Social Class,Ingroup–Outgroup Comparison,and Citizen Evaluations: Is User Satisfaction Linked to Outcome Disparities?

Abstract

Research on citizen satisfaction has emphasized the role of service quality (including expectations for quality) in shaping citizens’ evaluations of public services. This article considers an understudied but important aspect of public service delivery—equity—and investigates how disparities in service outcomes between disadvantaged and advantaged groups affect citizens’ evaluations of service providers. This study also examines whether citizens with different socioeconomic status (SES) have different perceptions toward the outcome disparities. Using individual-level data from secondary schools, we find that service users appear to recognize and care about a performance gap among social strata. Even when a student’s individual outcome is held constant, satisfaction varies with the school-level performance gap between high-SES and low-SES students. This finding suggests that service users are concerned about not just their outcomes but also the relative positioning of outcomes for those of their own social group compared with other groups.

Keywords

citizen satisfaction social equity performance gap ingroup–outgroup

Introduction

Citizens’ opinions about public service quality have been a central concern of modern governments, in part because citizens’ judgments about public services have important implications for democracy and accountability. For example, citizens’ positive assessments of service performance may lead to general support for government policy, and to the extent that public officials are motivated to seek such public support, they have an incentive to ensure public services operate well enough to satisfy their constituents. Noting such implications, scholars are beginning to devote more attention to studying what drives citizen satisfaction with government and what information citizens use to evaluate service performance. We know that citizens obtain information about government from a variety of sources, including the media, public meetings, and interactions with other citizens; but they likely draw upon firsthand experiences as their primary source of information when evaluating their governments. These bureaucratic encounters probably constitute the most tangible expression of government that citizens experience with any regularly, making them an important phenomenon for scholars and government officials to understand.

Much of the literature on citizen assessments of government performance has explored whether citizen satisfaction is associated with administrative measures of service quality. Some studies find that citizen satisfaction is uncorrelated or only weakly correlated with archival indicators of service quality, raising questions about the validity of citizens’ evaluations (e.g., Stipak, 1979, 1980). Others, however, find a positive and significant relationship between a broad range of administrative measures of service quality and citizen satisfaction, suggesting citizens take into account meaningful information about performance when forming evaluations of service quality (e.g., Favero & Meier, 2013; Van Ryzin, Immerwahr, & Altman, 2008). Given the mixed findings, our current understanding of how citizens evaluate the government organizations they regularly interact with is relatively limited.

Furthermore, several aspects of citizen evaluations remain unexamined. First, previous studies mainly focus on the effect of service quality on citizen satisfaction and pay less attention to another important aspect of public service delivery—equity. Public administration literature has long emphasized the multidimensional nature of government performance (L. B. Andersen, Boesen, & Pedersen, 2016; Rainey, 2009) and the potential trade-off between equity and efficiency (Jimenez, 2014; Meier, Wrinkle, & Polinard, 1999; Wilson, 1989). How citizens respond to the equity issue (e.g., disparities in service access or outcomes), therefore, is a theoretically and practically important question.

Second, little research has explored the potentially heterogeneous responses of citizens to performance information when evaluating public services. Researchers have noted differences in average satisfaction levels among demographic groups responding to surveys about public services (see B. Brown & Benedict, 2002; K. Brown & Coulter, 1983), but little is known about why some groups appear more satisfied than others. Citizens with low socioeconomic status (SES), for instance, may be less likely to experience high-quality services, resulting in lower satisfaction with public services. Or, citizens of different socioeconomic classes may experience a similar quality of public services, but they may reach different conclusions about a service because their evaluations are shaped by different sets of values or expectations.

This article aims to advance public administration literature by addressing these questions: How do citizens respond to disparities in service outcomes between disadvantaged and advantaged groups? Do high-SES and low-SES citizens have different responses to disparities in service outcomes? Drawing upon work in sociology and social psychology, which suggests social groupings affect how individuals perceive and respond to one another, this research hypothesizes that citizens interpret service provision through the lens of ingroup–outgroup comparisons. Specifically, we argue citizens recognize the gap between service outcomes of their own social group and that of others and express dissatisfaction when members of their own social group consistently experience worse outcomes than members of an outgroup.

The empirical subjects of this study are individual parents and students in secondary schools. Using an individual-level analysis, we investigate whether and how parents’ and students’ satisfaction with schools are associated with disparities in educational outcomes between high-SES and low-SES groups. The evidence from more than 55,000 parents and students over several years highlights the importance of considering how service outcomes vary among social strata. Service users appear to notice and care when their own group does not experience the same level of positive service outcomes as other groups served by the same organization. This finding suggests that citizens care about not just their own outcomes but also those of others with whom they socially identify.

Citizen Satisfaction and Government Performance

Increasing pressure to improve efficiency and promote accountability in the era of New Public Management has coincided with growing demand for reliable and valid measures of performance (Hjortskov, 2016; Osborne & Gaebler, 1992), driven by a belief that measuring performance is the first step toward improvement. No universally accepted performance measure, however, exists in the public sector as public agencies deal with multiple stakeholders (Boyne, 2003; Walker, Boyne, & Brewer, 2010) and have multiple performance dimensions (Boyne, 2002; Rainey, 2009). In addition, many public services are difficult to quantify, and it is hard to find comparable quality indicators that transfer across policy areas (Andrews, Boyne, & Walker, 2006; Rainey, 2009). Citizen assessments of public service, therefore, have been a valuable performance data source, especially in the public sector.

Citizen satisfaction is a potentially beneficial metric because it can capture elements of public service that are important to citizens (Andrews et al., 2006; Brudney & England, 1982) while reflecting information about the actual quality of public service. Empirical studies have supported the notion that citizens’ perceptual evaluations are correlated with archival performance indicators in a variety of settings, including education (Charbonneau & Van Ryzin, 2012; Favero & Meier, 2013; Song & Meier, 2018), health care (Cheon, Song, McCrea, & Meier, 2019; Jha, Orav, Zheng, & Epstein, 2008), and street cleanliness (Van Ryzin et al., 2008). These findings suggest that citizens are able to conduct meaningful evaluations of public service.

Surveys designed to gauge citizen satisfaction with government services also play a substantial role in making citizens’ voices heard (Kelly & Swindell, 2002, 2003) and inducing the government to be responsive. Although public choice theory suggests that public service provision in a democracy will be congruent with the preferences of the median voter, governments often leave some citizen demands unsatisfied (Salamon, 1987). Citizen surveys can function as an important instrument for those unsatisfied citizens to indicate their preferences. In addition to providing vital information on the desired service outcomes and processes, citizen surveys also function as a means of political participation allowing citizens to communicate with government officials and governing elites (Verba, 1996). Compared with other means of political engagement (e.g., voting, political activism), barriers to participating in a survey are relatively low, enabling participation from a broad slice of the public and potentially capturing the views of groups that often lack political voice (Verba, 1996).

