Distrust in Government Leaders, Demand for Popular ... 2011...Distrust in Government Leaders, Demand...
Transcript of Distrust in Government Leaders, Demand for Popular ... 2011...Distrust in Government Leaders, Demand...
ORI GIN AL PA PER
Distrust in Government Leaders, Demandfor Leadership Change, and Preferencefor Popular Elections in Rural China
Lianjiang Li
Published online: 12 June 2010
� Springer Science+Business Media, LLC 2010
Abstract This paper examines the relationship between distrust in incumbent
government leaders and demand for systemic changes in rural China. It finds that
individuals who distrust government leaders’ commitment to the public interest
have both stronger demand for leadership change and stronger preference for
popular elections. It argues that distrust in government leaders may have enhanced
the demand for leadership change, which in turn may have reinforced the preference
for elections. It further argues that distrust in incumbent leaders has in effect
induced a demand for systemic changes, as introducing popular election of gov-
ernment leaders would require a major constitutional amendment. The paper sug-
gests that two distinctive mechanisms may be at work in determining whether
distrust in current government authorities induces preference for systemic changes.
Whether citizens can engineer leadership change through existing channels influ-
ences the generation of idealistic wishes for a better political system. Perceived
availability of better and viable alternatives affects whether idealistic wishes
become a practical preference.
Keywords Political trust � Distrust in government leaders � Leadership change �Systemic changes
Introduction
Trust in incumbent government authorities enhances system support while distrust
generates demand for leadership and policy changes (Abravanel and Busch 1975,
p. 57; Muller and Jukam 1977; Chanley et al. 2000; Hetherington 2004). Scholars,
L. Li (&)
Department of Government and Public Administration, Chinese University of Hong Kong,
Shatin, N.T., Hong Kong, China
e-mail: [email protected]
123
Polit Behav (2011) 33:291–311
DOI 10.1007/s11109-010-9111-3
however, disagree about whether distrust in incumbent government leaders induces
preference for systemic changes. As regards the United States, one side argues that
distrust in current authorities may ultimately weaken public acceptance of
democratic principles and result in support for undemocratic alternatives (A. Miller
1974a, b; Caddell 1979). The other side argues that distrust in political authorities
does not undermine support for the democratic system (Citrin 1974; Abravanel and
Busch 1975, p. 80; W. Miller 1979). Further empirical studies on the US and other
democracies have not settled this debate. Some researchers observe that distrust in
current administration significantly weakens support for a democratic system
(Miller and Listhaug 1990; Waldron-Moore 1999; Bratton and Mattes 2001;
Seligson 2002a, p. 180; Sarsfield and Echegaray 2006; also see Avery 2006, 2007;
Weitz-Shapiro 2008). Others find that distrust in incumbent government authorities
has little impact on democratic support (Lipset and Schneider 1983; Citrin and
Green 1986; Finkel et al. 1989; Rose and Shin 2001).
These studies concentrate on democracies where free election of government
leaders has been institutionalized. The debate has thus focused on whether distrust
in freely elected government authorities weakens support for democracy. Only
recently have scholars begun to address another dimension of the theoretical puzzle,
which is whether distrust in political authorities who are not freely elected
stimulates demand for systemic changes in authoritarian and transitional societies.
Their findings, though, vary from place to place. On one hand, a study by Rose
(2007) finds that distrust in government leaders in post-soviet Russia, Ukraine, and
Belarus, where elections fall short of being free and fair, does not generate
preference for further democratization. On the other hand, Johnson (2005, pp. 79–
80) observes that distrust in political leaders in the quasi-authoritarian Ukraine
enhances support for democracy, while a more recent study by Jamal (2007) finds
that in the Arab world individuals who have less confidence in authoritarian
government have stronger support for democracy.
This study joins the discussion by examining the relationship between distrust in
current government leaders and demand for systemic changes in China. The case of
China is interesting for two reasons. First, the country has witnessed a growing
number of popular protests in the last two decades (Walker 2006; Cai 2008; O’Brien
2008), which suggests that trust in government authorities may have declined
(cf. Abravanel and Busch 1975; Muller and Jukam 1977). Second, public demand
for systemic changes seems to be on the rise. Disillusioned protest leaders, for
instance, have openly called for ending one-party rule (Li 2004, p. 239; Li and
O’Brien 2008, p. 23). More importantly, a three-wave survey conducted in the
Chinese capital city finds that ‘‘diffuse support for the regime had declined by more
than 6 percent’’ from 1995 to 1999 (Chen 2004, p. 184). Is distrust in government
leaders associated with demand for systemic changes in China? What can we learn
from the China case about the relationship between distrust in government
authorities and preference for systemic changes?
Drawing on survey data collected in the Chinese countryside, this paper
addresses these questions. It first examines how distrust in government leaders’
commitment to the public interest is associated with demand for leadership change
and preference for popularly electing government leaders. It then explores how
292 Polit Behav (2011) 33:291–311
123
distrust in leaders may foster idealistic wishes for a better political system and why
only certain idealistic wishes become practical demands for systemic changes. It
concludes with a discussion about the theoretical and methodological implications
of survey findings from rural China.
Data and Methods
This study relies on a local survey conducted in 2007. The field sites were Dongshan
county and Wuping county in Fujian province, Shangrao county in Jiangxi province,
and Yongjia county in Zhejiang province. The four counties were selected by
convenience. Sampling in each county was conducted in three stages. First, five
townships were selected with probability proportionate to size (PPS). Second, four
villages were selected from each township with PPS. Lastly, within each selected
village around 20 randomly chosen individuals over the age of 18 were interviewed,
regardless of village population size. Interviews were administered by advanced
undergraduate students in the Department of Sociology from a leading university in
South China under the supervision of their professor. Altogether 1,600 farmers were
interviewed. Cases are weighted according to the rural population size of the county.
To adjust for survey design effects, each selected township is treated as a stratum
and each village a cluster.
As with other survey studies (e.g., Zhu 1996; Berinsky 2004), this research faces
the unavoidable problem of missing responses. Altogether 7.7% of 1,600
respondents did not answer one or more of the 45 questions analyzed in this study.
Little’s Missing Completely at Random (MCAR) test (Little 1988) showed that the
data were not MCAR (N = 1,600; v2 = 3,437; df = 3,063; p = .000). Assuming
that the data were missing at random (MAR), missing values were multiply imputed
to increase the efficiency of estimation and to make inferences valid by reflecting
additional variability due to the missing values (Rubin 1987; Schafer and Olsen
1998; King et al. 2001). Five multiply imputed datasets were generated using the
Markov Chain Monte Carlo (MCMC) method (Schafer 1997; SPSS Inc 2008, p. 25).
The imputed datasets were analyzed with Mplus 5.0 (Muthen and Muthen 2006,
p. 315). Estimates and standard errors of direct and indirect effects obtained from
imputed datasets were then combined using the rules given by Rubin (Rubin 1987;
Schafer and Olsen 1998, pp. 556–557). For the purposes of comparison, the analytic
model was also fitted to the original data using listwise deletion and to a singly
imputed dataset generated with the Expectation Maximization (EM) algorithm
(Little and Rubin 1987). Results obtained from the three alternative treatments of
missing values are highly consistent.
The data analysis proceeded in two steps. Confirmatory factor analysis was used
to assure that multiple measures of latent theoretical constructs were reliable. A path
model was then estimated to examine the relationships between the outcome,
explanatory, mediator, control, and background variables. Ordinal and binary
indicators of latent constructs were treated as crude measurements of latent
continuous factors. The WLSMV (weighted least squares mean and variance
adjusted) estimator was used to correct for the bias that might result from having
Polit Behav (2011) 33:291–311 293
123
ordinal endogenous variables in the model (Muthen and Muthen 2006, p. 424).
