You are on page 1of 9

Research in Social Stratification and Mobility 71 (2021) 100565

Contents lists available at ScienceDirect

Research in Social Stratification and Mobility


journal homepage: www.elsevier.com/locate/rssm

Does school matter for students’ self-esteem? Associations of family SES,


peer SES, and school resources with Chinese students’ self-esteem
Weiwei Li a, Xihua Zeng b, e, *, Yun Wang c, *, Robin Curtis d, Erin Sparks d
a
Guangdong University of Education, Guangzhou, Guangdong, People’s Republic of China
b
Department of Psychology, School of Public Health, Southern Medical University (Guangdong Provincial Key Laboratory of Tropical Disease Research), Guangzhou,
Guangdong, People’s Republic of China
c
State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University, Beijing, People’s Republic of China
d
Department of Psychology and Neuroscience, Dalhousie University, Halifax, Canada
e
Department of Psychiatry, Zhujiang Hospital, Southern Medical Uniersity, Guangzhoudong, People’s Republic of China

A R T I C L E I N F O A B S T R A C T

Keywords: Educational equality has attracted many countries’ efforts to increase school resources to improve poor students’
Socioeconomic status nonachievement development. The present study investigated the associations between family socioeconomic
Self-esteem status (SES) and self-esteem, as well as their variances across school contextual variables (school location, peer
Peer SES
SES, school resources). This study used data from the National Children’s Study of China (NCSC), with a sample
School resources
of 23,627 students (4th – 9th grade, mean age = 12.76 years, 47.1 % female) from 604 schools. The multilevel
Multilevel analysis
linear model analyses mainly found the following: (1) family SES (objective income, subjective income, parental
education) was significantly and positively related to self-esteem; (2) peer parental education and teacher-related
resources were positively and significantly associated with self-esteem; and (3) most importantly, the positive
association between subjective income and self-esteem was stronger in urban schools than rural schools. The
positive association between parental education and self-esteem was stronger in schools with higher peer
parental education but was weaker in schools with higher teacher education. Based on these findings, we provide
suggestions for the efficient allocations of school resources and the promotion of programs for students’ self-
esteem.

1. Introduction influences the development of important life outcomes (Orth et al.,


2012). Self-esteem attained in childhood and adolescence has
Educational equality is an important issue in different countries. To far-reaching consequences on students’ development, especially for
realize educational equality, many countries attempt to increase poor students) (Mann, Hosman, Schaalma, & De Vries, 2004). For
educational resources to buffer the negative impact of family SES (so­ example, positive self-esteem predicts increased academic resilience
cioeconomic status) on poor children’s development, which raises an among poor students in disadvantaged situations (Borman & Overman,
increasing concern about the efficiency of school resource investment. 2004) and increased success in their future careers (Mann, Hosman,
The empirical studies on educational resource investment have focused Schaalma, & De Vries, 2004). Thus, self-esteem is a critical internal
intensively on the functions of peer family SES (average students’ family resource for students with low family SES to improve their disadvan­
SES from a school; van Ewijk & Sleegers, 2010) and school resources (all taged conditions. Unfortunately, most students with low family SES
the school material inputs; OECD, 2013), revealing their important roles have impaired self-esteem (e.g., Falci, 2011; Liu & Zhao, 2016; Twenge
in students’ achievement (e.g., Caldas & Bankston, 1997; Opdenakker & & Campbell, 2002; Veselska et al., 2010), which puts them at a disad­
Damme, 2007). However, scholars are increasingly concerned about the vantage and ultimately hinders intergenerational social mobility. A
association between school resources and students’ nonachievement limited Western study has found that as the proportions of middle- or
outcomes such as self-esteem. It has been documented that self-esteem high-income parents increased, low-income students experienced more

* Corresponding authors.
E-mail addresses: lweiwei01@gdei.edu.cn (W. Li), zxihuar@163.com (X. Zeng), wangyun@bnu.edu.cn (Y. Wang), Robin.Curtis@Dal.Ca (R. Curtis), Erin.Sparks@
Dal.Ca (E. Sparks).

https://doi.org/10.1016/j.rssm.2020.100565
Received 29 April 2018; Received in revised form 28 August 2020; Accepted 21 October 2020
Available online 24 October 2020
0276-5624/© 2020 Elsevier Ltd. All rights reserved.
W. Li et al. Research in Social Stratification and Mobility 71 (2021) 100565