Despite the importance and popularity of citizen surveys, concerns exist regarding their reliability and validity due to their fundamental subjectivity. Scholars argue that using citizen surveys as a performance indicator can be problematic because citizens may have limited knowledge about service provision, especially for services that do not directly affect them (Andrews et al., 2006; Stipak, 1979, 1980). Furthermore, survey questions on performance are sometimes vague and ill-defined (L. B. Andersen, Heinesen, & Pedersen, 2016), and we often do not know what information citizens use when assessing performance (Stipak, 1979). Some empirical studies also suggest that citizens’ perceptual evaluations may not reflect actual service quality (e.g., K. Brown & Coulter, 1983; Kelly, 2003). These studies generally assume that archival performance measures are objective and accurately reflect service quality, although these assumptions may not always hold (Andrews et al., 2006; Schachter, 2010).

Another weakness of subjective satisfaction measures is that they are vulnerable to bias. Recent research finds that citizens’ cognitive biases shape their judgments about government performance (e.g., S. C. Andersen & Hjortskov 2015; James & Van Ryzin, 2017; Marvel, 2015). Marvel (2015), for example, shows that citizens implicitly relate public agencies to inefficiency and inflexibility, which can make citizens’ evaluations unconsciously biased (especially in a negative way). S. C. Andersen and Hjortskov (2015) present a dual-process model of cognitive thought and show that citizen perceptions of performance are better explained by an intuitive model rather than a rational model. Citizen satisfaction can be shaped by information cues about performance (James, 2011) and by the choice of a positive or negative label description (Olsen, 2015). Partisanship can also influence the way citizens assess public services when it comes to politically contentious areas (James & Van Ryzin, 2017) or when there is clear responsibility among political actors (Jilke & Baekgaard, 2019). These significant framing effects raise questions about whether citizens’ perceptions of performance accurately represent service performance.¹

The Role of Performance Information in Citizen Satisfaction

Although the pros and cons of using citizen satisfaction as a performance indicator have been discussed, a growing number of governments in the United States and other countries are using citizen surveys to assess the quality of public services (Van Ryzin, Muzzio, Immerwahr, Gulick, & Martinez, 2004). Scholars have also provided a theoretical framework to better understand how citizen satisfaction is formed. The Expectancy Disconfirmation Model (EDM) has become the predominant framework for studying citizen satisfaction with public services (e.g., Jacobsen, Snyder, & Saultz, 2015; James, 2007; Poister & Thomas, 2011; Van Ryzin, 2004, 2006, 2013). The model suggests citizens form judgments by comparing perceived service quality with prior expectations—when service quality falls short of expectations, negative performance-expectations gap will lead to dissatisfaction whereas when service quality exceeds expectations, the positive gap will lead to satisfaction (Morgeson, 2013; Van Ryzin, 2013).

Empirical studies have supported this model in various contexts such as the United States (Morgeson, 2013; Roch & Poister, 2006; Van Ryzin, 2004, 2006), England (James, 2007, 2011), Denmark (Hjortskov, 2018), and Mexico (Petrovsky, Mok, & León-Cázares, 2017). Going one step further, recent research considers multiple types of expectations. James (2011) categorizes expectation into two types—predictive expectation of how the service will be and normative expectation of how the service should be. So far, empirical work suggests that normative interpretations of expectations are particularly salient for citizens (Hjortskov, 2016; see also Jacobsen et al., 2015).

Whereas scholars have extensively studied the effects of expectations on satisfaction, until recently little attention had been paid to the role of comparative performance information in shaping citizen satisfaction. Among the few such studies, Olsen (2017) highlights the relative nature of performance assessment and examines how citizens use historical and social reference points. The findings show that social aspiration has a much greater impact than historical aspiration in citizens’ evaluations. In a similar vein, Barrows, Henderson, Peterson, and West (2016) show that citizens’ evaluations of performance can be influenced by how local service quality (local school performance) ranks relative to that of other reference groups (school performance at the state, national, or international level). These findings suggest that a citizen’s performance evaluation is a fundamentally relative process.

Discussions of equity are notably absent from the citizen satisfaction literature, despite growing attention in recent decades to social equity as a fourth “pillar” of public administration (Frederickson, 1990). Even recent studies examining citizens’ use of comparative performance information have not considered whether citizen attention to such information is motivated by concerns about social equity. Another gap in the literature is the lack of consideration given to the potentially heterogeneous responses of citizens when it comes to evaluating government services. In particular, we know little about whether citizens of different demographic backgrounds respond differently to performance information. This research gap is surprising given a substantial amount of evidence demonstrating the significant effect of socioeconomic and demographic characteristics on citizens’ average satisfaction levels with public service (B. Brown & Benedict, 2002; K. Brown & Coulter, 1983).² It remains unexplored whether these topline differences among demographic groups are driven by divergent service experiences or by different methods of evaluating a similar service experience.

Social Groups as a Lens for Evaluating Performance

Service equity can be conceptualized in different ways. For example, one approach might consider equity as a lack of significant individual-level differences in service outcomes while another conceptualization might focus on group-level differences and see whether different social groups achieve similar levels of performance outcomes. A large literature in sociology and social psychology points to social groups as important units that help to define how individuals experience and evaluate social interactions (Allport, 1954; Tajfel, 1986). Social groupings—such as race/ethnicity, gender, and economic status—can form a strong basis for operationalizing social identities. These identities affect individuals’ social ordering and behavior, and they likely extend to selecting and evaluating public services. Social groups may affect not only how individuals evaluate government services but also how they respond to social disparities in service outcomes.

Particularly, theories of ingroup and outgroup comparison provide a useful lens to understand how individuals compare their own outcomes with those of others. People categorize others into several social groups based on their social identities (Allport, 1954; Tajfel, 1986) and often show favoritism to members of their own group (ingroup) compared with members of another group (outgroup) (Y. T. Lee, 1993). According to social identity theory, ingroup preference originates from the tendency for a member of a certain group to view his or her own group with more positive group sentiment than other groups (Tajfel, 1986). People adopt ingroups as psychologically primary groups with regard to intimacy and affection, which naturally leads to a positive perspective on ingroups (Allport, 1954). Ingroup preferences also affect behaviors, as in the case of residential sorting. The literature on self-segregation indicates that people often move into neighborhoods largely composed of others who share their own race and/or SES (Bohmert & DeMaris, 2015; Burgess, Wilson, & Lupton, 2005; Lewis & Hamilton, 2011; Sigelman, Bledsoe, Welch, & Combs, 1996). In sum, ingroup biases and social identity theory suggest that individuals tend to have more positive perceptions of their own ingroup compared with an outgroup. Thus, we expect citizens will tend to give more favorable evaluations to service providers when their ingroup members perform better than outgroup members.