Since it relies on a local probability sample, the research focuses on exploring
relationships between variables (Manion 1994).
The Outcome Variable: Preference for Elections as an Indicatorof Demand for Systemic Changes
The outcome variable is preference for popularly electing government leaders.
Preference for elections is used as an indicator of demand for systemic changes in
China for two reasons. First, an expressed preference for elections may signify a
certain degree of disapproval of the current political system, under which ordinary
citizens have little say in the selection of government leaders (Burns 1989; Chan
2004). Second, a preference for popular elections may indicate a demand for an
institutional change that will require a major constitutional amendment. The
Constitution of the People’s Republic of China (adopted 1982, last amended 2004)
stipulates that government leaders of all levels are elected by deputies to the
corresponding levels of people’s congress rather than ordinary citizens (Arts. 79.1;
101.1) (see O’Brien 1990, pp. 61–64, 127–137; Manion 2000, 2008).1
Preference for elections was tapped by asking whether government leaders
should be directly elected by the people through one-person one-vote. Three choices
were provided to the respondents: (1) should not; (2) should but conditions are not
ripe; (3) should and elections can be held now. The question was asked,
respectively, about the state chairman (i.e., the Chinese president), provincial
governor, city mayor, county head, and township head (see Appendix for variable
descriptions). Individuals who found elections undesirable were expected to choose
‘‘should not,’’ those who wished to elect government leaders but found it
impractical were expected to choose ‘‘should but conditions are not ripe,’’ and
those who found elections both desirable and feasible were expected to choose
‘‘should and can be held now.’’
The measure was relevant and stimulating. Excepting for a few experimental
elections of township heads in Sichuan province and Guangxi province (Li 2002;
Dong 2006), no popular election of government leaders had been held at any level in
China. Equally important, the measure made good sense to Chinese farmers, many
1 It is worth noting that preference for elections is not equivalent to support for democratization, not to
mention demand for overall regime change. As important as it is, popular election of government leaders
is only one of several major dimensions of a democratic system. Moreover, election without multiparty
competition can hardly be considered a criterion for democracy. Even if government leaders at all levels
are elected with one-person one-vote, China will at best become an ‘‘electoral authoritarian’’ country
(e.g., Diamond 2002; Schedler 2006) rather than a democracy if such elections are limited to and run by a
single ruling party and exclude all potential organized opposition. Recent survey studies of political
support in urban China have used more direct measures to tap popular demand (or the lack of it) for
regime change, e.g., to what extent an individual feels proud to live under the current political system and
feels obliged to support the current political system, if an individual believes that the communist-led
multi-party system should be changed, and whether an individual finds political stability more important
than democratization (Chen 2004, p. 23; Tang 2005, pp. 70–76).
294 Polit Behav (2011) 33:291–311
123
of whom had directly elected villagers’ committees through one-person one-vote for
nearly a decade by the time of the survey (Manion 1996, 2006).2
Preference for elections turned out to be fairly widespread (Table 1). Over two-
thirds of the respondents agreed that leaders of all five levels of government should
be directly elected through one-person one-vote, either immediately or when
conditions are mature. A preference for elections was strongest in regard to
township head, and decreased steadily regarding higher levels of government
leaders. While 86% of the respondents agreed that township heads should in
principle be popularly elected, 70% thought the same about the state chairman.
Similarly, a practical preference for elections was much stronger for lower level
government leaders than for higher levels. Nearly 39% of the respondents thought
that township heads should and could be popularly elected right away, when only
16% thought the same about the state chairman.
The Explanatory Variable: Distrust in Government Leaders’ Commitmentto the Public Interest as an Indicator of Distrust in Government Authorities
The explanatory variable is distrust in incumbent government leaders, which is
multi-dimensional as its objects include leaders’ commitment, competence,
character, equity, honesty, and responsiveness (Abramson 1972, p. 1245;
Weatherford 1984, pp. 188–189; Citrin and Muste 1999). This study focuses on
distrust in government leaders’ commitment to rule in the interests of the governed
(Levi and Stoker 2000, p. 476), because popular elections can improve the incentive
compatibility or ideological congruence between elected elites and their constit-
uents (Achen 1978; Dalton 1985; Huber and Powell 1994; Manion 1996).
Individuals who distrust the current leaders’ commitment to the public interest
are expected to prefer popular elections, because it will enable them to remove
untrustworthy leaders imposed from above and to choose trustworthier ones on their
own.
Distrust in government leaders’ commitment to the public interest was tapped
with three indicators. Respondents were asked whether they believed that
Table 1 Do you think the following government leaders should be directly elected by the people through
one-person one-vote?
Should not Should but
conditions not ripe
Should and can
be held now
Non-response
The state chairman 28.8 53.7 16.4 1.1
Provincial governor 26.1 55.8 16.9 1.2
City mayor 24.2 56.1 18.7 1.1
County head 16.9 52.9 29.0 1.2
Township head 12.1 47.9 38.9 1.1
Note: N = 1,600. Row entries are percentages, which may not add up to 100 due to rounding errors
2 Villagers’ committees are not a level of government but ‘‘mass organizations of self-government.’’
Polit Behav (2011) 33:291–311 295
123
government leaders: (1) put their own interests before those of farmers; (2) do not
care whether farmers will agree when they make policies; and (3) care primarily
about the powerful and rich and neglect the interests of ordinary people. Provided
choices were: (1) fully disbelieve; (2) disbelieve; (3) half believe and half doubt;
(4) believe; and (5) fully believe. The question was asked about all five levels of
government leaders: central, provincial, city, county, and township.
Measured by these three indicators, popular distrust was highest for township
leaders and decreased for higher levels. Over 22% of the respondents had no
confidence in township leaders in all three aspects, while 15% had no trust in central
leaders. Conversely, over 35% of the respondents were confident about central
leaders in all three aspects, while less than 20% had full confidence in township
leaders. The observation corroborates a previous finding that the Chinese people
have less trust in lower levels of government than in higher levels (Shi 2001; Li
2004).
Confirmatory factor analysis showed that for each level of government leaders
the responses to the three questions constituted a single latent construct (standard-
ized factor loadings ranged from .712 to .898). Five latent continuous variables were
then constructed to measure distrust in central, provincial, city, county, and
township leaders (see Appendix).
Mediator Variable: Demand for Leadership Change
Demand for leadership change is assumed to be the mediator through which distrust
in leaders affects preference for elections. First, distrust in the incumbent leaders
does not always induce demand for leadership change. Some Chinese farmers, for
instance, do not want to see corrupt officials replaced on grounds that ‘‘when a full
tiger leaves, he will inevitably be replaced by a hungry wolf’’ (Li and O’Brien 1996,
p. 34). Second, distrust in incumbent leaders may induce demand for leadership
change without generating demand for any systemic changes. An individual who
wants leadership change may remain confident in the existing system of top-down
appointment and may look to higher levels for the desired change. Only those
individuals who both want leadership change and no longer have confidence in top-
down appointment are likely to favor popular election as an alternative means of
choosing government leaders.
To tap their demand for leadership change, respondents were asked whether they
would vote for the incumbent government leader if, hypothetically, a democratic
election was held. Provided choices were: (1) yes; (2) unsure; and (3) no. The
question was repeated about the state chairman, provincial governor, city mayor,
county head, and township head. Reservations about voting for the incumbent were
interpreted as an indication of demand for leadership change.
Demand for leadership change turned out to be remarkably widespread (Table 2).