psychosocial problems such as low self-esteem (Crosnoe, 2009). In­ high social status, which also affects others’ evaluations of their children
dividuals’ self-esteem is more contingent on social contexts in Eastern because of the Chinese belief that “a tiger has no dog son”. As the re­
cultures, such as China (Hwang, 2012; Pelham, 1995). The impact of flected appraisals model assumes, children can perceive high social
family SES on self-esteem could vary across schools with different peer status when others positively evaluate them because of their parents’
SES and school resources. To take advantage of school resources to high level of education, which leads to high self-esteem. Moreover,
improve students’ self-esteem, it is necessary to examine whether and well-educated parents’ have better knowledge and parenting practices
how different school resources affect students’ self-esteem and how they and can support and protect children’s self-esteem.
buffer the association of family SES to self-esteem. Second, compared with objective SES, subjective SES has a stronger
The current study is based on data from the National Children’s connection to the perceptions of the social class (Chen et al., 2018, 2019;
Study of China (NCSC). It aims to test (1) the associations between in­ Operario et al., 2004; Piko & Fitzpatrick, 2001). According to reference
dividual students’ family SES (objective income, parental education, group theory, subjective SES makes individuals perceive their social
subjective income) and self-esteem; (2) the associations between school positions by comparing their wealth with that of the referent group in
location, peer family SES, school resources (material resources, teacher- the immediate social context (Chen, 2019). Irrespective of their objec­
related resources) and students’ self-esteem; and (3) whether the asso­ tive SES, people feel richer or poorer than the referent group, which
ciations between individual family SES (by means of different SES in­ leads to positive or negative self-evaluations. For example, recent
dicators) and self-esteem vary across school contextual variables (school empirical evidence has found that students with a lower subjective SES
location, peer family SES, school resources). This study aims to reveal (taking over from parents’ subjective SES) were more vulnerable to low
not only the impacts of school contextual variables on students’ self- self-esteem (Zeng et al., 2017).
esteem but also their moderating roles in the associations between To summarize, the association between family SES and self-esteem is
family SES and self-esteem. This study can provide more informed comprehensive because of the differentiation between the objective and
suggestions, not only for China but also for other countries, about the subjective indicators, as well as the differentiation among objective SES
efficient allocation of school resources. indicators. However, there is a shortage of evidence that includes all the
indicators in one study. To obtain a comprehensive understanding of the
2. Theoretical framework impact of family economic context on students’ self-esteem, this study
seeks to differentiate between objective SES (objective income, parental
2.1. The association between family SES and students’ self-esteem education) and subjective income to test the links between family SES
and self-esteem at the individual and school levels.
Self-esteem is defined as an individual’s overall positive or negative
evaluation of the self, emerging from interactions with one’s peers, 2.2. The associations between school contexts and students’ self-esteem
family, and other members of society (Rosenberg et al., 1995). Re­
searchers have documented that family SES is an important family According to self-categorization theory, peer family SES and school
contextual factor relating to self-esteem (Twenge & Campbell, 2002; resources could make a difference in self-esteem. Self-categorization
Wiltfang and Scarbecz, 1990). Studies conducted in Western countries theory assumes that categorizing oneself as a part of a certain group can
and mainland China have provided evidence that higher family SES (as accentuate the perceived similarities (e.g., identity, attitudes) between
ascribed by parents) positively predicted children’s and adolescents’ the self and the in-group members and accentuate the perceived dif­
higher self-esteem (Falci, 2011; Feng & Zhang, 2008; Herman et al., ferences between the self and the out-group members (Stets & Burke,
2011; Zeng et al., 2017). 2000; van Dick et al., 2005v). Usually, schools with high peer SES and
In the empirical studies to date, researchers have tested family SES with rich school resources are known for their superior students and
through different indicators, namely, objective family SES and subjec­ better teaching resources (Chiu, 2010). In the schools with high peer
tive family SES (Kraus et al., 2012). Specifically, objective family SES family SES or rich school resources, students can categorize themselves
consists of self-reports of an individual’s family or their income, edu­ as a member of the group of privileged people and enjoy a sense of group
cation, and occupation (Gottfried, 1985). Subjective family SES is glory and pride derived from the distinguished group members, thereby
assessed by an individual’s psychological identification as a member of gaining high self-esteem (e.g., Major & O’Brien, 2005; Marsh et al.,
their class after comparing their resources with the resources of others, 2000; van Dick et al., 2005v). Speaking to this possibility, students in the
or with their past income (e.g., Adler et al., 2000). schools with high peer family SES reported significantly higher
Objective and subjective SES independently link to self-esteem in self-esteem than did children in the schools with low peer family SES
different ways (Zeng et al., 2017). First, objective family SES can relate (Farrell et al., 2009). In another Chinese study, Zheng (2013) compared
to self-esteem through internalizing others’ perceptions about ourselves, students from advantaged schools with those from disadvantaged
referred to as the ‘looking-glass self’ process (i.e., the reflected ap­ schools (based on school types, not contextual variables), finding that
praisals model, Twenge & Campbell, 2002). If we perceive that others students in the advantaged schools experienced higher self-esteem.
see us as lower/higher class by our visible wealth, we are likely to see With the growing public awareness of educational investment in
ourselves that way as well and therefore experience lower/higher schooling and teaching, most studies have tested the associations be­
self-esteem. In the empirical studies, objective family SES has been tween school resources and students’ achievement (e.g., Caldas &
tested with different objective indicators, such as objective income, Bankston, 1997; Opdenakker & Damme, 2007), yet very few studies
occupation, and education (e.g., Conger et al., 2010). Objective income focus on students’ self-esteem. Based on the self-categorization process,
is the most visible label for self-evaluation. For example, students easily people share in the high value and pride that is derived from the
experience lower/higher self-esteem among peers when others evaluate distinguished groups (e.g., Marsh et al., 2000). We expect that high peer
their worn-out/fancy clothes. Parental occupation, a measure of family SES and rich school resources could be a link to students’ high
parental working position in the society, is closely connected with par­ self-esteem, which needs to be tested by empirical study. Moreover, as
ents’ working conditions (Wang et al., 2018). It indirectly influences mentioned above, family SES can be linked to self-esteem by means of
students’ outcomes by relating to someone who is close to the child different SES indicators. It remains unknown whether and how these
(Bronfenbrenner, 1979). Parental education can make differences in the school contextual variables weaken or strengthen the relationships be­
perception of social class in some societies, such as China. The Chinese tween family SES (with different indicators) and self-esteem.
people highly stress the importance of education, and scholars have
enjoyed high social reputation and status since ancient China (Han,
2013). Within such societies, well-educated parents themselves enjoy