Empirical studies support the notion that ingroup preferences affect citizens’ perceptions of and preferences for public services. In the context of education, parents of different SES tend to have different preferences for the demographic composition of schools (Weiher & Tedin, 2002). Specifically, parents tend to prefer schools where most students share their own demographic characteristics, which often leads to school segregation in the United States (Billingham & Hunt, 2016; Glazerman & Dotter, 2017). This ingroup preference appears to be stronger for parents of higher SES and in the racial majority (Billingham & Hunt, 2016; Buckley & Schneider, 2009; Saporito, 2003). Billingham and Hunt (2016), for example, show that White parents are particularly concerned about the racial composition of their children’s school in the District of Columbia, and they prefer schools composed mainly of White students.³ Beyond an ingroup preference regarding the composition of a student body, there is some evidence that parents care about performance gaps between their own group and other groups. Using survey experiments with a nationally representative sample of U.S. adults, Valant and Newark (2016) find that individuals tend to care more about a test score gap that is directly related to their own group. African Americans show more concern about a Black–White achievement gap than about a Hispanic–White achievement gap, and low-income individuals tend to care more about poor–wealthy gaps than race-based gaps.⁴

Given the salience of group identities for shaping individuals’ perceptions in a number of social settings, we argue that citizens will look at public service performance through the lens of social group identity. Citizens belonging to low social-status groups may be particularly sensitive to service outcome inequality because such groups often experience subpar service from governments. Specifically, based on the ingroup preference, this study hypothesizes that a performance gap in outcomes for high-SES versus low-SES groups has a relatively strong association with the satisfaction of low-SES citizens (who are assumed to generally experience worse outcomes than high-SES citizens). On the contrary, the performance gap is expected to have a relatively weak relationship with satisfaction for high-SES citizens.

Hypothesis 1: Citizens from a disadvantaged group will be less satisfied with a service provider when there is a larger service outcome gap between their own group and other groups.

Hypothesis 2: Citizens from an advantaged group will be relatively indifferent to a service outcome gap among social groups.

Empirical Context

The empirical context of this article is secondary schools in South Korea. Korean schools provide a suitable setting for testing our hypotheses for several reasons. First, SES and academic achievement are highly salient issues in Korea because a higher level of educational attainment has played a significant role in obtaining privileged occupations and high income in society. For high-SES groups, educational attainment has been regarded as a means of retaining their SES (S. Lee & Brinton, 1996), while education has provided opportunities for low-SES groups to obtain better social status.⁵ Given the importance of education, the South Korean government has enforced a centralized equalization policy to try to ensure equal educational opportunities (Seth, 2002); however, high-SES parents express preferences for the stratified school system for their children to receive high-quality education (Park, 1988), and inequalities in educational outcomes remain a serious concern.

Second, there are high levels of transparency regarding performance in Korean schools, and this setting allows students and parents to share performance information. Korean parents and students take academic achievement very seriously (Mintrom & Walley, 2013), and most schools provide parents and students with not only an individual student’s standardized test results but also the school’s average scores. Also, the government has pursued a uniformity of education to ensure equal educational opportunity, and this system makes it easier for parents and students to compare their achievement to that of others. These unique characteristics bring heightened salience to students’ relative positions on educational achievement.

Third, studying citizen satisfaction in the Korean context can enhance the external validity of the theory due to its political and cultural setting (O’Toole & Meier, 2014). South Korea has a centralized education system with limited school choice, and it provides a unique opportunity to test whether theories developed from decentralized contexts can be generalizable to centralized contexts. In addition, unlike many Western countries with an individualist culture, Korea has a collectivist culture which endorses ingroup preferences (Chung & Jin, 2011). The collectivism increases positive perceptions of ingroups compared with outgroups (Chung & Jin, 2011); therefore, examining the role of ingroup–outgroup social comparisons in shaping citizen satisfaction in Korea can also contribute to the literature.

Data and Methods

This study uses an individual-level education data set from Seoul City (Capital City of South Korea). Since 2010, the Seoul Metropolitan Office of Education has conducted the Seoul Education Longitudinal Study (SELS) to collect data from all levels of schools and multiple stakeholders to facilitate effective education policy. During the 2010-2015 data collection, students, parents, teachers, and principals at every sampled school were asked to take the annual survey. Students in the sample take an annual standardized test in Korean, English, and math in early July of each year (the tests covering material taught from March to June). In addition to the survey responses and test scores, the SELS data include an archival database on school characteristics. All sampled schools provide various information, including the number of students and teachers, student demographic composition, school facilities and equipment, school budget and resources, and school types. These statistics are also reported annually.

The SELS is a very representative sample of the Seoul City school population. Sampling is conducted using a stratified two-stage sampling technique. In the first stage, the number of schools is determined based on the total number of schools in each school district. In the next stage, schools are randomly selected from within each district, and students are also randomly selected within each school. The final sample of SELS includes 289 schools, 16,059 students, 15,603 parents, and 1,915 teachers in the first year of data collection.

Following the first year of data collection, the Office of Education returned to each sampled student, parent, and school to gather additional data. Unique school and student IDs allow us to link individual-level information for students and parents across years. The response rates of the surveys are very high. In the second year, the response rate was 92.0% for elementary school students, 95.7% in middle school, 94.8% in general high schools, and 94.0% in vocational high schools. The response rate for the third year is above 85.0% and for all following years is above 70.0%.

An advantage of SELS data is that information is collected at both the individual level (students and parents) and the aggregate level (schools). This data structure permits us to investigate how service users’ perceptions are affected by both individual-level and aggregate-level factors. The data include individuals’ SES (parents’ educational attainment) as well as student test scores, which enables us to incorporate a more precise measure of citizen reactions to performance discrepancy between high-SES and low-SES groups. The individual-level measure can also avoid aggregation bias, whereas the aggregated measure of satisfaction may disguise substantial dissatisfaction by some groups.

The unit of analysis is the student-year for all models (both parent and student models). This study uses multiple regression analysis. Our analysis includes fixed effects for years and standard errors clustered at the school level to address potential serial correlation and heteroscedasticity.

Measures

Parent and Student Satisfaction

The subjective well-being of students and parents is significantly influenced by school policies and culture (Beck, Maranto, & Lo, 2014). This study employs student and parent satisfaction as our two dependent variables and tests whether they vary depending on the performance gap in schools. For each year of data collection, parents and students answer how satisfied they are with the school on a 5-point scale (ranging from strongly disagree to strongly agree). Specifically, parents are asked to evaluate schools regarding student learning, course variety, career consultation, educational facilities, school safety, and overall satisfaction. These items show high internal consistency (Cronbach’s α = .9) and our factor analysis finds that all items are loaded on a single factor with an eigenvalue of 4.66 (for details, see Table A1 in the appendix). Although the parent survey asks about various aspects of schools, the student survey is relatively limited.⁶ Students answer the question “Are you satisfied with your school?” using a 5-point scale, and we measure student satisfaction with this item (see Table A2 in the appendix). It is notable that using a single item for measuring satisfaction has been a common method in prior studies (James & Moseley, 2014; Morgeson, 2013; Van Ryzin, 2004).