Over 40% of the respondents showed reservations about voting for the incumbent
state chairman. Demand for leadership change was stronger for provincial governor,
city mayor, county head, and township head, as over 60% of the respondents were
either unsure or determined not to vote for the incumbent.
296 Polit Behav (2011) 33:291–311
123
Three hypotheses were formulated. First, distrust in government leaders should
lead to a stronger demand for leadership change. Second, demand for leadership
change should lead to a stronger preference for elections. Lastly, distrust in leaders
should have a positive indirect effect on preference for elections by inducing a
stronger demand for leadership change.
Control Variables
Since it included a mediator variable, the analysis had two sets of control variables.
Life satisfaction and perception of corruption were controlled to determine if
distrust in current leaders’ commitment to the public interest had an independent
effect on demand for leadership change. Individuals who were satisfied with their
lives might give credit to incumbent government leaders (cf. Catterberg and Moreno
2006, p. 43) and prefer they stay in office, while those who were unsatisfied might
blame current leaders (cf. Kinder 1981) and want leadership change. In the
meantime, individuals who were more acutely aware of political corruption were
expected to have a stronger demand for leadership change. Life satisfaction was
tapped with three measures: (1) self-assessment of change in family income in the
last 2 years; (2) expectation of change in family income in the coming 2 years; and
(3) satisfaction with one’s life in general (see Appendix). The three indicators
constituted a single latent construct (standardized factor loadings ranged from .650
to .796). Perception of corruption was also tapped with three indicators.
Respondents were asked if township officials, county officials, and provincial and
city government officials had in the last decade engaged in corruption. Provided
choices were: (1) no; (2) do not know; (3) it is said so; and (4) yes. The three
indicators constituted a single latent construct (standardized factor loadings ranged
from .728 to .970).3 Both life satisfaction and perception of corruption were
expected to have a direct impact on demand for leadership change. They were also
expected to indirectly affect demand for leadership change by influencing distrust in
government leaders.
Table 2 Would you vote for the incumbent leader if a democratic election was held?
Yes Unsure No Non-response
The state chairman 58.1 37.7 3.8 .4
Provincial governor 37.8 57.0 4.8 .5
City mayor 32.9 61.0 5.7 .4
County head 30.8 59.0 10.0 .2
Township head 29.1 57.7 13.0 .2
Note: N = 1,600. Row entries are percentages, which may not add up to 100 due to rounding errors
3 The answer ‘‘do not know’’ was treated as a valid response rather than a missing value because it was
read out as an alternative answer during the interview and it indicates a level of perception of corruption
which is stronger than answering ‘‘no’’ but weaker than answering ‘‘it is said so.’’
Polit Behav (2011) 33:291–311 297
123
Five variables were controlled to determine if demand for leadership change had
an independent effect on one’s preference for elections. First, life satisfaction was
controlled, as it has been observed to affect diffuse support for the political system
in China (Chen 2004, p. 111). Individuals who were unsatisfied with their lives were
expected to have a stronger preference for elections. Second, perception of
corruption was controlled, as individuals who were more acutely aware of
corruption were expected to have a stronger preference for popularly electing
government leaders. The third control variable was internal political efficacy, i.e.,
the sense of competence in understanding public affairs and participating in politics
(Abramson 1972; Balch 1974; Craig and Maggiotto 1982; Craig et al. 1990). To tap
their sense of efficaciousness, respondents were asked, hypothetically, if they would
be able to: (1) judge whether a county head has done a good job after listening to
his/her work report; (2) judge which one is better if two candidates for county head
debated each other; and (3) judge whether decisions by a county head benefit or
harm their interests (see Appendix). The three indicators constituted a single latent
construct (standardized factor loadings ranged from .855 to .960). More efficacious
individuals were expected to have a stronger preference for elections in that they
might be more confident of their abilities to work the new system.
The fourth control variable was assessment of elected village leaders. In rural
China, elections of villagers’ committees were designed to serve as a democratic
training class (see O’Brien and Li 2000, pp. 469–470). It has been argued that as
farmers realize that they can use their ballots to oust unpopular village leaders, they
will wish to do the same to government officials (Shi 1999, pp. 394–395). To see if a
preference for popularly electing government leaders might derive from the
appreciation of the instrumental value of village elections, respondents were asked
to compare elected village cadres with appointed ones in four regards: (1) honest
and clean; (2) eager to serve farmers; (3) dare to speak for farmers and resist wrong
decisions of the township government; and (4) handle affairs impartially. The four
indicators constituted a single construct (standardized factor loadings ranged from
.861 to .911). Individuals who found elected village leaders more honest,
responsive, and impartial than appointed ones were expected to have a stronger
preference for popular election of government leaders.
Lastly, social trust was controlled as it might increase confidence in universal
suffrage and ease worries about ‘‘tyranny of the majority’’ (Dahl 1971; Putnam
1993; Reisinger et al. 1994, p. 206). In rural China, social trust may be particularly
important in cultivating a preference for elections, as village politics often involves
long-standing lineage rivalries (Manion 2006, pp. 312–313; Tsai 2007, p. 360). To
measure their trust in other people, respondents were asked whether they would in
general believe what the following people told them, including: (1) fellow villagers;
(2) people introduced by relatives and good friends; and (3) most people in society
(see Appendix). The three indicators tapped the same latent factor (standardized
factor loadings ranged from .691 to .931).
Four demographic features, i.e., gender, age, education, and membership in the
Chinese Communist Party (CCP), were assumed to affect all attitudinal variables
described above. Three regional dummy variables, i.e., Wuping county in Fujian
province, Shangrao county in Jiangxi province, and Yongjia county in Zhejiang
298 Polit Behav (2011) 33:291–311
123
province were included to control for the effect of unobserved differences between
the four selected counties. Dongshan county in Fujian was used as the reference.
The county dummies were assumed to affect all endogenous variables.
Model and Results
A path model was constructed based on the theoretical considerations discussed
above. Exogenous variables listed in the rectangular box on the left-hand side of
Fig. 1 were allowed to influence all endogenous variables on the right-hand side of
the bracket. Distrust in government leaders’ commitment to the public interest was
assumed to affect demand for leadership change, which in turn was assumed to
affect preference for elections. Distrust in leaders was assumed to have no direct
impact on attitudes toward election. Life satisfaction, perception of corruption,
internal political efficacy, assessment of elected village leaders, and social trust
were allowed to directly influence preference for elections. As was explained above,
life satisfaction and perception of corruption were also allowed to influence
preference for elections indirectly by affecting distrust in leaders and demand for
leadership change. The model has five versions, which are identical except that they
estimate, respectively, the relationships between distrust in leaders, demand for
leadership change, and preference for elections regarding five different levels of
government leaders in China.
Gender Age Education CCP membership
Wuping county Shangrao county Yongjia county
Demand for leadership
change
Distrust in leaders
Preference for elections
Perception of corruption
Internal efficacy
Assessing elected cadres
Social trust
Life satisfaction
Fig. 1 The relationship between distrust in government leaders, demand for leadership change, andpreference for elections. Note: Exogenous variables listed in the rectangular box are allowed to affect allendogenous ones on the right side of the bracket. Indicators of latent constructs and error/residual terms ofendogenous variables are suppressed for the sake of clarity
Polit Behav (2011) 33:291–311 299
123
The model is not an exact fit. Analyses of individual datasets showed that the chi-
square statistic (v2) ranged from 35 to 50 with 19 or 20 degrees of freedom (DF),
which indicated an imperfect fit (p \ .01) (not shown). However, three widely-
accepted close fit indexes reported in Table 3, i.e., Comparative Fit Index (CFI),
Tucker-Lewis Index (TLI, also called Non-Normed Fit Index (NNFI)), and Root
Mean Squared Error of Approximation (RMSEA) met conventional cut-off criteria,
which are, respectively, CFI [ .950, TLI [ 950, and RMSEA \ .060 (see Hu and
Bentler 1999; also see Marsh et al. 2004). The goodness-of-fit test results assure us
that estimated coefficients are acceptable descriptions of the data and can be used to
test hypotheses.