2
W. Li et al. Research in Social Stratification and Mobility 71 (2021) 100565

2.3. School contexts as moderators for the associations between family family SES and self-esteem, depending on the types of school contextual
SES and self-esteem variables and family SES indicators in question. Moreover, because
family SES may be linked to self-esteem through different family in­
Many studies have attempted to test the variations in the associations dicators, it is necessary to explore how the school contextual variables
between family SES and students’ outcomes across schools. Researchers interact with different family SES indicators to relate to self-esteem.
have proposed three hypotheses to explain this moderation effect,
namely, the social reproduction hypothesis, the complementary intangible 2.4. The association between school location and students’ self-esteem
hypothesis, and the public resources substitution hypothesis (Chiu & Zeng,
2008, for a review). Although these hypotheses are proposed based on Chinese rapid economic development and urbanization has unsur­
studies about achievement (Greenwald, Hedges, & Laine, 1996; Harris, prisingly widened the economic disparity between urban areas and rural
2010; Rumberger & Palardy, 2005; Ryabov, 2011) and subjective areas (e.g., Treiman, 2012). School location (urban schools versus rural
well-being (Diener & Biswas-Diener, 2002), they can be used to elabo­ schools) represents the urban-rural disparity to a larger extent and may
rate on the moderation effect of school context on the family relate to students’ self-esteem.
SES/self-esteem link because self-esteem is one of the important aspects Chinese urban areas enjoy better welfare resources, public goods
for the development of children and adolescents, closely relating to resources, and educational resources than rural areas (Treiman, 2012).
achievement, well-being, etc. Compared with rural schools, urban schools spend far more on educa­
First, the social reproduction hypothesis assumes that the effect of tion and have better health care. In line with the self-categorization
family SES on students’ self-esteem may not vary across school resources process, students in urban schools could be more likely to categorize
(Blossfeld & Shavit, 1991). As mentioned above, students can form a themselves as members of a privileged group, leading to higher
self-evaluation by observing the way that others evaluate by their visible self-esteem.
wealth (the reflected appraisals model). No matter how peer family SES The associations between family SES and self-esteem may differ be­
and school resources vary, richer students in one school can still show off tween urban versus rural schools. On the one hand, urban schools can
their visible family wealth (e.g., fancy clothes and learning materials) provide rich educational and public resources, which can buffer the
and obtain higher self-esteem. For example, it has been documented that impact of family SES on self-esteem (as the public resources substitution
school material resources in different institutions outside of the family hypothesis assumed). On the other hand, the educational market has
(e.g., countries) did not moderate the effect of individual family SES on grown faster in Chinese urban areas. Urban parents have to spend much
outcomes (Chiu & Zeng, 2008). Thus, the social reproduction hypothesis more than rural parents on children’s educational support, such as skill
indicates that family SES (especially objective family income) could be training and extracurricular learning; thus, family SES is more important
constant across schools. for students’ development in rich areas (Chiu & Zeng, 2008) (as the
Second, the complementary intangible hypothesis presumes that complementary intangible hypothesis assumed). At present, many par­
school environments may strengthen the association between family SES ents have attempted to seek urban schooling for children; it is necessary
and students’ self-esteem. Many studies have found the positive impact to explore whether family SES is more strongly related to self-esteem in
of peer family SES on students’ outcomes (e.g., Perry & McConney, urban schools.
2010); in particular, when students with high levels of parental educa­ To summarize, due to economic disparity and educational inequity,
tion aggregate together, schools are more likely to provide advanced it is important to examine whether and how school contexts could
teaching and a complicated curriculum. It has been documented that as change the link between family SES and self-esteem, which can help us
peer SES increases, schools are more likely to provide small groups improve the efficiency of educational investment and increase educa­
working on a scientific experiment, hands-on practical activities, etc. tional equality. As seen in Fig. 1, we refer to reflected appraisal theory,
(Tao et al., 2013). If students can receive the advanced knowledge and referent group theory, and self-categorization theory to assume that
academic support of well-educated parents, they can easily fulfill the family SES and school contextual variables positively relate to self-
schooling requirement and build up higher self-esteem. Thus, this hy­ esteem. This is the first study to use the hierarchical linear model
pothesis reflects the growing importance of family SES (especially (HLM) of multilevel analysis to test cross-level interactions between
family intangible resources such as parental education) on self-esteem in family SES and school contextual variables on self-esteem. Based on the
schools with a high peer SES. three hypotheses of the school context effect, we predict that the cross-
Third, the public resources substitution hypothesis (Blossfeld & level interactions could be neutrally, positively, or negatively related to
Shavit, 1991) assumes that schools with better teacher-related resources self-esteem, depending on the types of school resources and family SES
may buffer the impact of family SES on student’s self-esteem. For years, indicators.
the question about what kinds of school resources can positively relate
to students’ outcomes has not been confidently settled (e.g., Hanushek, 3. Data and methods
1997), but researchers have thought highly of teacher-related resources
(e.g., Darling-hammond, 2000). Teacher-related resources, such as 3.1. Data
teacher education level and teacher/student ratio, describe the varia­
tions in the instructional support across classrooms (Hanushek, 1997). Data for this study came from the National Children’s Study of China
Highly educated teachers can provide effective teaching tactics and form (NCSC). NCSC is a nationally representative standardized assessment
intimate relationships with students, which would reduce the impor­ developed and conducted by the National Key Laboratory of Cognitive
tance of parental tutoring and instruction (Rjosk et al., 2014) and buffer Neuroscience and Learning at Beijing Normal University. The objective
the parents’ negative evaluations and parenting practices associated of the NCSC is to assess the academic, cognitive, and psychosocial
with low family SES (Li et al., 2008; Liu et al., 2015). Compared with the development of children and adolescents aged 6–15 years in China
null or strengthened moderation effect assumed, the public resources (Dong & Lin, 2011). The NCSC sample was selected from 100 counties in
substitution hypothesis may provide practical advice on what kind of 31 provinces of the Chinese mainland using multistage, stratified, and
efforts a school can make, particularly for students with low family SES. unequal probability methods. More details about the design of the NCSC
To summarize, the three hypotheses and related evidence provide a can be found on its website (http://cddata-china.org/main/index.aspx)
different perspective to explain the moderation effects of the school or in the general report of NCSC (Dong & Lin, 2011). The current study
environment. Building on the three hypotheses about the effect of the used the social development database from the NCSC and the sample
school environment, we predict that school contextual variables may comprised 23,627 students (11,868 students from 229 primary schools
neutrally, positively, or negatively moderate the associations been and 11,759 from 375 middle schools), covering 4th to 9th grade

3
W. Li et al. Research in Social Stratification and Mobility 71 (2021) 100565

Fig. 1. Hypotheses for the associations between family SES, school contexts, their interactions, and self-esteem. a. We predicted that school-level variables interacted
with an individual-level variable to relate to self-esteem, which includes three possible hypotheses. b = zero cross-level interaction on the association between the
family SES and self-esteem. c = the positive cross-level interaction on the association between family SES and self-esteem. d = the negative cross-level interaction on
the association between family SES and self-esteem. The arrows with dashed lines indicated the cross-level interactions between individual and school-level variables.

students. The mean age of participants was 12.76 years [standard de­
viation (SD) = .17], and 47.1 % (N = 11,118) of participants were fe­ Table 1
Measurements of variables.
male. Consent was obtained from the participants’ parents and school
principals. Self-esteem Overall evaluation of self.

Individual-level
variables
3.2. Measures Objective income Annual income per family over the past year.
Parental education Years of schooling that father or mother has received,
Table 1 shows the details of the measures used in the current study, including 6 years (i.e., primary schools), 9 years (i.e.,
which are arranged into individual-level and school-level variables. middle schools), 12 years (i.e., high schools), above 15
years (i.e., college and university). The longer years
indicate a higher education level.
3.2.1. Student-related variables Subjective income Parents’ perception of their family income in comparison
Self-esteem was measured by the nine-item Chinese version of the with the average level of their local community.
Rosenberg self-esteem scale (Rosenberg & Pearlin, 1978). According to Control variables
Gender Female coded as 0, male coded as 1.
the scale, it measured global self-esteem, which is defined as an overall
Age Students’ biological age.
attitude about the self with an evaluation component, such as ‘I take a Behind age Normal schooling age coded as 0, older than the most
positive attitude towards myself; I am able to do things as well as most frequent or usual age for the grade coded as 1.
other people’. Responses were rated on a four-point scale, ranging from School-level variables
1 (strongly disagree) to 4 (strongly agree). With negatively worded items Peer family income Averaged individual objective income from a school.
Peer parental Averaged individual parental education from a school.
reversed, all items were averaged to form the scale score. Cronbach’s
education
alpha was .85 for this scale in the present study. Peer subjective Averaged individual subjective income from a school.
Socioeconomic status (SES) was measured by three indicators: objec­ family income
tive income, parental education, and subjective income. These variables School location Urban coded as 1, rural coded as 0.
were all reported by parents. Objective income was reported by the exact Book collections The book collection that schools have (‘how many books
are there in your school library?’), with response options
combined income of all family members per year and converted to an including 1 (less than 500), 2 (500− 2000), 3
ordinal scale. The scale ranged from 1 (less than 3000 RMB, approxi­ (2001− 5000) and 4 (more than 5000). Higher scores
mately equal to 453 U.S. dollars) to 9 (200,001 RMB or more, approx­ indicate more school resources.
imately equal to 30,199 U.S. dollars). Parental education was measured Equipment The equipment that schools have (‘how many classes (the
approximate rate) are equipped with multimedia,
by the years of schooling that parents have received. Responses included
projection, etc.?’), responded by 1 (0− 20%), 2 (21 %–40
6 years (i.e., primary school), 9 years (i.e., middle school), 12 years (i.e., %), 3 (51 %–60 %), 4 (61 %–80 %), and 5 (80 %–100 %).
high school), and 15 years or more (i.e., college and university), with the Higher scores indicate more equipment applied in
longer years indicating higher education levels. Subjective income was schools.
reported according to the parents’ perception of their family income in Teacher education A number computed from the percentage of teachers with
a bachelor’s, master’s or Ph.D. degree in a given school.
comparison with the average income level of their local community. Teacher/student ratio The number of teachers divided by the number of students
Responses were measured on a 5-point scale ranging from 1 (far worse in a given school.
than average) to 5 (far better than average).
Control variables included students’ demographic variables,
including gender, age, and below normal schooling age. We used the age
and grade variable to create the ‘normal schooling age group’ variable. If
students were as old as the normal schooling age, they were categorized
as a group of normal schooling students. If students were older than the