Performance Gap by Social Class

The key independent variable of this study is the performance gap by social class. The performance gap is a significant measure of equity in schools because high-SES students tend to achieve higher test scores in general, and students with higher scores are more likely to be assigned to advanced classes with more resources and have further opportunities. Although the performance gap is not officially reported, it is highly plausible that students and parents perceive the gap as average test scores are, in most cases, reported or noticed publicly. Also, students and parents interact and share such information with one another.

To measure the performance gap between high-status and low-status groups, social class should be first operationalized. The literature on SES suggests that education level is one of the most stable aspects of social status because it tends to remain relatively constant over time (in adulthood) and is highly correlated with income level (Sirin, 2005).⁷ Our data include both father’s and mother’s education level for each student, and we create a binary indicator of whether a student’s parents have at least a 4-year bachelor’s degree. The social class variable is coded as 1 if at least one parent (either father or mother) has a bachelor’s degree or higher degree (master or doctorate), and 0 otherwise. Using this coding, about 55% of our sample is considered high SES and 45% is considered low SES.

Finding a solid performance indicator is the second step in creating the performance gap measure. Although educational performance has multiple dimensions, academic achievement has been one of the most important performance indicators. This indicator is a particularly salient aspect of student performance in the Korean education system because standardized test scores influence further educational opportunities and future careers (Seth, 2002). To measure student academic performance, we calculate each student’s average test score among three subjects (Korean, English, and math) and then standardize this variable. A measure of school-level performance by social class is created by aggregating the individual-level performance by each group. Specifically, we calculate the average score in School A among students who have at least one parent with a bachelor’s degree and also the average score for students in that school whose parents do not have a bachelor’s degree. Then, we create a measure of performance discrepancy by subtracting the mean score of low-SES students ( $P_{low}$ ) from the mean for high-SES students ( $P_{high}$ ). Simply,

Performance gap = P_{high} - P_{low} .

Figure 1 shows the distribution of our performance gap measure. As the previous literature suggests, high-SES students tend to achieve higher test scores than low-SES students which results in mostly positive values in the gap measure.

Figure 1.

Distribution of performance gap by social class.

Control Variables

We include four sets of variables to control for other factors that can affect student and parent satisfaction. We first control for socioeconomic factors based on the previous literature suggesting SES plays an important role in shaping citizen satisfaction (B. Brown & Benedict, 2002; K. Brown & Coulter, 1983). The logged household income and parent education (averaged across both parents) at the individual level are included.

The second group of controls captures school resources. Highly educated and experienced teachers are important human resources that schools can use to achieve high performance. Teacher education is measured by the percentage of teachers with either a master’s or doctorate degree in each school, while teacher experience is measured by the school-level average years of teaching experience. Also, the models include the percentage of full-time teachers and class size (a student-teacher ratio) to control for school resources.

A third set of controls reflects various school characteristics. The percentage of low-achievement students (in Korean, English, and math) is included to capture clientele characteristics of the school.⁸ School size is measured by the total number of students (logged). We also control for school type (elementary, middle, general high, and vocational high) and school ownership (public vs. private).

The last group of controls includes objective performance measured both at the student level and the school level. Recent studies suggest that objective performance indicators (e.g., administrative records of performance) are significantly associated with citizen satisfaction (Charbonneau & Van Ryzin, 2012; Favero & Meier, 2013; Van Ryzin et al., 2008). We control for students’ standardized test scores at the individual level. We also include a school-level performance measure, which is the mean score within each school (Lyons, Lowery, & DeHoog, 1992).⁹ Descriptive statistics of all variables are shown in Table A3 in the appendix.

Findings

Table 1 shows the results for our models of parent satisfaction. The first column shows pooled results for both high-SES and low-SES groups while the second and third columns split the samples by SES. The gap between the performance of high-SES and low-SES students is not a significant predictor of parent satisfaction in any of the models. The largest (but still statistically insignificant) coefficient for the gap measure is in the third model, which is restricted to low-SES parents—those who we predicted would be most responsive to the performance gap.

Table 1.

Performance Gap and Parent Satisfaction.

Independent variables	Pooled model	SES
Independent variables	Pooled model	High SES	Low SES
Performance gap	−0.005 (0.020)	0.003 (0.022)	−0.024 (0.026)
Household income	−0.020* (0.008)	0.005 (0.013)	−0.042*** (0.011)
Parents education	−0.002 (0.007)	0.020 (0.015)	−0.061*** (0.018)
Average teacher education	0.004*** (0.001)	0.003** (0.001)	0.004*** (0.001)
Average teacher experience	−0.005^† (0.003)	−0.010** (0.004)	0.000 (0.003)
Full-time teachers	0.001 (0.002)	0.002 (0.003)	0.001 (0.002)
Class size	−0.012* (0.005)	−0.011^† (0.006)	−0.011* (0.006)
Low-performance students	−0.020*** (0.004)	−0.027*** (0.006)	−0.017*** (0.003)
Number of students	−0.103** (0.038)	−0.148** (0.046)	−0.047 (0.038)
Public school	−0.190*** (0.030)	−0.198*** (0.039)	−0.176*** (0.032)
Middle school	−0.483*** (0.044)	−0.448*** (0.051)	−0.518*** (0.050)
General high school	−0.524*** (0.062)	−0.409*** (0.077)	−0.657*** (0.064)
Vocational high school	−0.173* (0.083)	0.026 (0.101)	−0.330*** (0.088)
Student test scores	0.063*** (0.006)	0.071*** (0.008)	0.052*** (0.009)
School mean scores	0.182*** (0.029)	0.173*** (0.032)	0.151*** (0.035)
F	49.324	42.056	36.394
Prob. > F	.000	.000	.000
Adjusted R²	.086	.093	.081
N	55,814	30,880	24,934

Note. The dependent variable is parent satisfaction. Reference category for school type is elementary school. Year fixed effects and constant not shown. Standard errors are clustered at the school level and shown in parentheses. SES = socioeconomic status.

†

p < .10. *p < .05. **p < .01. ***p < .001. (Two-tailed test)

Several of the control variables, however, are significantly related to parent satisfaction with schools. Higher income parents are generally less satisfied, perhaps because such parents have higher expectations of schools. The average parent education level does not appear to be associated with satisfaction in the pooled model, although the third model indicates that higher levels of education are associated with lower levels of satisfaction within the low-SES group (those where both parents do not have a bachelor’s degree). Thus, there might be a nonlinear effect of educational attainment on parent satisfaction. Parents are more satisfied when a school’s teachers have higher educational attainment, but teacher experience appears to have a null or negative effect on satisfaction, depending on the model. The percentage of full-time teachers has no significant effect, but parents are less satisfied in schools with larger class sizes. Parents are generally less happy at schools with more low-performing students, larger schools, and public-sector schools. They are generally more satisfied with elementary schools compared with middle and high schools.

Consistent with prior findings, standardized test scores have a strong and positive effect on parent satisfaction. In all three models, the average test score at the school is a stronger predictor of a parent’s satisfaction than their own child’s score. In the pooled model, the effect size of school-level average scores is 3 times that of the individual-level test score measure.