The results appearing in Table 3 can be grouped into three general findings. First,
distrust in current government leaders’ commitment to the public interest was
positively associated with demand for leadership change. All else being equal,
individuals who had stronger distrust in government leaders were more likely to
have stronger reservations about voting for the incumbent government leaders if
democratic elections were held. The positive correlation between the two variables
was highly significant regarding all five levels of government leaders (p B .001,
one-sided test).
Second, demand for leadership change was positively associated with a
preference for elections, and the correlation was significant regarding the state
chairman, provincial governor, city mayor, and township head (p B .05, one-sided
test). All else being equal, individuals who had a stronger demand for leadership
change were more likely to agree that government leaders should be popularly
elected. Furthermore, individuals who were more eager to see leadership change
were more likely to agree that elections could be held right away. The effect of
demand for leadership change on preference for elections was statistically
insignificant regarding county head, probably because individuals who had similar
attitudes toward replacing the incumbent county head shared similar ideas about the
desirability and feasibility of popularly electing the county head.
Lastly, distrust in government leaders’ commitment to the public interest had a
positive indirect impact on a person’s preference for elections by positively
affecting demand for leadership change. All else being equal, the more distrustful
one felt about government leaders’ commitment to the public interest, the stronger
one’s demand for leadership change, and the stronger one’s preference for popular
elections. The indirect effect was statistically significant regarding the state
chairman, provincial governor, and township head (p B .05, one-sided test). The
result offered additional support to the hypothesis that distrust in government
leaders might reinforce a preference for elections by enhancing demands for
leadership change.
Due to the lack of longitudinal data, this research cannot establish causal
relationships between distrust in government leaders, demand for leadership change,
and preference for elections. In fact, alternative models that posit different relations
between the key variables would fit the data equally well (see Stelzl 1986). Based on
empirical and ‘‘theoretical grounds’’ (Warwick 1998, p. 597), however, it is
reasonable to believe that distrust in government leaders may have enhanced
demand for leadership change rather than the other way around, and it is also
300 Polit Behav (2011) 33:291–311
123
Ta
ble
3D
eter
min
ants
of
pre
fere
nce
for
elec
tions
and
dem
and
for
lead
ersh
ipch
ange
Ou
tco
me
vari
able
Pre
fere
nce
for
elec
tio
nsa
Th
est
ate
chai
rman
Pro
vin
cial
go
ver
no
rC
ity
may
or
Cou
nty
Hea
dT
ow
nsh
ipH
ead
Ind
epen
den
tva
riab
les
Dem
and
for
lead
ersh
ipch
ange
(wea
kto
stro
ng)
.235***
(.059)
.131**
(.047)
.073
�(.
04
3)
.04
0(.
03
9)
.10
0*
(.0
42
)
Lif
esa
tisf
acti
on
(lo
wto
hig
h)
.01
7(.
05
6)
-.0
18
(.0
61
)-
.00
7(.
05
6)
-.0
02
(.0
57
)-
.03
7(.
06
7)
Per
cep
tio
no
fco
rrup
tio
n(w
eak
tost
rong
)-
.00
1(.
04
9)
.06
2(.
04
3)
.07
1(.
04
2)
.05
6(.
04
5)
.02
7(.
04
9)
Ass
essm
ent
of
elec
ted
vil
lage
lead
ers
(lo
wto
hig
h)
.05
8(.
05
0)
.06
4(.
05
6)
.06
5(.
05
7)
.09
6(.
05
9)
.03
0(.
05
1)
Inte
rnal
effi
cacy
(lo
wto
hig
h)
.03
8(.
05
1)
.04
4(.
05
5)
.04
3(.
05
5)
.13
6*
(.0
62
).1
15
*(.
05
5)
So
cial
tru
st(l
ow
toh
igh
)-
.08
3(.
07
6)
-.0
99
(.0
78
)-
.08
1(.
07
5)
-.1
29�
(.0
70
)-
.12
5�
(.0
69
)
Gen
der
(0=
fem
ale;
1=
mal
e).1
18
(.0
94
).1
73
�(.
08
9)
.14
7�
(.0
78
).0
55
(.0
53
).0
42
(.0
68
)
Ag
e(1
8–
88
yea
rs)
.00
0(.
00
3)
-.0
02
(.0
03
)-
.00
2(.
00
3)
-.0
02
(.0
03
)-
.00
1(.
00
3)
Ed
uca
tio
n(0
–2
1y
ears
)-
.01
2(.
01
3)
-.0
11
(.0
14
)-
.01
1(.
01
4)
-.0
05
(.0
15
).0
05
(.0
14
)
CC
Pm
emb
ersh
ip(0
=n
on
-mem
ber
,1
=m
emb
er)
-.0
34
(.1
07
)-
.08
2(.
10
9)
-.0
31
(.1
04
)-
.00
3(.
11
5)
.00
0(.
09
9)
Wu
pin
gco
un
tyin
Fu
jian
pro
vin
ce(0
=n
o;
1=
yes
).2
18*
*(.
08
4)
.14
2(.
09
2)
.07
8(.
10
6)
.16
6�
(.0
90
).1
79
�(.
09
4)
Sh
ang
rao
cou
nty
inJi
angx
ip
rov
ince
(0=
no
;1
=y
es)
.19
6�
(.1
05
).2
12
*(.
10
4)
.13
5(.
12
0)
.00
7(.
12
6)
.07
6(.
10
9)
Yongji
aco
unty
inZ
hej
iang
pro
vin
ce(0
=n
o;
1=
yes
).3
26*
*(.
11
1)
.40
6*
**
(.1
11
).4
04
**
*(.
11
8)
.47
9*
**
(.0
95
).3
37
**
(.1
03
)
Indir
ect
effe
cts
of
inte
rest
Dis
trust
inle
ader
svia
dem
and
for
lead
ersh
ipch
ange
.063**
(.021)
.023
�(.
01
3)
.01
8(.
01
2)
.00
9(.
01
1)
.01
9�
(.0
10
)
Per
cepti
on
of
corr
upti
on
via
dem
and
for
lead
ersh
ipch
ange
.012
(.016)
.030*
(.012)
.019
�(.
01
1)
.01
0(.
01
2)
.04
1*
(.0
17
)
Pse
ud
oR
2.0
86
.06
0.0
46
.06
5.0
57
Med
iato
rva
riab
leD
eman
dfo
rle
ader
ship
chan
geb
Th
est
ate
chai
rman
Pro
vin
cial
go
ver
no
rC
ity
may
or
Co
un
tyH
ead
To
wnsh
ipH
ead
Ind
epen
den
tva
riab
les
Dis
tru
stin
lead
ers
(wea
kto
stro
ng
).2
69
**
*(.
05
0)
.18
4*
**
(.0
56
).2
56*
**
(.0
62
).2
53*
**
(.0
47
).1
87
**
*(.
05
3)
Polit Behav (2011) 33:291–311 301
123
Ta
ble
3co
nti
nu
ed
Med
iato
rva
riab
leD
eman
dfo
rle
ader
ship
chan
geb
Th
est
ate
chai
rman
Pro
vin
cial
go
ver
no
rC
ity
may
or
Cou
nty
Hea
dT
ow
nsh
ipH
ead
Lif
esa
tisf
acti
on
(lo
wto
hig
h)
.09
5(.