4
W. Li et al. Research in Social Stratification and Mobility 71 (2021) 100565

most frequent or usual age for the grade they attended, they were Table 2
categorized as a group of left-behind students. Descriptions of individual-level and school-level variables.
Variables Mean (SD)/percentage
3.2.2. School-related variables
Dependent variable
School location was reported by principals, who indicated whether Self-esteem 27.555(4.521)
their school was located in an urban area or a rural area (rural = 0, urban Individual level (N ¼ 23,627)
= 1). Based on the administrative division of China (The State Council, Family income 3.57 (1.909)
2006), the urban areas belong to the jurisdiction residents committee Parental education 2.484 (1.134)
Subjective family income 2.89 (.725)
and the rural areas belong to the village committee. Urban area refers to Age 12.765 (1.732)
the area related to the actual construction of residential committees and Behind age (normal age) 96.7 %
other residential buildings in cities, district-level cities, districts, and Gender (Boys = 1) 52.9 %
municipalities. Rural area refers to the area outside the town, defined by School level (N ¼ 604)
School location (urban school) 62.1 %
the regulation of the Chinese administrative division.
Peer objective family income 3.595 (.2771)
Peer family SES was averaged by individual objective income, indi­ Peer parental education 2.493 (.797)
vidual subjective income, and individual parental education from a Peer subjective family income 2.891 (.277)
given school, which was standardized at the school level (Marsh & Hau, Book collections 2.527 (1.156)
2003). Equipment 2.347 (1.763)
Teacher education .462 (.353)
School material resources included equipment and books, which were Teacher/student ratio .061 (.023)
all reported by principals. Equipment was reported with the application
of teaching equipment (e.g., multimedia, projection) in the school,
namely, ‘how many classes (the approximate rate) are equipped with
Table 3
multimedia, projection, etc.?’. Responses were rated on a five-point
Estimated parameters of hierarchical multilevel regression predicting self-
scale from 1 (0− 20%) to 5 (80 %–100 %); higher scores indicated that
esteem by individual family SES (objective income, parental education, sub­
more equipment was available in schools. Books were measured with jective income), school contextual variables (school location, peer family SES,
one item: ‘how many books are there in your school library?’. Responses school resources), and interactions between individual family SES and school
were rated on a four-point scale from 1 (less than 500) to 4 (more than levels.
5000). Again, higher scores indicated more school resources.
M0 M1 M2 M3
Teacher-related resources included teacher education and teacher/
Fixed effect
student ratio. Teacher education was reported by the percentage of
Individual Level
teachers who hold at least a bachelor’s degree. The teacher/student ratio Family income (FI) .055*** .038*** .029**
was calculated by dividing the number of teachers by the number of (.007) (.007) (.011)
students in a school. Subjective income .036*** .032*** .017(.011)
(SI) (.007) (.007)
Parental education .130*** .089*** .102**
3.3. Analytical strategies
(PE) (.008) (.008) (.012)
School level
In the present study, the percentages of missing data ranged from 0.1 Peer parental .138*** .123***
% for self-esteem to 18.5 % for books. We addressed missing data using education (.019) (.019)
EM (Expectation-Maximization) in SPSS 17.0 before conducting a Teacher education .047*** .056***
(.012) (.012)
multilevel analysis. To explore our research questions, we used hierar­ Teacher/student .037*** .036***
chical linear model analysis to explore the relations between SES and ratio (.010) (.010)
self-esteem at the student level and school level. We estimated param­ Cross-Level
eters with a restricted iterative generalized least square method by HLM School location x .043*(.020)
subjective income
version 6.08 (Muthén & Muthén 2012). Before estimation, we stan­
Peer parental .049***
dardized all measures of variables across the entire sample (see Marsh, education x parental (.014)
1987). Peer family SES was determined by taking the average of indi­ education
vidual SES in each school (but with no restandardization so that both Teacher education − .025***
individual and school-level scores were in the same metric). x parental education (.009)
Random effect
We explored our research questions by the following models. First, School-level variance .119(.346) .062*** .034*** .033***
we computed a null model (M0) with a random intercept but no (.250) (.185) (.182)
explanatory variables. In M0, the total variances in the dependent var­ Student-level variance .889(.943) .881(.939) .880(.937) .878(.937)
iable (self-esteem) were decomposed into variations among students and Objective income .001(.026)
slope variance
schools. Second, we tested the family SES model (M1) by adding the
Subjective income .001(.026)
three SES indicators (objective, subjective family income, parental ed­ slope variance
ucation) simultaneously, but with fixed slopes. M1 can test whether a Parental education .002(.040)
significant amount of the variance in self-esteem could be explained by slope variance
family SES after accounting for other individual characteristics. Third, in Number of estimated 2 2 2 11
parameters
model 2 (M2), we added school location, peer family SES, and school − 2LL 65361.172 64897.949 64657.778 64778.137
resources as the predictors for the level-2 intercept, with family SES ΔR2 a 0.785 % 45.289 % 33.591 %
slopes fixed. M2 can test the main effects of school contextual variables
Note. Only significant standardized regression coefficients with standard error in
on self-esteem. Fourth, in model 3 (M3), we added school location, peer
parentheses are shown in table. M0 = the null model without any predictors;
family SES and school resources as predictors for the random slopes of M1= the family SES model with three family indicators added simultaneously;
family SES, interacting with all three family SES indicators simulta­ M2 = the school context model with school variables added into the school level
neously. M3 can test whether the effects of individual family SES on self- intercept; M3 = the cross-interaction model with interactions between school
esteem varied by the school variables. variables and family SES. a: ΔR2 indicated the effect size for the variables added
to the models.
* p < .05, ** p < .01, *** p < .001.

