Content area
Full Text
Conventional wisdom maintains that Chinese citizens have relatively high levels of trust in the central government but low levels of trust in local government.1 As the story goes, the centre benefits from widespread perceptions that Beijing's policies are well intentioned and that local officials are to blame for improper implementation. Initial analysis of our survey of villagers supports this contention. When nonresponse is taken into consideration, however, we find little difference between trust in central and local officials (see Figure 1).
Figure 1:
Trust in State Leaders with and without Nonresponse
Source:
HEPS survey.
[Figure omitted. See PDF]
When we include nonresponse (i.e. do not know, refuse to answer or missing), the proportion of respondents who stated that they trusted government officials was stable across each level of government: 62 per cent for central leaders; 53 per cent for provincial leaders; 50 per cent for county leaders; and 63 per cent for village leaders. Remarkably, both central and village leaders received higher proportions of respondents expressing trust in them than did provincial or county leaders, complicating our understanding of political trust in China.
To examine item nonresponse in China, this report utilizes the Health, Environmental Protection and Society Survey (HEPS). This original survey of over 800 villagers in rural China was conducted by the authors in 2012. We also conducted a similar analysis with four additional datasets that have been cited as evidence of high political trust in China:2 the 1993 Survey on Social Mobility and Social Change in China (SSMSC); the 2008 China Survey; the 2011 Asian Barometer (Wave 3); and Lianjiang Li's 2014 survey.3
From our HEPS survey, we find that item nonresponse can be significantly reduced through careful questionnaire design and enumerator training. We find low rates of non-substantial response (NSR) on politically sensitive questions in all the data we analysed, which indicates that survey research is useful even in an authoritarian setting. Marginalized individuals are more likely to give a non-substantial response, but the relationship between cosmopolitanism and NSR is unclear. We also find that marginalized individuals have different perceptions of trust compared to their privileged counterparts. Finally, we find that nonresponse may have led scholars to overestimate political trust and exaggerate the gap between trust in central...