05
8)
.01
4(.
06
2)
-.0
55
(.0
67
)-
.06
2(.
05
3)
-.0
93�
(.0
50
)
Per
cep
tio
no
fco
rrup
tio
n(w
eak
tost
rong
).0
57
(.0
70
).2
40
**
*(.
06
2)
.26
8*
**
(.0
58
).3
02
**
*(.
04
2)
.39
8*
**
(.0
48
)
Gen
der
(0=
fem
ale;
1=
mal
e)-
.00
3(.
09
5)
.01
0(.
09
6)
-.0
32
(.0
89
)-
.01
4(.
08
2)
.03
8(.
08
8)
Ag
e(1
8–
88
yea
rs)
-.0
05
(.0
04
)-
.00
4(.
00
4)
-.0
01
(.0
04
)-
.00
1(.
00
3)
.00
3(.
00
3)
Ed
uca
tio
n(0
–2
1y
ears
)-
.01
3(.
01
4)
.00
5(.
01
0)
.01
4(.
01
1)
.01
5(.
01
1)
.01
2(.
01
3)
CC
Pm
emb
ersh
ip(0
=n
on
-mem
ber
,1
=m
emb
er)
-.1
82
(.1
14
)-
.10
2(.
10
5)
-.1
37
(.1
00
)-
.13
7�
(.0
86
)-
.22
9*
*(.
07
7)
Indir
ect
effe
cts
of
inte
rest
Per
cep
tio
no
fco
rru
pti
on
via
dis
tru
stin
lead
ers
.00
7(.
01
2)
.03
0*
(.0
12
).0
44
**
*(.
01
3)
.06
6*
**
(.0
16
).0
53
**
*(.
01
6)
Lif
esa
tisf
acti
on
via
dis
tru
stin
lead
ers
.00
0(.
01
0)
-.0
07
(.0
08
)-
.01
3(.
01
0)
-.0
23*
(.0
11
)-
.01
6�
(.0
09
)
Pse
ud
oR
2.1
21
.10
6.1
30
.14
7.1
86
Mo
del
fit
ind
exes
CF
I.9
91
.99
1.9
89
.98
9.9
87
TL
I.9
88
.98
8.9
85
.98
5.9
84
RM
SE
A.0
26
.02
6.0
28
.02
9.0
30
No
tes:
Cel
len
trie
sar
eunst
andar
diz
edord
ered
pro
bit
coef
fici
ents
,w
ith
stan
dar
der
rors
inpar
enth
eses
ben
eath
them
.C
oef
fici
ents
of
county
dum
mie
son
the
med
iato
r
var
iab
lear
en
ot
sho
wn.
N=
1,6
00
.M
issi
ng
dat
aar
em
ult
iply
imp
ute
d.
Dat
aar
ew
eig
hte
d�
pB
.05
on
e-si
ded
;*
pB
.05
two
-sid
ed;
**
pB
.01
two-s
ided
;***
pB
.00
1tw
osi
ded
aP
refe
rence
for
elec
tions
ism
easu
red
by
‘‘D
oyou
thin
kth
efo
llow
ing
gover
nm
ent
lead
ers
should
be
dir
ectl
yel
ecte
dby
the
peo
ple
thro
ugh
one-
per
son
on
e-v
ote
?’’
1=
sho
uld
no
t;2
=sh
ould
bu
tco
nd
itio
ns
are
no
tri
pe;
3=
should
and
elec
tions
can
be
hel
dnow
bD
eman
dfo
rle
ader
ship
chan
ge
ism
easu
red
by
‘‘W
ou
ldy
ou
vo
tefo
rth
ein
cum
ben
tle
ader
ifa
dem
ocr
atic
elec
tio
nw
ash
eld
?’’
1=
yes
;2
=u
nsu
re;
3=
no
302 Polit Behav (2011) 33:291–311
123
reasonable that demand for leadership change may have reinforced preference for
elections rather than the other way around. First, it is more plausible to assume that
distrust in government leaders occurs prior to the formation of demand for
leadership change. It is true that an individual who has a pre-formed demand for
leadership change may be inclined to view incumbent leaders with skepticism such
that one’s demand for leadership change enhances one’s distrust in government
leaders. In China, however, this is not a very likely scenario. For one, it is hard to
see how individuals would develop an interest in leadership change without having
doubts about incumbent leaders. Second, most people are socialized into having
faith in political authorities, particularly in national leaders. Political education in
Chinese schools is designed to cultivate faith in the ruling party and the regime.
Cultural myths about ‘‘benign emperors’’ and ‘‘parent-like mandarins’’ may also
help inculcate trust in political authorities. Lastly, the observation of corrupt
behavior such as using power for private gain and practicing favoritism may induce
distrust in government leaders (Chanley et al. 2000; Seligson 2002b; Anderson and
Tverdova 2003; Bowler and Karp 2004), but it does not necessarily create demand
for leadership change. In light of these considerations, it is more likely that an
individual finds government leaders untrustworthy and then develops an interest in
leadership change.
It is also more plausible to assume that demand for leadership change occurs
prior to the formation of a preference for elections in China than otherwise. It is true
that pro-election individuals may be more likely to view unelected government
leaders with distrust and want leadership change because the leaders have not been
chosen through procedures that they believe are appropriate. But this is not very
likely in the Chinese countryside. The surveyed rural residents had no prior
experience with democratic election of government leaders, and Chinese farmers
have relatively little knowledge of democratic political systems and philosophies.
They are unlikely to have been socialized into pro-election views prior to interaction
with government officials. It is more likely that Chinese farmers’ interactions with
government authorities affect their trust in government leaders and their demand for
leadership change, which in turn condition their preference for elections.
The two arguments are supported by two other empirical findings. First, all else
being equal, individuals who were more acutely aware of corruption had stronger
distrust in government leaders at provincial, city, county, and township levels
(p B .001, one-sided test) (not shown in Table 3) as well as stronger demand for
leadership change at the provincial, city, county, and township levels (p B .001,
one-sided test). Moreover, perception of corruption had a significant indirect impact
on preference for popularly electing the provincial governor, city mayor, and county
head by affecting distrust in these three levels of leaders (p B .05, one-sided test).
Although it is possible that one’s preference for elections leads to a demand for
leadership change, which in turn induces a stronger perception of corruption, it is
more likely that perception of corruption occurs prior to distrust in government
leaders, demand for leadership change, and preference for elections.
Second, individuals who were less satisfied with their lives were generally more
distrustful of government leaders’ commitment to the public interest and thereby
had stronger demands for leadership change. The direct impact of life satisfaction
Polit Behav (2011) 33:291–311 303
123
on distrust was significant regarding county and township leaders (p B .01, one-
sided test), and its indirect effect on demand for leadership change by increasing
distrust was also significant regarding county and township leaders (p B .01,
one-sided test), probably because these two levels of government are most directly
responsible for the local economy and governance. This finding supports the
argument that distrust in leaders occurs prior to demand for leadership change. It is
possible that individuals who are more eager to see county and township heads
replaced are more likely to develop distrust in county and township leaders, which
in turn makes them more unsatisfied with their lives. It is, however, much more
plausible that life satisfaction affects confidence in township and county leaders,
which in turn affects demand for leadership change at these two levels.