In Table 2, we explain variation in student satisfaction levels. The test score gap appears to affect student evaluations of the school. In the pooled model, the gap between low-SES and high-SES average scores is significant at the .10 level ( $β$ = −.034, p < .094). The negative coefficient indicates that student satisfaction decreases as the disparity between low-SES and high-SES students grows. When the sample is split, we can see that this result appears to be driven primarily by the low-SES students. The third model shows that the gap is significantly and negatively related to the satisfaction of low-SES students. The magnitude of this coefficient ( $β$ = −.059, p < .023) is relatively small, but it suggests that low-SES students are marginally less satisfied with their schools when the performance gap increases. For high-SES students (Model 2), the coefficient is still negative but smaller in magnitude and statistically insignificant.

Table 2.

Performance Gap and Student Satisfaction.

Independent variables	Pooled model	SES
Independent variables	Pooled model	High SES	Low SES
Performance gap	–0.034 ^† (0.021)	–0.021 (0.024)	–0.059* (0.026)
Household income	−0.013^† (0.007)	−0.001 (0.010)	−0.024* (0.010)
Parents education	0.008 (0.006)	0.017 (0.011)	−0.009 (0.014)
Average teacher education	0.002** (0.001)	0.002 (0.001)	0.003** (0.001)
Average teacher experience	−0.005^† (0.003)	−0.009** (0.003)	−0.002 (0.003)
Full-time teachers	0.000 (0.002)	−0.000 (0.002)	0.001 (0.002)
Class size	−0.019*** (0.004)	−0.014** (0.005)	−0.025*** (0.005)
Low-performance students	−0.005^† (0.003)	−0.009* (0.004)	−0.004 (0.003)
Number of students	−0.020 (0.036)	−0.094* (0.044)	0.064 (0.040)
Public school	0.003 (0.031)	0.017 (0.039)	−0.009 (0.034)
Middle school	−0.704*** (0.044)	−0.650*** (0.054)	−0.758*** (0.045)
General high school	−1.005*** (0.055)	−0.907*** (0.069)	−1.106*** (0.061)
Vocational high school	−0.730*** (0.081)	−0.704*** (0.106)	−0.825*** (0.090)
Student test scores	0.105*** (0.006)	0.104*** (0.008)	0.102*** (0.009)
School mean scores	0.158*** (0.025)	0.152*** (0.027)	0.149*** (0.033)
F	97.974	70.023	72.165
Prob. > F	.000	.000	.000
Adjusted R²	.160	.163	.151
N	56,743	31,246	25,497

Note. The dependent variable is student satisfaction. Reference category for school type is elementary school. Year fixed effects and constant not shown. Standard errors are clustered at the school level and shown in parentheses. SES = socioeconomic status.

†

p < .10. *p < .05. **p < .01. ***p < .001. (Two-tailed test)

Relationships with control variables are fairly similar to what we saw for parents, which is not surprising considering that we would expect parent and student opinions to be strongly intertwined. Unlike their parents, students do not appear to care much about school size and school sector. Middle school and high school students are less satisfied with their schools compared with elementary school students. Standardized test scores are also an important predictor of student satisfaction, but students’ own scores appear to matter more to their satisfaction than to their parents’ satisfaction. Although the coefficients for individual-level test scores were about a third the size of the coefficients for school-level averages in the parent models (Table 1), the own-score coefficients are about two thirds the size of the school-level average coefficients in the student models (Table 2).

Although we use multiple survey items to measure parent satisfaction, a single item is used to capture student satisfaction. Given that student satisfaction models have an ordered dependent variable, we also tried running ordered logit regressions (results are shown in Table A4 in the appendix). The results of the ordered logit models are substantively similar to the results from the linear regression models, with the gap measure producing a negative coefficient that is significant at the .10 level in the pooled model and at the .05 level when the sample is restricted to low-SES students.

Robustness

This study conducts several robustness tests to show that our baseline results are robust. We first check how sensitive our results are to some outliers in the performance gap measure. To see whether the outliers drive the results, we drop outliers that are 3 standard deviations (SDs) above the mean or 3 SDs below the mean and rerun the regression. Results are very similar to what we present here. We also try restricting the sample to when the gap measure has values within 2 SDs of the mean, and the main results remain the same (results not shown).

We also consider whether results are driven by the negative values of the performance gap measure. Theoretically, the gap measure is more likely to have positive values given that it represents the performance discrepancy between high-SES and low-SES students. Although the number of negative values is small compared with the positive values, there are some negative values in the gap measure (about 16.4% of the total observations), which indicates that low-SES students outperform high-SES students in some schools. Even when we drop those negative values, results remain similar (results not shown).

As another sensitivity test, we try treating parents’ educational attainment as a categorical variable because parent education is one of our key variables. We include dummies of father’s and mother’s education level as controls and rerun the analysis (results are not shown). In every estimation, there is a robust, negative relationship between performance gap and student satisfaction, in line with our argument.

Discussion and Conclusion

Social categories play a powerful role in shaping both how we perceive others and how others perceive us. Based on the notion of social aspirations (Cyert & March, 1963), which suggests that citizens’ evaluation of performance is fundamentally a relative process (see Olsen, 2017), this article investigates how comparing outcomes between different social groups may shape citizens’ perceptions of public service providers. Building upon the literature discussing social ingroups and outgroups, this research argues that individuals belonging to traditionally underserved social categories are more likely to look at outcome disparities between their group and the dominant group when evaluating service providers. Our empirical analysis using individual-level education data provides some evidence supporting the theoretical expectations—clients with low SES are less satisfied with service providers when there are high levels of outcome disparities while clients with high SES do not show such a pattern.

Incorporating a consideration of ingroup–outgroup perspectives and the possibility for heterogeneous responses among citizens to government performance is theoretically and practically important because both scholars and government officials have largely adopted aggregate measures of citizen satisfaction (mostly looking at the average response) as indicators of citizens’ assessments of government services. Although these topline average measures provide some information about citizen satisfaction, they can also mislead by disguising substantial (dis)satisfaction by certain subgroups of respondents (Ringquist, 2005). Put differently, there is no “typical” or “average” citizen in the real world, and such averaged-up reporting may provide an inaccurate picture of information to public officials and mislead them about what their constituents want. This article highlights that it is important to understand that individuals from different social groups are likely to hold different views of government performance, and they may provide accurate and reliable information to public administrators. Thus, citizens can help public administrators see various nuances that collectively tell a more complete story regarding performance, which may suggest a need to rethink how to set up service priorities and allocate public resources to better serve all citizens.