5
W. Li et al. Research in Social Stratification and Mobility 71 (2021) 100565

4. Results stronger in urban schools, and the association between parental edu­
cation and self-esteem was stronger in schools with higher peer parental
Table 2 shows the descriptive statistics for level-1 and level-2 vari­ education but was weaker in schools with higher teacher educational
ables. Table 3 shows the main analysis results of our multilevel regres­ level.
sion models. The intraclass correlation coefficient (ICC) of the null
model suggests that 11.8 % of the variance in students’ self-esteem was 5.1. The positive associations between family SES and self-esteem
explained by school variables. With the M1, M2, M3 models, we
sequentially analyzed the associations between family SES, school We find that students’ family SES (objective income, parental edu­
contextual variables, and their interactions with self-esteem. We added cation, and subjective income) significantly and positively related to
three SES indicators simultaneously at the individual level, school level, self-esteem, which is similar to previous findings (Varga et al., 2014). In
and cross-level interactions. In this way, the full model results became line with the reflected appraisals model, we found that high objective
greatly complicated for the 18 cross-level interaction variables. To income and well-educated parents were an accessible way to identify
simplify the presentation of the results, we only showed the significant privilege, improving self-esteem. As we predicted, we also found the
results of all models as follows. positive role of subjective SES in self-esteem. According to referent
group theory, when students compare themselves with a referent group
4.1. The associations between family SES and student’s self-esteem on family wealth, they can feel themselves in a certain position of the
social hierarchy in the immediate context and form a self-evaluation
Table 3 shows that with family SES indicators simultaneously added about their social class, thereby forming self-esteem. Interestingly, we
into M1, the three indicators (objective income, subjective income, found that parental education accounted for a larger proportion of
parental education) of family SES significantly and positively related to variances in self-esteem than family income did. This result is similar to
self-esteem (objective income βFI=.055, p<.001; subjective income βSI another Chinese study showing that intangible resources (i.e., parental
=.036, p < .001; parental education βPE = .130, p < .001), which education) count more than material resources for students’ outcomes
respectively accounted for 7.326 %, 1.587 %, and 32.031 % of the (Liu & Xie, 2015). Well-educated parents have higher social status,
individual-level variance in self-esteem. which leads to others’ positive evaluations of their children. Further­
more, they have a better knowledge of parenting practices and more
4.2. The associations between school contextual variables and students’ positive expectations for interactions with their children (Liu & Xie,
self-esteem 2015), benefiting their children’s self-esteem development. In other
words, these results reflect that different family SES indicators matter
Overall, model 2 (Table 3) shows the significant predictors of self- differently to students’ self-esteem, which suggests that educational
esteem at the school level. Specifically, peer parental education (β = practitioners should put more attention on students who subjectively
.138, p < .001), teacher education (β = .047, p < .001), and teacher/ identify as poor students and those whose parents are undereducated.
student ratio (β = .037, p < .001) significantly and positively related to
self-esteem, which respectively accounted for 12.423 %, 3.710 %, and 5.2. The positive associations between school contexts and students’ self-
3.056 % of the school-level variance in self-esteem. However, school esteem
location, peer family income, peer subjective family income, book col­
lections, and equipment did not significantly relate to self-esteem. As one of our major concerns, we explored the associations between
school contextual variables and students’ self-esteem. We found that
4.3. Variance in the associations between family SES and self-esteem peer parental education and teacher-related resources (teacher educa­
across school contextual variables tion and teacher/student ratio) significantly and positively related to
self-esteem. According to self-categorization theory, individuals’ self-
Finally, we simultaneously added school contextual variables into esteem can derive from memberships in social groups, together with
the cross-level interaction model to interact with each of the three family the value and overall evaluation attached to those memberships
SES indicators (M3, Table 3) and found that school location positively (Crocker et al., 1994). Because of the high social status of scholars in
interacted with subjective income (β = .043, p < .05), which accounted Chinese societies, being associated with the groups of well-educated
for 15.054 % of the variance in the slope of subjective family income on parents enables students to identify self as a member of the high social
self-esteem. Peer parental education positively interacted with parental class. Consequently, students can share the high self-esteem of a
education (β = .049, p < .001), which accounted for 13.684 % of the distinguished group. Schools with a high quality and quantity of
variance in the slope of parental education on self-esteem. However, teacher-related resources also have a great reputation for their superior
teacher education negatively interacted with parental education (β = schooling and teaching. Thus, students in those schools are more likely
− .025, p < .001), accounting for 7.865 % of the variance in the slope of to categorize themselves as members of these privileged groups, thereby
parental education on self-esteem. All the remaining school contextual deriving positive self-evaluations, resulting in high self-esteem.
variables did not significantly interact with family SES (objective in­ However, we did not find significant associations between school
come, parental education, subjective income). location, peer objective income, peer subjective income, equipment,
book collections, and self-esteem. The reason why there is no effect of
5. Discussion peer objective and peer subjective income could be related to the rapid
development of the Chinese economic and education market. As the
To test the moderation effect of school resources on the link between Chinese economy has grown, the educational market has become
family and self-esteem, this study used data from a national survey of competitive and has expanded. Most Chinese parents, regardless of their
Chinese children and adolescents to explore the associations between wealth, are willing to invest in children’s education, such as extracur­
family SES and self-esteem, as well as the extent to which these associ­ ricular training and electronic equipment. Peer family income may not
ations vary across school contexts. The HLM analysis showed that family be accessible enough to identify privileged groups and did not signifi­
SES (measured by objective income, parental education, and subjective cantly matter for self-esteem. In this study, book collections were re­
income) positively and significantly related to Chinese students’ self- ported based on the approximate number of books in a school’s library
esteem. At the school level, peer parental education, teacher educa­ in total, rather than exact numbers for each student, decreasing the
tion, and teacher/student ratio are positively related to self-esteem. The accuracy of this variable. Thus, the number of books as measured in this
positive association between subjective income and self-esteem was study may not be sensitive enough to predict the variance in self-esteem