In sum, it seems that Chinese farmers became doubtful about government
leaders’ commitment to the public interest if they knew or heard that sub-national
authorities were corrupt and/or if they were unsatisfied with their lives. And then as
a consequence of this, they wanted to see untrustworthy leaders replaced and
preferred to effect the leadership change through popular elections.4 This line of
interpretation is supported by the finding that individuals who were more confident
of their abilities to assess and choose county government head tended to have a
stronger preference for elections. The effect of internal efficacy was significant on
preference for popularly electing county and township government heads (p B .01,
one-sided test). This finding suggests that Chinese farmers base their choice of
leadership selection systems on assessment of their own political competence. They
favor popular elections if they feel more confident about working the new system.
Three minor findings of this research are worth mentioning. First, individuals
who found elected village community leaders more honest, impartial, and
responsive than appointed ones were more likely to favor popular election of
government leaders. This effect, though, was statistically insignificant after
controlling for the effects of other variables. The finding thus offers limited support
to the ‘‘spill-over’’ argument that village elections may enhance demand for popular
election of government leaders (Shi 1999).
Second, contrary to what has been observed in industrialized democracies, social
trust in rural China had a negative effect on preference for elections, and the effect
was significant in regard to electing county and township heads (p B .05, one-sided
test). It is possible that interpersonal trust and political trust are positively correlated
in rural China, as social capital theory suggests. The finding offers some support to
the argument that social trust in undemocratic societies is associated with support
for authoritarian patterns of rule (Jamal 2007).
Lastly, preference for elections varied considerably from one location to another.
Farmers from Yongjia county, Zhejiang province had a significantly stronger
preference for popular elections than those from the other three counties, probably
because Yongjia had the highest level of economic development. It will be
4 It ought to be noted that many unobserved factors might have affected distrust in government leaders,
demand for leadership change, and preference for elections. Failed attempts to defend one’s lawful rights
and interests through non-electoral channels such as petitioning and administrative litigation, for
example, may result in stronger distrust in government leaders, stronger demand for leadership change, as
well as stronger preference for elections.
304 Polit Behav (2011) 33:291–311
123
interesting to see how the local economy, political history, and government
performance in different Chinese counties affect political attitudes and values of the
local populace. For this study, the fact that correlations between distrust in current
government leaders, demand for leadership change, and preference for elections
remain highly significant after controlling for the effects of county dummies
suggests that the key findings reported above may be generalizable to other parts of
rural China.
Conclusions
This research finds that distrust in incumbent government leaders’ commitment to
the public interest is associated with stronger demand for leadership change, which
is in turn associated with stronger preference for popular elections of government
leaders in rural China. It argues that distrust in government leaders’ commitment to
the public interest may have enhanced demands for leadership change, which in turn
may have reinforced a preference for popular election as an alternative mechanism
of leadership selection. It further argues that distrust in current government leaders
may have in effect induced a demand for systemic changes. The belief that
government leaders should be popularly elected may indicate disapproval of the
existing leadership selection system, which is characterized by top-down appoint-
ment and limited popular participation. The belief that government leaders should
and can be popularly elected right away may indicate a demand for immediately
changing the existing system of leadership selection. Yet introducing popular
election of government leaders, as noted above, will require a major constitutional
amendment.
As regards the theoretical debate about whether distrust in government
authorities induces demand for systemic changes, the China case suggests that
two mechanisms may be at work. First, whether ordinary people can engineer
leadership changes through existing channels affects the generation of idealistic
wishes for a better alternative system. Second, perceived availability of better and
viable alternatives affects whether an idealistic wish becomes a practical demand.
The distinction of the two mechanisms helps explain why authoritarianism may be
more vulnerable to the corrosive effect of distrust in incumbent government leaders
than democracy is. Under authoritarian rule ordinary people have little institution-
alized recourse to get rid of untrustworthy government leaders, so they are more
likely to feel frustrated with the existing system and develop idealistic wishes for a
better one (cf. Anderson and LoTempio 2002, p. 349). In the meantime, their
idealistic wishes can readily develop into a practical preference for elections, which
have proven elsewhere to be a better and more viable system of leadership selection.
Democracy, by contrast, has somewhat stronger immunity to the corrosive effect of
distrust in incumbent government authorities (see Sztompka 2000, pp. 140–150). It
enables the people to engineer leadership changes through regular elections, thereby
generating less frustration with the overall system and weaker wishes for a better
one. Equally important, it is harder for idealistic wishes for a system better than
democracy to become a practical demand for systemic changes, because better and
Polit Behav (2011) 33:291–311 305
123
viable alternatives may not be in sight (Rose 2007, p. 112; Rose and Shin 2001,
p. 340). Viewed from this perspective, both sides of the theoretical debate may be
correct on whether distrust of freely elected leaders induces a preference for
undemocratic alternatives. Distrust in freely elected government leaders may foster
idealistic wishes for an even better system. Such wishes, however, may never
materialize into practical demands due to the lack of ‘‘real’’ choices (Fraser 1970,
p. 415).
Methodologically, the research reaffirms the importance of specifying the object
of political trust in survey research (e.g., Citrin and Muste 1999; Levi and Stoker
2000). If the dependence on global measures has led to the overestimation of
distrust in the United States (Hill 1981), then it may have the opposite effect in the
study of political trust in China. As is shown in this research, measures of trust need
to be administered at all five levels of government. Asking only about trust in
central leaders would have missed distrust in local government authorities; asking
broadly about local government leaders would have missed the fact that public trust
decreases significantly for every lower level of the government hierarchy. Since
popular trust in Chinese central leaders may remain robust even when trust in local
authorities is declining, focusing exclusively on the central government would
unveil the tip of an iceberg of public distrust in government leaders but fail to
expose its growing body.
Acknowledgements This project was funded by the Research Grants Council of the Hong Kong
Government (Grant No. CUHK2440/06H). I thank my collaborators for administering the survey. For
insightful comments and suggestions, I thank the editors, two anonymous reviewers, Pierre Landry,
Xiaobo Lu, Jeremy Wallace, and especially Kevin O’Brien and Melanie Manion.
Appendix
See Table 4.
Table 4 Description of variables
Mean SD Std. factor loading
Outcome variables
Observed preference for elections
Do you think the following government leaders should be directly elected by the people through one-
person one-vote?
1 = should not; 2 = should but conditions are not ripe; 3 = should and elections can be held now
The state chairman 1.879 .668
Provincial governor 1.911 .657
City mayor 1.946 .659
County head 2.119 .674
Township head 2.264 .670
306 Polit Behav (2011) 33:291–311
123
Table 4 Description of variables
Mean SD Std. factor loading
Explanatory variables
Do you believe the following statements?
1 = fully disbelieve; 2 = disbelieve; 3 = half believe and half doubt; 4 = believe; 5 = fully believe
Latent distrust in central leaders’ commitment to the public interest
They put their own interests before those of farmers 2.703 1.240 .765
They do not care if farmers agree when they make policies 2.745 1.220 .823
They care primarily about the powerful and rich and neglect the
interests of ordinary people
2.598 1.236 .898
Latent distrust in provincial leaders’ commitment to the public interest
They put their own interests before those of farmers 2.901 1.065 .751
They do not care if farmers agree when they make policies 2.860 1.068 .792
They care primarily about the powerful and rich and neglect the
interests of ordinary people
2.746 1.073 .863
Latent distrust in city leaders’ commitment to the public interest
They put their own interests before those of farmers 3.025 1.020 .712
They do not care if farmers agree when they make policies 2.929 1.009 .768
They care primarily about the powerful and rich and neglect the
interests of ordinary people
2.838 1.017 .871
Latent distrust in county leaders’ commitment to the public interest
They put their own interests before those of farmers 3.175 1.069 .741
They do not care if farmers agree when they make policies 3.046 1.049 .791
They care primarily about the powerful and rich and neglect the
interests of ordinary people
2.980 1.069 .863
Latent distrust in township leaders’ commitment to the public interest
They put their own interests before those of farmers 3.223 1.116 .762
They do not care if farmers agree when they make policies 3.103 1.096 .775
They care primarily about the powerful and rich and neglect the
interests of ordinary people
3.001 1.121 .848
Mediator variables
Observed demand for leadership change
Would you vote for the incumbent leader if a democratic election was held? 1 = yes; 2 = unsure;
3 = no
The state chairman 1.459 .574
Provincial governor 1.672 .566
City mayor 1.730 .561
County head 1.792 .605
Township head 1.838 .630
Control variables
Latent life satisfaction
How do you compare your family’s economic situation with 2 years ago?