This article also highlights the multidimensional nature of performance by measuring performance outcomes in three different ways: (a) individual-level outcomes, (b) the average client outcome for an organization, and (c) an outcome gap between the high-SES and low-SES clients of an organization. The first two measures can be considered rough measures of service effectiveness (one at the individual level and one at the organizational level) while the third measure captures service equity. Across our models, a measure of overall performance (the average school exam score) is consistently the strongest predictor of parent and student satisfaction. This finding suggests that clients look first and foremost to the overall performance level of an organization when reporting their satisfaction with the organization. The second-strongest predictor of satisfaction is the client’s own outcome. For students, we find evidence that a third predictor—the performance gap between low-SES and high-SES students—also affects satisfaction levels. Specifically, low-SES students are less likely to offer favorable assessments of their school when there is a bigger academic achievement gap between low-SES and high-SES students, even after controlling for the student’s performance and the school’s overall average performance. For parents, we find no significant relationship between this gap and satisfaction levels.

Access to information may be an important mechanism underlying the pattern of results that we observe. Public discussion of school exam results in the media and other forums seems to focus mostly on overall average performance as this is the simplest single indicator by which to evaluate the quality of a school. To the extent that performance information filters down to how individual citizens evaluate schools, it is probably a simple overall summary statistic (like the average score or the percentage of students at a school who pass an exam) that has the highest chance of reaching the most citizens. This may help explain why overall performance dominates individual exam performance in our models.

The differences between parents and students may also reflect differences in information access. Although parents may primarily assess a school by relying on secondhand resources such as what they hear from other parents, the general reputation of the school, or media sources who have reacted to overall performance summary statistics, students have more firsthand experience to draw on—even if they are less sophisticated evaluators because of cognitive limitations associated with age. This could explain why our coefficient estimates are larger for students than for parents when it comes to individual and gap-based measures. Furthermore, students may be more acutely aware than their parents of class-based differences, particularly if parents are self-segregated such that they rarely interact with parents or students of a different SES. Students, on the contrary, regularly see such students at school even if they do not form strong social bonds with those outside their ingroup.

When considering how our results might generalize to other settings, several unique features of Korean schools are worth considering. Compared with other types of public organizations, schools are particularly data-rich environments from the perspective of clients. Families (with school-age children) interact with schools on a near-daily basis for much of the year, and standardized exam results from schools are often well-publicized, particularly in South Korea. Thus, high levels of transparency exist on at least one generally objective measure of school performance. Yet, official information about the equity of school outcomes is probably not well-publicized, suggesting that school students may be relying on anecdotal information about equity when forming opinions about a school. Clients of other public organizations may have less access to information about service quality, but we still expect that certain residents will care about equity to whatever extent they are able to assess it. For example, if people notice that roads or streetlights seem to be better maintained in certain areas of the city, they may conclude that the city government is inequitable in their provision of services. The media can also play a role in informing the public about service quality and equity. For instance, when people are frequently exposed to news coverage indicating that low-income individuals and racial minorities regularly experience relatively poor public service-related outcomes (e.g., higher infant mortality rates), the public may take notice and partially attribute blame for such disparities to government organizations.

These examples illustrate that the kind of heterogeneity of citizen evaluations that we find here is also relevant for other policy areas and other governments. In particular, many city and county government in the United States use citizen surveys to gauge citizen satisfaction with city/county services and to learn about concerns in the community. It is important to understand that—in addition to absolute outcomes—relative outcomes based on ingroup–outgroup comparisons may play an important role in shaping citizens’ evaluations of public services such as public transportation, trash collection, and community safety.

Theoretical and practical implications of this study are not limited to SES based on educational attainment. In theory, ingroup–outgroup comparisons can apply to any social group which experiences lower service performance than other groups. Given that our data came from a country that is relatively homogeneous in terms of race/ethnicity, this study focuses on SES as a potentially salient social identity. But future studies might consider how inequity in outcomes affects perceptions among ethnic minority groups. Depending on the service being considered, other identities such as veteran status, gender, sexual orientation, or citizenship status might assume particular salience.

When it comes to practice, our study suggests that public administrators should take care to consider how opinions might differ among social groups when assessing citizen satisfaction with their services. One practical recommendation is that when administrators examine results from a satisfaction survey, they should consider breaking down responses by subgroups to determine whether sentiments differ among various social groups. For example, suppose that there is an organization which enjoys fairly high average levels of satisfaction among the full population, but for which there is discontent among a minority group who feels they are being underserved. Unless responses to the survey results are summarized by subgroup, administrators will be unaware of the potential equity issue being flagged by their respondents. Designers of citizen or client surveys might also consider including survey items that directly ask about perceptions of equity to more clearly assess the extent to which residents feel they are being treated equally. Carefully examining survey results can then help to inform practitioners about potential problem areas within the organization that deserve further investigation and—perhaps—corrective actions.

Footnotes

Appendix

Table A4.

Performance Gap and Student Satisfaction: Ordered Logit Analysis.

Independent variables	Pooled model	SES
Independent variables	Pooled model	High SES	Low SES
Performance gap	–0.075 ^† (0.044)	–0.046 (0.052)	–0.132* (0.055)
Household income	−0.018 (0.016)	0.003 (0.022)	−0.039^† (0.021)
Parents education	0.019 (0.012)	0.034 (0.023)	−0.007 (0.031)
Average teacher education	0.006** (0.002)	0.004^† (0.002)	0.008** (0.002)
Average teacher experience	−0.011* (0.006)	−0.019** (0.007)	−0.003 (0.006)
Full-time teachers	−0.000 (0.004)	−0.001 (0.005)	0.002 (0.004)
Class size	−0.041*** (0.009)	−0.031** (0.011)	−0.054*** (0.011)
Low-performance students	−0.011^† (0.006)	−0.019* (0.008)	−0.007 (0.007)
Number of students	−0.047 (0.079)	−0.201* (0.098)	0.128 (0.088)
Public school	−0.008 (0.066)	0.017 (0.080)	−0.031 (0.073)
Middle school	−1.620*** (0.097)	−1.493*** (0.119)	−1.753*** (0.102)
General high school	−2.261*** (0.122)	−2.037*** (0.155)	−2.500*** (0.136)
Vocational high school	−1.705*** (0.180)	−1.650*** (0.233)	−1.919*** (0.200)
Student test scores	0.216*** (0.013)	0.214*** (0.017)	0.208*** (0.020)
School mean scores	0.344*** (0.053)	0.330*** (0.059)	0.328*** (0.072)
Wald χ²	1,815.11	1,305.06	1,327.33
Prob. > χ²	.000	.000	.000
Log pseudolikelihood	−70,763.89	−38,914.63	−31,784.52
N	56,743	31,246	25,497

†

p < .10. *p < .05. **p < .01. ***p < .001. (Two-tailed test)

Authors’ Note

An earlier version of this article was presented at the Public Management Research Conference in Washington, D.C., June 8 to 10, 2017. We thank three anonymous reviewers for their constructive comments and suggestions.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Miyeon Song

Notes

Author Biographies

Miyeon Song is an assistant professor of Public Administration and Policy in the Department of Political Science at the University of South Carolina. Her research interests include public management, government performance, public-private comparison, organizational theory and behavior, and bureaucratic politics.

Minjung Kim is a PhD candidate in the School of Public Affairs, American University. Her research interests include public and nonprofit management, participatory governance, performance management, race and ethnicity, and organizational behavior.

