Chinese Sociological Review

2024, VOL. 56, NO. 3, 313–337

https://doi.org/10.1080/21620555.2024.2321122

Is shadow education a myth? How schools

affect private tutoring in China
Zhidi Pan
Zhejiang University, Hangzhou, China

ABSTRACT
Shadow education, known as private tutoring, is
expanding rapidly worldwide. While mainstream
studies explain shadow education from a family-cen-
tered perspective, this study focuses on the role of
the school. Using data from the China Panel Education
Survey, this study discerns that active communica-
tions between schools and families can elicit private
tutoring involvement. The school-initiated communi-
cations yield more positive and significant effects
when schools impose heightened assessment pres-
sure on teachers, suggesting an inextricable linkage
between shadow education and formal schooling.
The school effect is also more pronounced for eco-
nomically disadvantaged families. These findings pro-
vide important insights for understanding how
schools shape social stratification in China.

Over the past decades, shadow education, commonly known as private

tutoring, has expanded worldwide and has become a multibillion-dollar
industry (Entrich 2020; Byun, Chung, and Baker 2018). The term “shadow
education” was first coined by Stevenson and Baker (1992) to denote fee-
based organized learning activities conducted outside the formal school
hours (Jansen, Elffers, and Jak 2021; Bray 1999). Shadow education has
long been prominent in East Asian countries, including China, Japan, and
South Korea (W. Zhang and Yamato 2018; Entrich 2017; Lee and Shouse
2011). Its proliferation extends to South Asia, Western and Southern
Europe, North America, and other parts of the world (Bray 2021; Kim and
Jung 2019). Many studies have consistently highlighted the correlation

CONTACT Zhidi Pan zhidipan@zju.edu.cn Department of Sociology, 866 Yuhangtang Rd,

Zhejiang University, Hangzhou, 310000, China.
© 2024 Taylor & Francis Group, LLC
314 Z. PAN

between private tutoring and students’ academic achievement (Kuan 2018;

Y. Zhang and Xie 2016; Choi and Park 2016). Nevertheless, shadow edu-
cation has sparked public concerns about educational inequality. Privileged
families can mobilize economic, social, and cultural capital to seek com-
petitive advantages from private tutoring, potentially reproducing and
exacerbating the existing educational inequality (Jansen, Elffers, and Jak
2021; Byun, Chung, and Baker 2018; W. Zhang and Bray 2018; Lucas and
Byrne 2017).
Previous research has demonstrated that various family characteristics
can influence private tutoring. From a rational choice perspective, parents
invest in private tutoring because they can afford it and believe their chil-
dren can benefit. Since socioeconomically advantaged parents possess rich
resources, they are more likely to enroll their children in private tutoring
than their disadvantaged counterparts, leading to a widening social class
gap in shadow education investment (Byun, Chung, and Baker 2018;
Entrich 2017; Park et al. 2016). Furthermore, parental decisions and prac-
tices are not simply rational but shaped by social norms and peer influ-
ences. Norms established by educational policies and status competition
among families can contribute to an intensive involvement in private
tutoring (W. Zhang and Bray 2020; Lee and Shouse 2011; Schaub 2010).
This strand of research predominantly focuses on a family-centered expla-
nation, assuming that shadow education parallels mainstream schooling
and exerts an independent impact on the student learning net of formal
schooling processes.
However, formal schooling and private tutoring are not mutually exclu-
sive but inextricably connected. To examine these linkages from an orga-
nizational perspective, researchers have investigated how teachers and
staff elicit parental investment in private tutoring. Prior studies report
that schools with high teaching quality will reduce families’ need for sup-
plementary and remedial tutoring outside of school (Kim and Park 2010;
Ireson 2004). Additionally, schools may establish formal or informal part-
nerships with private tutoring organizations to encourage parental involve-
ment. For example, research reveals that when public schools face financial
constraints, some teachers are willing to take part-time positions in pri-
vate tutoring companies to boost their income even though such action
is prohibited by regulations (W. Zhang 2014; Kobakhidze 2014). Under
this scenario, shadow education becomes an essential and extended part
of formal schooling.
In China, “shadow education” does not parallel mainstream schooling
processes but is the by-product of a highly competitive and exam-ori-
ented educational system. China boasts one of the most extensive and
rapidly expanding markets for shadow education. Statistics show that
from 2014 to 2019, China’s extracurricular tutoring market revenue
 Chinese Sociological Review 315

increased by 67%, from 284.5 to 475.1 billion yuan (Deloitte 2018).