6
W. Li et al. Research in Social Stratification and Mobility 71 (2021) 100565

across schools. The positive results of peer parental education and schools is inhibited. Therefore, the moderation effect by peer parent
teacher-related resources suggest the important role of invisible re­ education on parental education/self-esteem links might not relate to
sources, rather than material resources, in self-esteem. students’ achievement. However, to clarify the peer parental education
effect, students’ achievement could be included to improve the expla­
5.3. The variances of associations between family SES and self-esteem nation in the future.
across school contexts Finally, in line with the public resources substitution hypothesis, we
found that teacher education (rather than the teacher/student ratio)
Given the positive effect of school context on self-esteem, we buffered the association between parental education and self-esteem.
continued to explore whether the associations between family SES and Teachers with high academic degrees are much more likely to provide
self-esteem vary across school contexts. As we predicted, we found that positive and warm support and to employ constructive teaching prac­
the cross-level interactions were neutral, positively, or negatively tices to relieve poor students’ negative self-evaluations. Particularly in
related to self-esteem, depending on the types of school contextual China, teachers are expected to play a role similar to parents, guiding
variables and family indicators. and caring for students as much as their children (Lo & Lo, 2001; Tian
First, we found that the association between family income and self- et al., 2013). However, we did not find a moderating effect of the
esteem did not vary as a function of peer family income or schools’ teacher/student ratio on self-esteem, which again indicates that the
material resources, such as books and equipment. As the social repro­ quality, not the quantity, of teacher-related resources is important.
duction hypothesis assumed, students can still perceive their social Nonetheless, we admit that the effect of teacher education could relate
status as the way that others evaluate them by their visible wealth (e.g., to the endogeneity issue. Highly educated teachers may gather in
fancy clothes, iPad), regardless of how the school resources and peer SES schools with well-behaved students, which could improve an in­
vary. Thus, it is reasonable that the impact of objective family income dividual’s self-esteem. However, the allocation of educated teachers in
was constant. China is mostly driven by teacher employment and management policy.
Second, in line with the complementary intangible hypothesis, we Chinese tenured policy (bianzhi) has predominantly managed teachers’
found that both school location and peer parental education strength­ employment. With the tenured position, teachers can enjoy permanent
ened the positive link between family SES (subjective SES, parental employment, with welfare and insurance financially supported by the
education) and self-esteem. The positive effect of subjective income on government(Wu & Chen, 2018). In this case, teachers would be more
students’ self-esteem was larger in urban schools (rather than rural likely to work in schools with tenured positions rather than in schools
schools), which is similar to the findings of a previous study (Chiu & without these positions. Quite often, some private schools can have
Zeng, 2008). This result reflects the Chinese disparity of the economy much more capable students, even though they do not have tenure po­
and educational market between urban-rural areas. Chinese urban areas sitions to hire highly educated teachers. Consequently, it is still unclear
enjoy better welfare resources, public good resources, and educational whether the moderation effect of teacher education levels on family
resources than rural areas (Treiman, 2012). Moreover, compared with SES/self-esteem link relates to students’ achievement. Student
rural schools, urban schools gather more people from various social achievement should be included in a future study to clarify this issue.
statuses. In this case, according to referent group theory, the referent To summarize, because of different types of school contextual vari­
group can be more accessible for urban parents to subjectively feel their ables and family SES indicators, the links between family SES and self-
social status. In particular, Chinese self-evaluation considerably depends esteem complicatedly vary across school contextual variables. With
on the views of one’s surrounding group (Chen et al., 2018). Thus, different moderation patterns by school contexts, we can provide prac­
parents would be more likely to identify the self through social com­ tical advice on how to efficiently employ school resources according to
parison in urban areas. Parents’ perception of family wealth can be students’ family SES background.
passed on to children. Therefore, after controlling for objective income,
students whose parents had higher subjective income experienced 6. Strengths, limitations, and future study
higher self-esteem in urban schools. The Chinese educational market
and related competition has become much more intense. Both in urban Based on the National Children’s Study of China, the present findings
and rural schools, parents would like to invest substantially in their have theoretical and practical implications and strengths that we would
children’s education, regardless of their wealth or education levels (as like to highlight. First, considering the important role of self-esteem in
the social reproduction hypothesis assumed). This is why school location social mobility, we focused on the effect of family SES on self-esteem
did not interact with objective income and parental education to relate rather than achievement. In addition, we simultaneously tested the
to self-esteem. objective and subjective SES indicators, finding their independent roles
We also found that the positive association between parental edu­ in self-esteem. These results can enrich the empirical evidence on the
cation (not family income) and self-esteem was larger in schools with role of family SES in students’ self-esteem, helping improve students’
higher peer parental education. Schools with high peer SES usually self-esteem programs. Second, we explore not only the associations be­
provide scientific and complicated learning curriculum and tactics (Tao tween school contextual variables and self-esteem but also the moder­
et al., 2013). Students from well-educated families are more likely to ations of school contexts on the associations between family SES and
succeed in schools and thus have more possibilities to experience higher self-esteem. The results can contribute to theories explaining the
self-esteem (in line with the complementary intangible hypothesis). school contexts’ effect on self-esteem and provide guidance on what
However, the effect of peer parental education might suggest the types of school effort can be made to improve the self-esteem of students
endogeneity problem. For example, parents with high education levels from different SES families. Based on our findings, some suggestions can
might reflect on their choices for schools with high achievement stu­ be provided for school practice. First, we show that teacher education
dents, which greatly relates to their children’s self-esteem. If so, schools can buffer the negative influences of parental education on self-esteem,
should equip more complicated schooling tasks for well-performing suggesting that more school investment should be used to improve
students, which still requires parents’ education practice. Alterna­ teachers’ education and ability, especially teachers in less advantaged
tively, the more able students could achieve due to their ability rather schools. High-quality teaching practices will benefit students who lack
than parental practice, which conflicts with our results. Moreover, parents’ tutoring and support. Second, with the roles of subjective SES in
parental choice for better schools is not allowed in China. According to self-esteem, teachers should direct students to focus on ability growth
the Compulsory Education Law of the People’s Republic of China (Na­ rather than wealth when they compare themselves with the referent
tional Peoples’ Congress, 1986), school placement should be decided on group. As parental education has a strong effect on students’ self-esteem,
the basis of proximity. School choice for primary schools and junior more attention and assistance should be allocated to improve parenting