Nathan Favero is an assistant professor in the School of Public Affairs at American University. His research interests include citizen satisfaction, equity, internal management, quantitative methodology, and education policy.

References

Allport

G. W.

(1954). The nature of prejudice: Abridged. Garden City, NY: Doubleday.

Andersen

L. B.

Boesen

Pedersen

L. H.

(2016). Performance in public organizations: Clarifying the conceptual space. Public Administration Review, 76, 852-862.

Andersen

L. B.

Heinesen

Pedersen

L. H.

(2016). Individual performance: From common source bias to institutionalized assessment. Journal of Public Administration Research and Theory, 26, 63-78.

Andersen

S. C.

Hjortskov

(2015). Cognitive biases in performance evaluations. Journal of Public Administration Research and Theory, 26, 647-662.

Andrews

Boyne

G. A.

Walker

R. M.

(2006). Subjective and objective measures of organizational performance: An empirical exploration. In Boyne

G. A.

Meier

K. J.

O’Toole

L. J.

Walker

R. M.

(Eds.), Public service performance: Perspectives on measurement and management (pp. 14-34). Cambridge, UK: Cambridge University of Press.

Baekgaard

Serritzlew

(2016). Interpreting performance information: Motivated reasoning or unbiased comprehension. Public Administration Review, 76, 73-82.

Barrows

Henderson

Peterson

P. E.

West

M. R.

(2016). Relative performance information and perceptions of public service quality: Evidence from American school districts. Journal of Public Administration Research and Theory, 26, 571-583.

Beck

D. E.

Maranto

W.-J.

(2014). Determinants of student and parent satisfaction at a cyber charter school. The Journal of Educational Research, 107, 209-216.

Billingham

C. M.

Hunt

M. O.

(2016). School racial composition and parental choice: New evidence on the preferences of White parents in the united states. Sociology of Education, 89, 99-117.

10.

Bohmert

M. N.

DeMaris

(2015). Interracial friendship and the trajectory of prominority attitudes: Assessing intergroup contact theory. Group Processes & Intergroup Relations, 18, 225-240.

11.

Boyne

G. A.

(2002). Concepts and indicators of local authority performance: An evaluation of the statutory frameworks in England and Wales. Public Money & Management, 22(2), 17-24.

12.

Boyne

G. A.

(2003). What is public service improvement? Public Administration, 81, 211-227.

13.

Brown

Benedict

W. R.

(2002). Perceptions of the police: Past findings, methodological issues, conceptual issues and policy implications. Policing: An International Journal of Police Strategies & Management, 25, 543-580.

14.

Brown

Coulter

P. B.

(1983). Subjective and objective measures of police service delivery. Public Administration Review, 43, 50-58.

15.

Brudney

J. L.

England

R. E.

(1982). Urban policy making and subjective service evaluations: Are they compatible? Public Administration Review, 42, 127-135.

16.

Buckley

Schneider

(2009). Charter schools: Hope or hype? Princeton, NJ: Princeton University Press.

17.

Burgess

Wilson

Lupton

(2005). Parallel lives? Ethnic segregation in schools and neighbourhoods. Urban Studies, 42, 1027-1056.

18.

Charbonneau

É.

Van Ryzin

G. G

. (2012). Performance measures and parental satisfaction with New York City schools. The American Review of Public Administration, 42, 54-65.

19.

Cheon

Song

McCrea

A. M.

Meier

K. J.

(2019). Health care in America: The relationship between subjective and objective assessments of hospitals. International Public Management Journal. Advance online publication. doi:10.1080/10967494.2018.1553808

20.

Chung

J.-E.

Jin

(2011). In-group preference as opportunism governance in a collectivist culture: Evidence from Korean retail buyer-supplier relationships. Journal of Business & Industrial Marketing, 26, 237-249.

21.

Cyert

R. M.

March

J. G

. (1963). A behavioral theory of the firm. Englewood Cliffs, NJ: Prentice Hall.

22.

Duncan

O. D.

Featherman

D. L.

Duncan

(1972). Socioeconomic background and achievement. New York, NY: Seminar Press.

23.

Favero

Meier

K. J.

(2013). Evaluating urban public schools: Parents, teachers, and state assessments. Public Administration Review, 73, 401-412.

24.

Frederickson

H. G.

(1990). Public administration and social equity. Public Administration Review, 50, 228-237.

25.

Glazerman

Dotter

(2017). Market signals: Evidence on the determinants and consequences of school choice from a citywide lottery. Educational Evaluation and Policy Analysis, 39, 593-619.

26.

Hjortskov

(2016). Citizen satisfaction: Political voice and cognitive biases (Doctoral thesis). Forlaget Politica, Aarhus, Denmark.

27.

Hjortskov

(2019). Citizen expectations and satisfaction over time: Findings from a large sample panel survey of public school parents in Denmark. The American Review of Public Administration, 49, 353-371. doi:10.1177/0275074018765822

28.

Jacobsen

Snyder

J. W.

Saultz

(2015). Understanding satisfaction with schools: The role of expectations. Journal of Public Administration Research and Theory, 25, 831-848.

29.

James

(2007). Evaluating the expectations disconfirmation and expectations anchoring approaches to citizen satisfaction with local public services. Journal of Public Administration Research and Theory, 19, 107-123.

30.

James

(2011). Managing citizens’ expectations of public service performance: Evidence from observation and experimentation in local government. Public Administration, 89, 1419-1435.

31.

James

Moseley

(2014). Does performance information about public services affect citizens’ perceptions, satisfaction, and voice behaviour? Field experiments with absolute and relative performance information. Public Administration, 92, 493-511.

32.

James

Van Ryzin

G. G.

(2017). Incredibly good performance: An experimental study of source and level effects on the credibility of government. The American Review of Public Administration, 47, 23-35.

33.

Jha

A. K.

Orav

E. J.

Zheng

Epstein

A. M.

(2008). Patients’ perception of hospital care in the United States. New England Journal of Medicine, 359, 1921-1931.

34.

Jilke

Baekgaard

(2019). The political psychology of citizen satisfaction: Does functional responsibility matter? Journal of Public Administration Research and Theory. Advance online publication. doi:10.1093/jopart/muz012

35.

Jimenez

B. S.

(2014). Externalities in the fragmented metropolis: Local institutional choices and the efficiency-equity trade-off. The American Review of Public Administration, 46, 314-3–36.

36.

Kelly

J. M.

(2003). Citizen satisfaction and administrative performance measures is there really a link? Urban Affairs Review, 38, 855-866.

37.

Kelly

J. M.

Swindell

(2002). A multiple-indicator approach to municipal service evaluation: Correlating performance measurement and citizen satisfaction across jurisdictions. Public Administration Review, 62, 610-621.

38.

Kelly

J. M.

Swindell

(2003). The case for the inexperienced user: Rethinking filter questions in citizen satisfaction surveys. The American Review of Public Administration, 33, 91-108.