According to a sizeable educational survey conducted by the People’s
Forum Questionnaire Center in 2019, covering kindergarten to high
school levels, 76.4% of families enrolled their children in private tutoring
classes (Shi and Yin 2020). Moreover, Chinese schools must comply with
administrative assessment rules and meet families’ needs to improve their
children’s exam performance. Schools strive to mobilize all available orga-
nizational resources to enhance students’ test abilities, which is consistent
with the goal of private tutoring institutions. Therefore, teachers are
essential in eliciting private tutoring behaviors across different school con-
texts. Teachers’ proactive interactions and communications with students
and parents may explain why families invest in private tutoring.
Using data from the China Panel Education Survey (CEPS), this
research delves into how schools affect parental decisions on private
tutoring through the lens of organizational systems. This study examines
two overarching research questions. First, how do teacher-parent commu-
nications influence students’ participation in private tutoring, net of social
background factors? Second, how do teacher-parent communications’
effects on private tutoring involvement vary by school and family charac-
teristics? This study contributes to the existing literature by focusing on a
school-family interaction framework, offering a more comprehensive
understanding of shadow education. More importantly, this study pro-
vides theoretical and empirical insights into comprehending the role of
the school process in eliciting private tutoring, discerning whether schools
serve as equalizers that promote social mobility or hidden institutions
that perpetuate educational inequality.

Theoretical framework
Rational choice perspective
Prior research in Western countries demonstrated that families invest
in shadow education out of rational consideration (Entrich 2017;
Boudon 1974). Their decision-making process involves assessing the
costs, risks, and benefits of educational investment, guided by the avail-
able information and resources. Since household resources vary across
social classes, parental investment and children’s achievements also vary
across social groups (Liu, Jiang, and Chen 2020; Davies 2024). As
demonstrated in social stratification research, families of higher socio-
economic status are more inclined to engage in extracurricular tutoring
to increase academic success, potentially contributing to educational
inequality (Jansen, Elffers, and Jak 2021; Byun, Chung, and Baker 2018;
Park et al. 2016; Buchmann, Condron, and Roscigno 2010).
316 Z. PAN

Specifically, advantaged families have a more profound understanding

of educational returns than their disadvantaged counterparts, thereby
developing high educational expectations and active parenting attitudes
(Carolan and Wasserman 2015; Bodovski 2010). Additionally, advan-
taged families can have persistent educational investments to garner
higher returns (Entrich and Lauterbach 2019). With rich economic and
cultural capital, they enhance children’s cognitive abilities through
high-quality extracurricular classes (Vincent and Ball 2007). Furthermore,
many upper-middle-class parents adopt intensive parenting practices to
increase educational success (Fingerman et al. 2012; Chua 2011; Barg
2019; Milkie and Warner 2014; Lareau 2011). In this way, advantaged
families can always expect higher educational returns with increased
investment.
In contrast, disadvantaged families have less faith in the returns from
educational investment. Due to the relatively insufficient economic and
social resources, they struggle to provide sustained support in shadow
education. Even if they can afford short-term private tutoring, they often
lack the knowledge and time to instruct and accompany their children
(Day and Dotterer 2018; McLanahan and Percheski 2008). Their under-
investment in parenting styles may impede children’s academic develop-
ment, leading to poor expectations of educational returns. Disadvantaged
families, thus, are less willing to invest in shadow education.
Consistent with the theory of critical sociology of education, the above
analysis suggests that rational choice on shadow education ultimately
solidifies the existing social structure (Bourdieu 1973; Bowles and Gintis
1976). The rapidly growing private tutoring market has facilitated the
advantaged social groups to monopolize high-quality educational resources
and maintain their elite status. Furthermore, the dominant class dissemi-
nates its morality and ideology throughout educational involvement,
forming a “cultural hegemony”—a legitimized dominant power that fur-
ther deepens educational inequality.

Social norm perspective

Another strand of research, represented by Japanese and South Korean
scholars, contends that families’ educational decisions are not solely ratio-
nal but shaped by social norms (Yang 2017; Woo and Hodges 2015).
Social norms regulate interpersonal relationships, construct people’s
understanding of the world, and shape and stabilize social behavior. Social
norms can legitimize specific behavior in people’s minds as natural and
desirable to be followed even without external incentives (Scott 2013;
Meyer 1977). This perspective helps explain how individual educational
choices evolve into group preferences.
 Chinese Sociological Review 317