1 = a lot worse; 2 = somewhat worse; 3 = same; 4 = somewhat better; 5 = a lot better
Polit Behav (2011) 33:291–311 307
123
Table 4 Description of variables
Mean SD Std. factor loading
How do you predict your family’s economic situation in the coming 2 years?
1 = worse; 2 = uncertain; 3 = no change; 4 = better
All in all, are you satisfied with your life?
1 = very unsatisfied; 2 = unsatisfied; 3 = so-so; 4 = satisfied; 5 = very satisfied
Family economic situation compared to 2 years ago 3.969 .865 .796
Family economic situation in the coming 2 years 3.407 .894 .715
Overall level of satisfaction with life 3.611 .784 .650
Latent perception of corruption
In the last decade did the following government officials engage in corruption?
1 = no; 2 = do not know; 3 = it is said so; 4 = yes
Township government officials 2.701 1.075 .728
County government officials 2.550 1.046 .970
Provincial and city government officials 2.427 .984 .759
Latent political efficacy
Do you think you can do the following?
1 = definitely can’t; 2 = can’t; 3 = unsure; 4 = can; 5 = definitely can
Judge a county head’s performance after listening to his/her work
report
3.173 .927 .855
Judge which one is better if two county head candidates debate
each other
3.182 .944 .960
Judge if a county head’s decisions benefit or harm self-interests 3.288 .963 .862
Latent assessment of elected village leaders
How do you compare elected village cadres with appointed ones in the following aspects?
1 = appointed ones are better; 2 = same; 3 = elected ones are better
Honest and clean 2.471 .591 .870
Eager to serve farmers 2.500 .584 .911
Dare to speak for farmers and resist wrong township decisions 2.527 .590 .901
Handle affairs impartially 2.472 .603 .861
Latent social trust
In general, do you believe what the following people tell you?
1 = fully disbelieve; 2 = disbelieve; 3 = half believe and half doubt; 4 = believe; 5 = fully believe
Fellow villagers 3.343 .679 .691
People introduced by relatives and good friends 3.146 .682 .931
Most people in the society 2.995 .735 .727
Demographic variables
Gender (0 = female; 1 = male) .516 .500
Age (range 18–88) 41.30 13.59
Education (0–21 years) 8.02 3.47
CCP membership (0 = non-member; 1 = member) .17 .38
Notes: N = 1,600; missing data are multiply imputed. Row entries are means, standard deviations, and
standardized factor loadings
308 Polit Behav (2011) 33:291–311
123
References
Abramson, P. R. (1972). Political efficacy and political trust among black schoolchildren: Two
explanations. Journal of Politics, 34, 1243–1275.
Abravanel, M. D., & Busch, R. J. (1975). Political competence, political trust, and the action orientations
of university students. Journal of Politics, 37, 57–82.
Achen, C. H. (1978). Measuring representation. American Journal of Political Science, 22, 475–510.
Anderson, C. J., & LoTempio, A. J. (2002). Winning, losing and political trust in America. BritishJournal of Political Science, 32, 335–352.
Anderson, C. J., & Tverdova, Y. V. (2003). Corruption, political allegiances, and attitudes toward
government in contemporary democracies. American Journal of Political Science, 47, 91–109.
Avery, J. M. (2006). The sources and consequences of political mistrust among African Americans.
American Politics Research, 34, 653–682.
Avery, J. M. (2007). Race, partisanship, and political trust following Bush versus Gore (2000). PoliticalBehavior, 29, 327–342.
Balch, G. I. (1974). Multiple indicators in survey research: The concept ‘‘sense of political efficacy’’.
Political Methodology, 1, 1–43.
Berinsky, A. J. (2004). Silent voices: Public opinion and political participation in America. Princeton,
NJ: Princeton University Press.
Bowler, S., & Karp, J. A. (2004). Politicians, scandals, and trust in government. Political Behavior, 26,
271–287.
Bratton, M., & Mattes, R. (2001). Support for democracy in Africa: Intrinsic or instrumental? BritishJournal of Political Science, 31, 447–474.
Burns, J. P. (Ed.). (1989). The Chinese communist party’s nomenklatura system. Armonk, NY: M. E.
Sharpe.
Caddell, P. H. (1979). Trapped in a downward spiral. Public Opinion, 2, 2–7, 52–55, 58–60.
Cai, Y. (2008). Local governments and the suppression of popular resistance in China. China Quarterly,193, 24–42.
Catterberg, G., & Moreno, A. (2006). The individual bases of political trust: Trends in new and
established democracies. International Journal of Public Opinion Research, 18, 31–48.
Chan, H. S. (2004). Cadre personnel management in China: The nomenklatura system, 1990–1998. ChinaQuarterly, 179, 703–734.
Chanley, V. A., Rudolph, T. J., & Rahn, W. M. (2000). The origins and consequences of public trust in
government—a time series analysis. Public Opinion Quarterly, 64, 239–256.
Chen, J. (2004). Popular political support in urban China. Stanford, CA: Stanford University Press.
Citrin, J. (1974). Comment: The political relevance of trust in government. American Political ScienceReview, 68, 973–988.
Citrin, J., & Green, D. P. (1986). Presidential leadership and the resurgence of trust in government. BritishJournal of Political Science, 16, 431–453.
Citrin, J., & Muste, C. (1999). Trust in government. In J. P. Robinson, P. R. Shaver, & L. S. Wrightsman
(Eds.), Measures of political attitudes (pp. 465–532). San Diego: Academic Press.
Craig, S. C., & Maggiotto, M. A. (1982). Measuring political efficacy. Political Methodology, 8, 89–110.
Craig, S. C., Niemi, R. G., & Silver, G. E. (1990). Political efficacy and trust: A report on the NES pilot
study items. Political Behavior, 12, 289–314.
Dahl, R. A. (1971). Polyarchy: Participation and opposition. New Haven, NJ: Yale University Press.
Dalton, R. J. (1985). Political parties and political representation: Party supporters and party elites in nine
nations. Comparative Political Studies, 18, 267–299.
Diamond, L. (2002). Thinking about hybrid regimes. Journal of Democracy, 13, 21–35.
Dong, L. (2006). Direct town elections in China: Latest developments and prospects. Journal ofContemporary China, 15, 503–515.
Finkel, S. E., Muller, E. N., & Seligson, M. A. (1989). Economic crisis, incumbent performance and
regime support: A comparison of longitudinal data from West Germany and Costa Rica. BritishJournal of Political Science, 19, 329–351.
Fraser, J. (1970). The impact of community and regime orientations on choice of political system.
Midwest Journal of Political Science, 14, 413–433.
Hetherington, M. J. (2004). Why trust matters. Princeton, NJ: Princeton University Press.
Polit Behav (2011) 33:291–311 309
123
Hill, D. B. (1981). Attitude generalization and the measurement of trust in American leadership. PoliticalBehavior, 3, 257–270.