39.

Korean Statistical Information Service. (2017). Available from http://kosis.kr/

40.

Lee

Brinton

M. C.

(1996). Elite education and social capital: The case of South Korea. Sociology of Education, 69, 177-192.

41.

Lee

Y. T.

(1993). Ingroup preference and homogeneity among African American and Chinese American students. The Journal of Social Psychology, 133, 225-235.

42.

Lewis

J. H.

Hamilton

D. K.

(2011). Race and regionalism: The structure of local government and racial disparity. Urban Affairs Review, 47, 349-384.

43.

Lyons

W. E.

Lowery

DeHoog

R. H.

(1992). The politics of dissatisfaction: Citizens, services, and urban institutions. Armonk, NY: M.E. Sharpe.

44.

Marvel

J. D.

(2015). Unconscious bias in citizens evaluations of public sector performance. Journal of Public Administration Research and Theory, 26, 143-158.

45.

Meier

K. J.

Wrinkle

R. D.

Polinard

J. L.

(1999). Equity versus excellence in organizations: A substantively weighted least squares analysis. The American Review of Public Administration, 29, 5-18.

46.

Mintrom

Walley

(2013). Education governance in comparative perspective. In Manna

McGuinn

(Eds.), Education governance for the twenty-first century: Overcoming the structural barriers to school reform (pp. 252-274). Washington, DC: The Brookings Institution.

47.

Morgeson

F. V.

(2013). Expectations, disconfirmation, and citizen satisfaction with the US federal government: Testing and expanding the model. Journal of Public Administration Research and Theory, 23, 289-305.

48.

Olsen

A. L.

(2015). Citizen(dis) satisfaction: An experimental equivalence framing study. Public Administration Review, 75, 469-478.

49.

Olsen

A. L.

(2017). Compared to what? How social and historical reference points affect citizens’ performance evaluations. Journal of Public Administration Research and Theory, 27, 562-580.

50.

Osborne

Gaebler

(1992). Reinventing government: How the entrepreneurial spirit is transforming the public sector. Reading, MA: Addison-Wesley.

51.

O’Toole

L. J.

Jr. Meier

K. J.

(2014). Public management, context, and performance: In quest of a more general theory. Journal of Public Administration Research and Theory, 25, 237-256.

52.

Park

B. K.

(1988). The state, class and educational policy: A case study of South Korea’s High School Equalization Policy (Doctoral dissertation). University of Wisconsin–Madison.

53.

Petrovsky

Mok

J. Y.

León-Cázares

(2017). Citizen expectations and satisfaction in a young democracy: A test of the expectancy-disconfirmation model. Public Administration Review, 77, 395-407.

54.

Poister

T. H.

Thomas

J. C.

(2011). The effect of expectations and expectancy confirmation/disconfirmation on motorists’ satisfaction with state highways. Journal of Public Administration Research and Theory, 21, 601-617.

55.

Rainey

H. G.

(2009). Understanding and managing public organizations. San Francisco, CA: John Wiley.

56.

Ringquist

E. J.

(2005). Assessing evidence of environmental inequities: A meta-analysis. Journal of Policy Analysis and Management, 24, 223-247.

57.

Robinson

(1994). Social status and academic success in South Korea. Comparative Education Review, 38, 506-530.

58.

Roch

C. H.

Poister

T. H.

(2006). Citizens, accountability, and service satisfaction: The influence of expectations. Urban Affairs Review, 41, 292-308.

59.

Rossi

P. H.

(1972). Community social indicators. In Campbell

Converse

P. E.

(Eds.), The human meaning of social change (pp. 87-126). New York, NY: Russell Sage Foundation.

60.

Salamon

L. M.

(1987). Of market failure, voluntary failure, and third-party government: Toward a theory of government-nonprofit relations in the modern welfare state. Journal of Voluntary Action Research, 16, 29-49.

61.

Saporito

(2003). Private choices, public consequences: Magnet school choice and segregation by race and poverty. Social Problems, 50, 181-203.

62.

Schachter

H. L.

(2010). Objective and subjective performance measures: A note on terminology. Administration & Society, 42, 550-567.

63.

Seth

M. J.

(2002). Education fever: Society, politics, and the pursuit of schooling in South Korea. Honolulu: University of Hawaii Press.

64.

Sigelman

Bledsoe

Welch

Combs

M. W.

(1996). Making contact? Black-White social interaction in an urban setting. American Journal of Sociology, 101, 1306-1332.

65.

Sirin

S. R.

(2005). Socioeconomic status and academic achievement: A meta-analytic review of research. Review of Educational Research, 75, 417-453.

66.

Song

Meier

K. J.

(2018). Citizen satisfaction and the kaleidoscope of government performance: How multiple stakeholders see government performance. Journal of Public Administration Research and Theory, 28, 489-505.

67.

Stipak

(1979). Citizen satisfaction with urban services: Potential misuse as a performance indicator. Public Administration Review, 39, 46-52.

68.

Stipak

(1980). Local governments’ use of citizen surveys. Public Administration Review, 40, 521-525.

69.

Tajfel

(1986). The social identity theory of intergroup behavior. Psychology of Intergroup Relations, 5, 7-24.

70.

Valant

Newark

D. A.

(2016). The politics of achievement gaps: U.S. public opinion on race-based and wealth-based differences in test scores. Educational Researcher, 45, 331-346.

71.

Van Ryzin

G. G

. (2004). Expectations, performance, and citizen satisfaction with urban services. Journal of Policy Analysis and Management, 23, 433-448.

72.

Van Ryzin

G. G

. (2006). Testing the expectancy disconfirmation model of citizen satisfaction with local government. Journal of Public Administration Research and Theory, 16, 599-611.

73.

Van Ryzin

G. G

. (2013). An experimental test of the expectancy-disconfirmation theory of citizen satisfaction. Journal of Policy Analysis and Management, 32, 597-614.

74.

Van Ryzin

G. G.

Immerwahr

Altman

. (2008). Measuring street cleanliness: A comparison of New York City’s scorecard and results from a citizen survey. Public Administration Review, 68, 295-303.

75.

Van Ryzin

G. G.

Muzzio

Immerwahr

Gulick

Martinez

. (2004). Drivers and consequences of citizen satisfaction: An application of the American customer satisfaction index model to New York City. Public Administration Review, 64, 331-341.

76.

Verba

(1996). The citizen as respondent: Sample surveys and American democracy presidential address, American Political Science Association, 1995. American Political Science Review, 90(1), 1-7.

77.

Walker

R. M.

Boyne

G. A.

Brewer

G. A.

(2010). Public management and performance: Research directions. Cambridge, UK: Cambridge University Press.

78.

Weiher

G. R.

Tedin

K. L.

(2002). Does choice lead to racially distinctive schools? Charter schools and household preferences. Journal of Policy Analysis and Management, 21, 79-92.

79.

Wilson

J. Q.

(1989). Bureaucracy: What government agencies do and why they do it. New York, NY: Basic Books.