The spread of educational norms is closely tied to a growing “schooled

society.” Individuals involved in formal schooling have constructed funda-
mental social concepts such as “discipline,” “profession,” and “qualifica-
tion,” signifying the importance of modern education. These concepts
imply a transformation of schools from small teaching places into the
cultural cradle of contemporary society, ultimately forming a “schooled
society” (Lee and Shouse 2011; Baker 2009). As the influence of the
“schooled society” extends globally, educational involvement evolves into
a shared value worldwide (Mori and Baker 2011).
In many countries, public policies actively encourage parental educational
involvement, transforming it into a normative behavior (Ishizuka 2019;
Schaub 2010; Bianchi, John, and Robinson 2006; Domina 2005). As a prom-
inent part of parental involvement, shadow education investment has grad-
ually become normalized and legitimized throughout students’ growth
experiences (Kumar and Chowdhury 2021; W. Zhang and Bray 2020; Baker
and LeTendre 2005). For instance, a prior study reported an exceptionally
high rate of tutoring participation in South Korea and found that these
families invest in shadow education primarily to follow the prevailing edu-
cational norms and seek recognition from others (Lee and Shouse 2011).
At the micro level, social norms are reproduced via social networks. Peer
networks, for instance, yield a substantial impact on students. When most
classmates engage in shadow education, students usually experience peer
pressure and are compelled to imitate this behavior (Kim, Jang, and Kim
2022; Pan, Lien, and Wang 2022; Entrich 2017). Similarly, parents’ reference
groups also shape their educational decisions. In South Korea, mothers
develop similar educational preferences through social networks, fostering
private tutoring involvement (Park, Byun, and Kim 2011; Heo 2022). These
studies underscore the significant association between shadow education
investment and the social norms conveyed through social networks.
Overall, both the rational choice and the social norm perspectives explain
shadow education investment. Nevertheless, current studies predominantly
treat shadow education as an extension of family education that parallels
mainstream schooling. To broaden the understanding of Chinese shadow
education, the author tries to go beyond the family-centered explanation
and introduce a school-family interaction framework.

The role of school

When examining the role of schools in shaping parental investment,
existing literature focuses on two structural factors: school quality and
school SES. First, scholars demonstrated that school quality is negatively
associated with private tutoring investment, supporting the rational choice
perspective (Ireson 2004; Kim and Park 2010). High-quality schools
318 Z. PAN

reduce parental additional educational expenses, but low-quality schools

may prompt parents to rationally invest in shadow education to compen-
sate for inadequate teaching resources (Kim 2004; Bray and Lykins 2012).
Second, researchers found that school SES positively correlates with pri-
vate tutoring participation, aligning with the social norm perspective
(Addi-Raccah and Dana 2015; Bray 2011). High SES schools may symbol-
ize an academically demanding environment, contributing to the norma-
tive pressure about competitiveness and excellence, thereby eliciting
private tutoring investment (de Castro and de Guzman 2010; Resh and
Dar 2012).
Current school studies highlight the structural factors shaping families’
educational investment, sometimes neglecting the process factors. However,
the structural-level analyses tend to depict schools as static resource pro-
viders or simple collections of similar social classes, implying a limited
capacity for schools to alter existing social stratification actively. In con-
trast, a process-level analysis perceives schools as organizations with
dynamic aims and actions (Hoy and Miskel 2013; Bidwell 2013). This per-
spective suggests that schools, through management strategies and interac-
tion processes, can intentionally impact teachers’ teaching methods and
families’ educational involvement, ultimately reshaping the social structure.
School-initiated communications with families represent one of the piv-
otal school processes that shape parental educational behaviors and achieve
schools’ educational goals (Yulianti et al. 2020; Epstein 2019; Islam 2019).
Thus, this study focuses on the impact of school-initiated teacher-parent
communications on families’ private tutoring investment. Regarding this
association, existing theories imply two competing hypotheses. The ratio-
nal choice perspective suggests that teacher-parent communications can
bolster parental trust in the school teaching process, reducing their
demands for supplementary tutoring resources (Hawrot 2022). Conversely,
the social norm perspective posits that teacher-parent communications
might arouse parental concern about children’s academic performance and
educational competition, potentially promoting parental involvement and
investment in children’s education (Yu 2020). Discerning which hypothesis
is more plausible in the Chinese context requires further investigation.

Effects of school process within the Chinese context

In China, schools initiate communications with families to reach specific
goals. Under the Chinese exam-oriented education system, schools must
improve students’ test scores and graduation rates to pass the governmen-
tal evaluations (Zhao 2015). Schools achieving superior performance stand
to be recognized with titles such as key schools and model schools,
accompanied by additional supporting resources (Ye 2015; You 2007). In
 Chinese Sociological Review 319