Hu, L.-T., & Bentler, P. M. (1999). Cutoff criteria for fit indices in covariance structure analysis:
Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55.
Huber, J. D., & Powell, G. B., Jr. (1994). Congruence between citizens and policymakers in two visions
of liberal democracy. World Politics, 46, 291–325.
Jamal, A. (2007). When is social trust a desirable outcome? Examining levels of trust in the Arab World.
Comparative Political Studies, 40, 1328–1349.
Johnson, I. (2005). Political trust in societies under transformation: A comparative analysis of Poland and
Ukraine. International Journal of Sociology, 35, 63–84.
Kinder, D. R. (1981). Presidents, prosperity, and public opinion. Public Opinion Quarterly, 45, 1–21.
King, G., Honaker, J., Joseph, A., & Scheve, K. (2001). Analyzing incomplete political science data: An
alternative algorithm for multiple imputation. American Political Science Review, 95, 49–69.
Levi, M., & Stoker, L. (2000). Political trust and trustworthiness. Annual Review of Political Science, 3,
475–508.
Li, L. (2002). The politics of introducing direct township elections in China. China Quarterly, 171, 704–
723.
Li, L. (2004). Political trust in rural China. Modern China, 30, 228–258.
Li, L., & O’Brien, K. J. (1996). Villagers and popular resistance in contemporary China. Modern China,22, 28–61.
Li, L., & O’Brien, K. J. (2008). Protest leadership in rural China. China Quarterly, 193, 1–23.
Lipset, S. M., & Schneider, W. G. (1983). The decline of confidence in American institutions. PoliticalScience Quarterly, 98, 379–402.
Little, R. J. A. (1988). A test of missing completely at random for multivariate data with missing values.
Journal of the American Statistical Association, 83, 1198–1202.
Little, R. J. A., & Rubin, D. B. (1987). Statistical analysis with missing data. New York: Wiley.
Manion, M. (1994). Survey research in the study of contemporary China: Learning from local samples.
China Quarterly, 139, 741–765.
Manion, M. (1996). The electoral connection in the Chinese countryside. American Political ScienceReview, 90, 736–748.
Manion, M. (2000). Chinese democratization in perspective: Electorates and selectorates at the town
level. China Quarterly, 63, 764–782.
Manion, M. (2006). Democracy, community, trust: The impact of elections in rural China. ComparativePolitical Studies, 39, 301–324.
Manion, M. (2008). 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, 195, 607–630.
Marsh, H. W., Hau, K.-T., & Wen, Z. (2004). In search of golden rules: comment on hypothesis-testing
approaches to setting cutoff values for fit indexes and dangers in overgeneralizing Hu and Bentler’s
(1999). Structural Equation Modeling, 11, 320–341.
Miller, A. H. (1974a). Political issues and trust in government, 1964–1970. American Political ScienceReview, 68, 951–972.
Miller, A. H. (1974b). Rejoinder to ‘‘comment’’ by Jack Citrin: Political discontent or ritualism?
American Political Science Review, 68, 989–1001.
Miller, W. E. (1979). Confidence in government or a crisis in leadership? Misreading the public pulse.
Public Opinion, 2, 9–15, 60.
Miller, A. H., & Listhaug, O. (1990). Political parties and confidence in government: A comparison of
Norway, Sweden and the United States. British Journal of Political Science, 20, 357–386.
Muller, E. N., & Jukam, T. O. (1977). On the meaning of political support. American Political ScienceReview, 71, 1561–1595.
Muthen, L. K., & Muthen, B. O. (2006). Mplus user’s guide (4th ed.). Los Angeles, CA: Muthen &
Muthen.
O’Brien, K. J. (1990). Reform without liberalization: China’s national people’s congress and the politicsof institutional change. New York: Cambridge University Press.
O’Brien, K. J. (Ed.). (2008). Popular protest in China. Cambridge, MA: Harvard University Press.
O’Brien, K. J., & Li, L. (2000). Accommodating ‘‘democracy’’ in a one-party state: Introducing village
elections in China. China Quarterly, 162, 465–489.
Putnam, R. D. (1993). Making democracy work: Civic traditions in modern Italy. Princeton, NJ:
Princeton University Press.
310 Polit Behav (2011) 33:291–311
123
Reisinger, W. M., Miller, A. H., Heslik, V., & Maher, K. H. (1994). Political values in Russia, Ukraine
and Lithuania: Sources and implications for democracy. British Journal of Political Science, 24,
183–223.
Rose, R. (2007). Learning to support new regimes in Europe. Journal of Democracy, 18, 111–125.
Rose, R., & Shin, D. C. (2001). Democratization backwards: The problem of third wave democracies.
British Journal of Political Science, 31, 331–354.
Rubin, D. B. (1987). Multiple imputation for nonresponse in surveys. New York: Wiley.
Sarsfield, R., & Echegaray, F. (2006). Opening the black box: How satisfaction with democracy and its
perceived efficacy affect regime preference in Latin America. International Journal of PublicOpinions Research, 18, 153–173.
Schafer, J. L. (1997). Analysis of incomplete multivariate data. London: Chapman and Hall.
Schafer, J. L., & Olsen, M. K. (1998). Multiple imputation for multivariate missing-data problems: A data
analyst’s perspective. Multivariate Behavioral Research, 33, 545–571.
Schedler, A. (2006). Electoral authoritarianism: The dynamics of unfree competition. Boulder, CO:
Lynne Rienner Publishers, Inc.
Seligson, M. A. (2002a). Trouble in paradise? The erosion of system support in Costa Rica, 1978–1999.
Latin American Research Review, 37, 160–185.
Seligson, M. A. (2002b). The impact of corruption on regime legitimacy: A comparative study of four
Latin American countries. Journal of Politics, 64, 408–433.
Shi, T. (1999). Village committee elections in China: Institutionalist tactics for democracy. WorldPolitics, 51, 385–412.
Shi, T. (2001). Cultural values and political trust: A comparison of the People’s Republic of China and
Taiwan. Comparative Politics, 33, 401–419.
SPSS Inc. (2008). SPSS missing values 17.0. Chicago, IL: SPSS Inc.
Stelzl, I. (1986). Changing a causal hypothesis without changing the fit: Some rules for generating
equivalent path models. Multivariate Behavioral Research, 21, 309–331.
Sztompka, P. (2000). Trust: A sociological theory. New York: Cambridge University Press.
Tang, W. (2005). Public opinion and political change in China. Stanford, CA: Stanford University Press.
Tsai, L. L. (2007). Solidary groups, informal accountability, and local public goods provision in rural
China. American Political Science Review, 101, 355–372.
Waldron-Moore, P. (1999). Eastern Europe at the crossroads of democratic transition: Evaluating support
for democratic institutions, satisfaction with democratic government, and consolidation of
democratic regimes. Comparative Political Studies, 32, 32–62.
Walker, K. L. (2006). Gangster capitalism and peasant protest in China: The last twenty years. Journal ofPeasant Studies, 33, 1–33.
Warwick, P. V. (1998). Disputed cause, disputed effect: The postmaterialist thesis re-examined. PublicOpinion Quarterly, 62, 583–609.
Weatherford, M. S. (1984). Economic stagflation and public support for the political system. BritishJournal of Political Science, 14, 187–205.
Weitz-Shapiro, R. (2008). The local connection: Local government performance and satisfaction with
democracy in Argentina. Comparative Political Studies, 41, 285–308.
Zhu, J.-H. (1996). ‘‘I don’t know’’ in public opinion surveys in China: Individual and contextual causes of
item non-response. Journal of Contemporary China, 5, 223–245.
Polit Behav (2011) 33:291–311 311
123