion Poll Average Standard Determinants Score Deviation Economic development 8.71 2.30 Environment protection 9.49 1.65 Convenient life 7.94 2.72 City image 8.55 2.41 Second Opinion Poll Average Standard Determinants Score Deviation Change Economic development 8.99 1.80 0.29 ** Environment protection 9.72 0.89 0.23 *** Convenient life 8.06 2.71 0.13 City image 8.59 2.22 0.04 Notes: ** and ***, respectively, represent statistical significance levels of 5%, 1%. Table 11 Changes in Understanding of the Projects First Opinion Poll Proportion of Project Item Item "Do Not Standard Number Know" Deviation The average of road and bridge 18 39.25 33.87 The average of block planning 3 36.63 39.90 The average of landscaping 6 37.95 38.39 The average of environmental 3 28.55 37.09 construction The average of all projects 30 37.66 32.98 Second Opinion Poll Proportion of "Do Not Standard Percentage Project Item Know" Deviation Change The average of 33.61 32.27 -5.64 ** road and bridge The average of 25.25 36.51 -11.39 *** block planning The average of 26.40 34.17 -11.55 *** landscaping The average of 14.85 29.36 -13.70 *** environmental construction The average of 29.46 29.60 -8.20 *** all projects Notes: ** and ***, respectively, represent statistical significance levels of 5%, 1%. Table 12 Citizens' Understanding of the Government First Opinion Poll Average Standard Score Deviation Do you think the government will seriously consider the suggestions? 7.51 2.93 Do you think the government will implement the result of the consultation? 7.39 2.86 Second Opinion Poll Average Standard Score Deviation Change Do you think the government will seriously consider the suggestions? 7.86 2.42 0.35 ** Do you think the government will implement the result of the consultation? 7.77 2.47 0.38 ** Source: 2005 Zeguo Township questionnaire, in James S. Fishkin, Baogang He, Robert C. Lushkin, and Alice Siu, "Deliberative Democracy in an Unlikely Place: Deliberative Polling in China." British Journal of Political Science, vol. 40, no. 2 (January 2010), pp. 435-448. Note: ** represents statistical significance level of 5% Figure 1 Views on Yielding Public Affairs Decisions to Experts and Public Servants 1 2 3 4 5 6 * 1st poll 16.8 27.7 8.4 9.9 5.4 31.7 ** 2nd poll 21.3 24.8 5.4 21.3 4.0 23.3 Note: 1-Strongly agree; 2-Agree; 3-Neutral; 4-Disagree; 5-Strongly disagree; 6-No opinion. Note: Table made from bar graph. Figure 2 Views on Average Citizen's Lack of Representation in Policymaking 1 2 3 4 5 6 * 1st poll 12.9 20.3 7.4 20.8 5.0 33.7 ** 2nd poll 13.9 18.3 5.4 31.7 6.9 23.8 Note: 1-Strongly agree; 2-Agree; 3-Neutral; 4-Disagree; 5-Strongly disagree; 6-No opinion. Note: Table made from bar graph.
|Printer friendly Cite/link Email Feedback|
|Author:||Su, Zhenhua; Le, Junjie; Zhang, Yongjing; Ma, Jun|
|Date:||Mar 22, 2012|
|Previous Article:||Organizing international security in Northeast Asia: Hegemony, concert of powers, and collective security.|
|Next Article:||Private-sector networks, democracy, and interstate relations: a case study of South Korea and Taiwan.|