Replication dataset for “Good Types in Authoritarian Elections: The Selectoral Connection in Chinese Local Congresses,” Comparative Political Studies, vol. 50, no. 3 (2017): 362–394.
Replication dataset for “Democracy, Community, Trust: The Impact of Chinese Village Elections in Context,” Comparative Political Studies, vol. 39, no. 3 (2006): 301–324.
Replication dataset for “The Electoral Connection in the Chinese Countryside,” American Political Science Review, vol. 90, no. 4 (1996): 736–748.
Appendix on Incomplete Nomination Data for “When Communist Party Candidates Can Lose, Who Wins? Assessing the Role of Local People’s Congresses in the Selection of Leaders in China,” China Quarterly, no. 195 (2008): 607–630.
2007–2009 Anhui, Hunan, Zhejiang Surveys
Data collection was supported by the National Science Foundation.
English-language questionnaire for county congress delegates
Chinese-language questionnaires for township and county and municipal delegates
Data for township, county, and municipal delegates
English-language questionnaire and data for constituents of township delegates
Surveys were conducted with delegates in people’s congresses at the municipal, county, and township levels in Zhejiang, Anhui, and Hunan provinces and with ordinary constituents of a sub-sample of township delegates surveyed. Provinces are a convenience sample, selected for feasibility of implementation. Within provinces, municipalities, counties, and townships are probability samples, selected with probability proportionate to population size. The sample is nested: counties within selected municipalities, townships within selected counties. Specifically, within each province we probabilistically selected two municipalities; these are large municipalities, with (county-level) urban districts within them. Within each of these municipalities, we probabilistically selected one urban district; we also probabilistically selected two counties or county-level cities under the governance of each of the selected municipalities. Within each of the selected counties or county-level cities (but not urban districts), we probabilistically selected two townships or towns. The result is a representative unbiased sample of congresses in the three provinces, comprising 6 municipal congresses, 18 county-level congresses, and 24 township or town congresses. We successfully surveyed 1,232 township congress delegates, 3008 county congress delegates, and 890 municipal congress delegates. Finally, because the project is interested in the relationship between delegates and their constituents, we also surveyed the mass public. For this, we selected two voting districts, with probability proportionate to size, in each selected township or town, resulting in a sample of 983 ordinary citizens in 48 voting districts. With voting district identifiers in survey data from congress representatives, a subset of township and town representatives can be matched to their constituents. For more detail on the survey, see Information for Autocrats: Representation in Chinese Local Congresses, Cambridge Studies in Comparative Politics (Cambridge: Cambridge University Press, 2015), Appendix A.
2005 Anhui Survey
I conducted this exploratory survey with Chinese colleagues at Anhui University, who also supervised fieldwork. I worked with my colleagues to design a survey consisting (almost) entirely of open-ended questions, trained a team of graduate students to survey a small sample of local people’s congress delegates, and worked together with the students to develop codes and to code and input the data. The survey was conducted in December 2005 in four townships and towns, located in a county and district under Hefei municipality in Anhui province. The localities are purposive samples, with different geographic features and in which the survey team had the contacts to ensure cooperation. Respondents were intended to be a census of congress representatives: that is, an effort was made to interview the entire roster of congress delegates in each of the four townships and towns. The decision to undertake a census rather than probabilistic sampling reflected my concern that an adequate number of respondents be interviewed; when the congress is not in session, finding respondents is a non-trivial task, as they are located in voting districts scattered across the townships and towns. Many delegates were not in the townships and towns on the days that interviewing was conducted. A total of 130 delegates were interviewed, which is 60 percent of congress delegates in the localities surveyed.
For my Chinese colleagues and their graduate students, the survey was an opportunity to learn more about survey methodology and statistical analysis. My main purpose was to explore the plausibility of a much larger probability sample survey of congress delegates. In particular, mindful of their official status and (for most) communist party membership, I wondered how frank and comfortable delegates were likely to be as interview subjects talking about relationships of power with the party and government. Partly for this reason, the survey focused especially on the most politically sensitive congress function—election of government leaders, which requires congresses to accept or reject candidates vetted and selected by the communist party committee one level up. This stage of research was also important in discovering appropriate language to design unambiguous, specific items presented in the forced-choice format that facilitates coding in a larger survey. This proof-of-concept exercise was instrumental in designing the instrument for the 2007–2009 largescale survey of local congress delegates and their constituents.
1990 and 1996 Surveys in Anhui, Hebei, Hunan, Tianjin
Data collection was supported by the Henry Luce Foundation, National Science Foundation, University of Michigan, and State Education Commission of the People’s Republic of China. Data and documentation for the 1990 mass public survey have been deposited at the Inter-University Consortium for Political and Social Research, University of Michigan at https://www.icpsr.umich.edu/ ICPSR 6805, Four-County Study of Chinese Local Government and Political Economy, 1990. Data and documentation for the 1990 elite survey and all 1996 data and documentation are available with the permission of mainland Chinese co-investigators. Please contact me by e-mail.
These surveys attempt to gauge the impact of a rapidly changing social, political, and economic environment. They are pairs in a panel study. They also systematically compare two different populations—villagers and local officials at various levels in the countryside. They are “omnibus surveys,” the product of negotiation among ten Chinese and American scholars with a variety of research agendas.
Survey data were collected in 4 counties, 20 townships, and 60 villages as part of a collaborative project begun in 1988 by political scientists at Peking University and the University of Michigan. Located in Hebei, Hunan, Anhui, and Tianjin, the counties are a non-probability sample selected to exhibit some regional variation from a sampling frame of about 20 counties in which, based on the experience of Chinese colleagues, we could reasonably expect agreement from local authorities to facilitate fieldwork. The townships and villages surveyed are probability samples, selected according to a nested sampling design. In each of the 4 counties, we drew a stratified probability sample of 5 townships, selected with probability proportionate to size and stratified by per capita income. In each selected township, we drew a sample of 3 administrative villages selected with probability proportionate to size. The sample of villagers is a systematic random sample using interval sampling of individuals drawn from official household registration lists (aggregated at the village level) after first eliminating from the lists individuals below age 18 and above age 80. In the 1996 survey, we attempted to interview again, if they could be found, the villagers interviewed in 1990, but also added respondents to fill out the sample, with the same sampling method used in 1990. Local officials are purposive samples of village, township, and county leaders. In 1990 we completed interviews with 252 local officials and 1,270 villagers; in 1996 we completed interviews with 242 local officials and 1,414 villagers, of whom 720 are re-interviews.