7
W. Li et al. Research in Social Stratification and Mobility 71 (2021) 100565

practices from low SES families. Farrell, L., Sijibengak, A., & Barrett, P. (2009). An examination of childhood anxiety,
depression and self-esteem across socioeconomic groups- A comparison study
While our findings have important implications, several limitations
between high and low socioeconomic status school communities. Advances in School
are inherent. First, family SES variables were reported by parents rather Mental Health Promotion, 2(1), 5–19.
than students. This means that the data cannot represent students’ Feng, X., & Zhang, X. (2008). Researches about the relationships among pride, explicit
perceptions of family socioeconomic status. In future studies, family SES self-esteem and depression in urban poverty adolescents. Psychological Development
and Education, 4, 100–105.
could be reported by both parents and students. Second, as discussed Gottfried, A. W. (1985). Measures of socioeconomic status in child development research
above, we cannot exclude the endogeneity problem. The moderating : Data and recommendations. Merrill-Palmer Quarterly, 31(1), 85–92.
roles of teacher education and peer parental education in the link of Greenwald, R., Hedges, L., & Laine, R. (1996). The effect of school resources on student
achievement. Review of Education, 66(3), 361–396. http://rer.sagepub.com/content/
family SES/self-esteem could relate to other factors, such as students’ 66/3/361.short.
achievement. Future studies could add student achievement to clarify Han, S. H. (2013). Confucian states and learning life: Making scholar-officials and social
the problems. Third, we aggregated SES from the given schools rather learning a political contestation. Comparative Education, 49(1), 57–71. https://doi.
org/10.1080/03050068.2012.740220.
than individual classes, which may bias our estimates of peer effects. In Hanushek, E. A. (1997). Assessing the effects of school resources on student performance:
future research, sample collection could include class data. An update. Educational Evaluation and Policy Analysis, 19(2), 14–164. https://doi.
. org/10.3102/01623737019002141.
Harris, D. N. (2010). How do school peers influence student educational outcomes?
Theory and evidence from economics and other social sciences. The Teachers College
Declaration of Competing Interest Record, 112(4), 1163–1197. https://doi.org/10.1017/CBO9781107415324.004.
Herman, K. C., Bi, Y., Borden, L. A., & Reinke, W. M. (2011). Latent classes of psychiatric
symptoms among Chinese children living in poverty. Journal of Child and Family
The authors report no declarations of interest.
Studies, 21(3), 391–402. https://doi.org/10.1007/s10826-011-9490-z.
Hwang, K. K. (2012). Life goals and achievement motivation in Confucian society. In
Acknowledgements K. K. Hwang (Ed.), Foundations of Chinese psychology: Confucian social relations
International and cultural psychology (pp. 219–264). Springer.
Kraus, M. W., Piff, P. K., Mendoza-Denton, R., Rheinschmidt, M. L., & Keltner, D. (2012).
Funded by MOE (Ministry of Education in China) Project of Hu­ Social class, solipsism, and contextualism: How the rich are different from the poor.
manities and Social Sciences, China (Project No.17YJCZH014);Foun­ Psychological Review, 119(3), 546–572. https://doi.org/10.1037/a0028756.
dation of "13th Five-Year Plan" for Guangzhou Philosophy and Social Li, X., Zou, H., Wang, R., & Dou, D. (2008). The developmental characteristics of
temporary migrant children. Self-Esteem and its relation to learning behavior and
Sciences Development, China (Foundation No. 2018GZYB70); Foun­ teacher-student relationship in Beijing. Journal of Psychological Science, 31, 909–913.
dation of "13th Five-Year Plan" for Guangdong Philosophy and Social https://doi.org/10.16719/j.cnki.1671-6981.2008.04.025.
Sciences Development, special project, China (Foundation No. Liu, A., & Xie, Y. (2015). Influences of monetary and non-monetary family resources on
children’s development in verbal ability in China. Research in Social Stratification and
2019GXJK004). Mobility, 40, 59–70. https://doi.org/10.1016/j.rssm.2015.02.003.
Liu, Y., Li, X., Chen, L., & Qu, Z. (2015). Perceived positive teacher-student relationship
References as a protective factor for Chinese left-behind children’s emotional and behavioural
adjustment. International Journal of Psychology, 50(5), 354–362. https://doi.org/
10.1002/ijop.12112.
Adler, N. E., Epel, E. S., Castellazzo, G., & Ickovics, J. R. (2000). Relationship of
Liu, X., & Zhao, J. (2016). Chines migrant children the moderating roles of group identity
subjective and objective social status with psychological and physiological
and types of school. PloS One, 1. https://doi.org/10.1371/journal.pone.0146559.
functioning: Preliminary data in healthy, White women. Health Psychology, 19(6),
Lo, R., & Lo, R. C. C. (2001). The role of class teachers in a key secondary school in
586–592. https://doi.org/10.1037/0278-6133.19.6.586.
Shanghai. Pastoral Care in Education, 19(1), 20–27.
Blossfeld, H. P., & Shavit, Y. (1991). Persisting barriers: Changes in educational. PhD
Major, B., & O’Brien, L. T. (2005). The social psychology of stigma. Annual Review of
Proposal. https://doi.org/10.1017/CBO9781107415324.004.
Psychology, 56, 393–421. https://doi.org/10.2307/3089335.
Borman, G. D., & Overman, L. T. (2004). Academic resilience in mathematics among
Mann, M., Hosman, C. M. H., Schaalma, H. P., & De Vries, N. K. (2004). Self-esteem in a
poor and minority students. The Elementary School Journal, 104(3), 177. https://doi.
broad-spectrum approach for mental health promotion. Health Education Research,
org/10.1086/499748.
19(4), 357–372. https://doi.org/10.1093/her/cyg041.
Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and
Marsh, H. W. (1987). The big-fish-little-pond effect on academic self-concept. Journal of
design. Harvard University Press.
Educational Psychology, 79(3), 280–295. https://doi.org/10.1037/0022-
Caldas, S. J., & Bankston, C. (1997). Effect of school population socioeconomic status on
0663.79.3.280.
individual academic achievement. The Journal of Educational Research, 90(5), 269.
Marsh, H. W., & Hau, K.-T. (2003). Big-fish–Little-pond effect on academic self-concept:
https://doi.org/10.1080/00220671.1997.10544583.
A cross-cultural (26-Country) test of the negative effects of academically selective
Chen, Y., & Fan, X. (2019). Subjective social status, income inequality and subjective
schools. The American Psychologist, 58(5), 364–376. https://doi.org/10.1037/0003-
perceptions of mobility (2003-2013). Social Sciences in China, 40(3), 70–88.
066X.58.5.364.
Chen, X., Fu, R., Liu, J., Wang, L., & Zarbatany, L. (2018). Social sensitivity and social,
Marsh, H. W., Kong, C.-K., & Hau, K.-T. (2000). Longitudinal multilevel models of the
school, and psychological adjustment among children across contexts. American
big-fish-little-pond effect on academic self-concept : Counterbalancing contrast and
Psychological Association, 54, 1124–1134. https://doi.org/10.1037/dev0000496.
reflected-glory effects in Hong Kong schools. Journal of Personality and Social
Chiu, Ming. Ming (2010). Effects of inequality, family and school on mathematics
Psychology, 78(2), 337–349. https://doi.org/10.1037//0022-3514.
achievement: country and student differences. Social Forces, 88(4), 1646–1675.
Muthén, L. K., & Muthén, B. O. (2012). Mplus User’s guide (No. 7; seventh ed). Muthén &
Chiu, M. M., & Zeng, Xihua (2008). Family and motivation effects on mathematics
Muthén.
achievement: Analyses of students in 41 countries. Learning and Instruction, 18(4),
OECD. (2013). PISA 2012 results : What makes schools successful? Resources, policies and
321–336. https://doi.org/10.1016/j.learninstruc.2007.06.003.
practices (Vol. IV).
Conger, R. D., Conger, K. J., & Martin, M. J. (2010). Socioeconomic status, family
Opdenakker, M., & Damme, J. Van (2007). Do school context, student composition and
processes, and individual development. Journal of Marriage and the Family, 72,
school leadership affect school practice and outcomes in secondary education?
685–704. https://doi.org/10.1111/j.l741-3737.2010.00725.x.
British Educational Research Journal, 33(2), 179–206. https://doi.org/10.1080/
Crocker, J., Luhtanen, R., Blaine, B., & Broadnax, S. (1994). Collective self-esteem and
01411920701208233.
psychological well-being, among white, black, and Asian college students.
Operario, D., Adler, N. E., & Williams, D. R. (2004). Subjective social status: Reliability
Personality and Social Psychology Review : an Official Journal of the Society for
and predictive utility for global health. Psychology & Health, 19(2), 237–246. https://
Personality and Social Psychology, Inc, 20(5), 503–513. https://doi.org/10.1177/
doi.org/10.1080/08870440310001638098.
0146167294205007.
Orth, U., Robins, R. W., & Widaman, K. F. (2012). Life-span development of self-esteem
Crosnoe, R. (2009). Low-income students and the socioeconomic composition of public
and its effects on important life outcomes. Journal of Personality and Social
high schools. American Sociological Review, 74(5), 709–730. https://doi.org/
Psychology, 102(6), 1271–1288.
10.1177/000312240907400502.
Pelham, B. W. (1995). Further evidence for a Jamesian model of self-worth: Reply to
Darling-hammond, L. (2000). Teacher quality and student achievement : A review of
Marsh. Journal of Personality and Social Psychology, 69(6), 1161–1165. https://doi.
state policy evidence previous research. Education, 8(1), 1–44. https://doi.org/
org/10.1037//0022-3514.69.6.1161.
10.1038/sj.clp.
Perry, L., & McConney, A. (2010). Does the SES of the school matter? An examination of
Diener, E. D., & Biswas-Diener, R. (2002). Will money increase subjective well-being? A
socioeconomic status and student achievement using PISA 2003. Teachers College
literature review and guide to needed research. Social Indicators Research, 57,
Record, 112(4), 1137–1162. http://www.tcrecord.org/content.asp?contentid
119–169.
=15662.
Dong, Q., & Lin, C. D. (2011). Handbook for databases on the national children’ study of
Piko, B., & Fitzpatrick, K. M. (2001). Does class matter? SES and psychosocial health
china. Science Press.
among Hungarian adolescents. Social Science & Medicine, 53(6), 817–830.
Falci, C. D. (2011). Self-esteem and mastery trajectories in high school by social class and
Rjosk, C., Richter, D., Hochweber, J., Lüdtke, O., Klieme, E., & Stanat, P. (2014).
gender. Social Science Research, 40(2), 586–601. https://doi.org/10.1016/j.
Socioeconomic and language minority classroom composition and individual
ssresearch.2010.12.013.

8
W. Li et al. Research in Social Stratification and Mobility 71 (2021) 100565

reading achievement : The mediating role of instructional quality. Learning and van Dick, R., Wagner, U., Stellmacher, J., & Christ, O. (2005v). Category salience and
Instruction, 32, 63–72. https://doi.org/10.1016/j.learninstruc.2014.01.007. organizational identification. Journal of Occupational and Organizational Psychology,
Rosenberg, M., & Pearlin, L. (1978). Social class and self-esteem among children and 78(2), 273–285. https://doi.org/10.1348/096317905X25779.
adults. The American Journal of Sociology, 84(1), 53–77. http://www.jstor.org/stabl van Ewijk, R., & Sleegers, P. (2010v). The effect of peer socioeconomic status on student
e/2777978. achievement: A Meta-Analysis. Educational Research Review, 5(2), 134–150. https://
Rosenberg, M., Schooler, C., Schoenbach, C., & Rosenberg, F. (1995). Global self-esteem doi.org/10.1016/j.edurev.2010.02.001.
and specific self-esteem : Different concepts, different outcomes. American Varga, S., Piko, B. F., & Fitzpatrick, K. M. (2014). Socioeconomic inequalities in mental
Sociological Association, 60(1), 141–156. well-being among Hungarian adolescents: A cross-sectional study. International
Rumberger, R. W., & Palardy,, G. J. (2005). Does segregation still matter? The impact of Journal for Equity in Health, 13(1), 100. https://doi.org/10.1186/s12939-014-0100-
student composition on academic achievement in high school. Teachers College 8.
Record, 107(9), 1999–2045. https://doi.org/10.1111/j.1467-9620.2005.00583.x. Veselska, Z., Madarasova Geckova, A., Gajdosova, B., Orosova, O., Van Dijk, J. P., &
Ryabov, I. (2011). Adolescent academic outcomes in school context: network effects Reijneveld, S. A. (2010). Socio-economic differences in self-esteem of adolescents
reexamined. Journal of Adolescence, 34(5), 915–927. https://doi.org/10.1016/j. influenced by personality, mental health and social support. European Journal of
adolescence.2010.12.004. Public Health, 20(6), 647–652. https://doi.org/10.1093/eurpub/ckp210.
Stets, J. E., & Burke, P. J. (2000). Identity theory and social identity theory. Social Wang, Q., Shen, J. J., Sotero, M., Li, C. A., & Hou, Z. (2018). Income, occupation and
Psychology Quarterly, 63(3), 224–237. education: Are they related to smoking behaviors in China? PloS One, 13(2), 1–18.
Tao, Y., Oliver, M., & Venville, G. (2013). A comparison of approaches to the reaching https://doi.org/10.1371/journal.pone.0192571.
and learning of science in Chinese and Australian elementary classrooms: Cultural Wiltfang, G. L., & Scarbecz, M. (1990). Social Class and adolescents’ self-esteem: Another
and socioeconomic complexities. Journal of Research in Science Teaching, 50(1, SI), look. Special Issue: Social Structure and The Individual, 53(2), 174–183.
33–61. https://doi.org/10.1002/tea.21064. Wu, Zhi-hui, & Chen, Chang-sheng (2018). The contradiction between supply and
The State Council. (2006). The regulations on the division of urban and rural areas. www. demand of public school teachers and its reform methods in compulsory education in
stats.gov.cn/tjsj/pcsj/rkpc/5rp/html/append7.htm. China. Educational Research, 8(463), 88–100.
Tian, L., Liu, B., Huang, S., & Huebner, E. S. (2013). Perceived social support and school Zeng, X., Li, W., Wang, Y., Li, J., Huang, X., & Li, X. (2017). Normative importance of
well-being among Chinese early and middle adolescents: The mediational role of money, family income, and self-esteem: A multilevel latent modeling analysis of data
self-esteem. Social Indicators Research, 113(3), 991–1008. https://doi.org/10.1007/ from Chinese early adolescents. Social Indicators Research, 1–16. https://doi.org/
s11205-012-0123-8. 10.1007/s11205-016-1243-3.
Treiman, D. J. (2012). Urban–rural disparities.pdf. Research in Social Stratification and Zheng, W. Y. (2013). High school students’ self-esteem development——Empirical research
Mobility, 30, 33–47. from WuXi in China. Nanjing Normal University.
Twenge, J. M., & Campbell, W. K. (2002). Self-esteem and socioeconomic status: A meta-
analytic review. Personality and Social Psychology Review, 6(1), 59–71. https://doi.
org/10.1207/S15327957PSPR0601_3.

You might also like