pursuit of these rewards, Chinese schools must exert control over various
aspects, including teachers’ work behaviors and parents’ educational atti-
tudes, to positively influence students’ learning processes. Schools now
link teacher assessment with student performance to determine teachers’
salaries, promotions, and retention (Walker, Qian, and Zhang 2011; Liu
and Onwuegbuzie 2012). Some schools also establish the goals, forms,
and frequencies of teacher-parent communications to convey schools’
educational values and shape parental beliefs (Chan et al. 2021).
These teacher-parent communications might foster parental involve-
ment and investment in education. Prior research documents that teach-
ers can promote parental educational involvement through deliberate and
effective communication with parents (Islam 2019; Santiago et al. 2016).
Hence, schools often initiate various interactions, such as parent-teacher
meetings and home visits, to positively influence the home environment
(Yulianti et al. 2020; Redding et al. 2004). In China, propelled by assess-
ment pressure, teachers must initiate active communications with parents
to underscore the value of education, raising parental concern about chil-
dren’s academic performance. Also, teachers tend to publicly praise stu-
dents with high grades at teacher-parent conferences and give special
reminders to students with poor grades (Zhao 2015; Pillet-Shore 2015),
potentially eliciting parental anxiety about social comparison and educa-
tional competition. Furthermore, teachers might advise students and par-
ents to utilize extra educational resources, like private tutoring, to enhance
academic advantages (W. Zhang and Bray 2015). In this scenario,
school-initiated communications create a social norm encouraging shadow
education investment. The following hypotheses are thus tested.
Hypothesis 1: School-initiated teacher-parent communications increase the
likelihood of families’ participation in shadow education.

Hypothesis 2: The effect of teacher-parent communications is more pro-

nounced when schools impose heightened performance pressure on
teachers.

Effects of school process across family backgrounds

To investigate the association between school process and social stratifi-
cation, the author further examines whether teacher-parent communica-
tions moderate the effects of family economic backgrounds on private
tutoring involvement. From the rational choice perspective, holding rich
economic and cultural capital, high-income families are more capable of
and willing to participate in private tutoring (Entrich and Lauterbach
2019; Bodovski 2010). They also clearly understand the returns of educa-
tion and develop high expectations, thereby inclined to support the
320 Z. PAN

school’s educational goals (Carolan and Wasserman 2015). On the con-

trary, low-income families usually find private tutoring unaffordable (Kim,
Goodman, and West 2021). Also, low-income working classes are less
involved in school activities and lack the experience to communicate
effectively with teachers, implying that school-initiated communications
may hardly change their educational behaviors (Hornby and Lafaele 2011).
The following hypothesis is thus tested.
Hypothesis 3: The effect of teacher-parent communications on private
tutoring is more potent for high-income families.

Nevertheless, the social norm perspective implies a different scenario.

For high-income strata, the emphasis on children’s education is already
inherent. Before school intervention, private tutoring involvement and
educational competition have become normative behaviors among
wealthy families (Addi-Raccah and Dana 2015; Bray 2011). Consequently,
schools may play a supplementary and additional role in shaping the
educational choices of advantaged families. In contrast, the low-income
working class usually lacks a habitus of educational involvement (Lareau
2011). However, their parenting attitudes might be switched by positive
norms established by schools, suggesting that schools could significantly
impact the educational behaviors of disadvantaged families. Moreover,
compared to the middle class, low-income parents are less likely to
challenge teachers’ authority and are more obedient to their suggestions
(Guo, Wu, and Liu 2019). In this scenario, the author concludes with a
counter-hypothesis.
Hypothesis 4: The effect of teacher-parent communications on private
tutoring is more potent for low-income families.

Data, measures, and methods

Data
This study analyzed data from the China Education Panel Survey
(CEPS). CEPS applied a stratified, multistage sampling design to gener-
ate a school-based national representative sample, collecting comprehen-
sive information on student behaviors, family backgrounds, and school
activities. The baseline survey for the 2013–2014 academic year (Wave
1) randomly selected 10,279 seventh-grade and 9,208 ninth-grade mid-
dle school students in 438 classrooms across 112 schools from 28 coun-
ty-level units in mainland China. The survey administered five
questionnaires to students, their parents, homeroom teachers, major
subject teachers, and principals. In the 2014–2015 academic year, CEPS
conducted the first follow-up survey (Wave 2) of the seventh-grade
 Chinese Sociological Review 321

students as they upgraded to eighth grade. A sample size of 9,449

eighth-grade students (92% response rate) was thus generated and
became the focus of this study. This study filled the individual-level
missing data with multiple imputations based on the chained equation
and the school-level missing data based on the Wave 2 survey data.
After excluding missing dependent variable values, the final sample
comprised 9,147 students.

Measures
The dependent variable in this study is private tutoring participation. The
key independent variable is teacher-parent communications. An array of
control variables, including demographic factors, family characteristics,
and school characteristics, are incorporated to distinguish the net effect
of teacher-parent communications. Table 1 presents the codings and dis-
tributions of all variables.

Methods
Given that the CEPS sample comprises students nested within schools,
this study employs multilevel models for data analysis. Families’ private
tutoring participation and parents’ and children’s sociodemographic char-
acteristics are measured at the individual level (Level 1); teacher-parent
communication and other school background variables are measured at
the school level (Level 2). The analyses proceed in four stages. First,
because the outcome variable is a count variable, the author adopts mul-
tilevel Poisson regression to estimate the general effect of teacher-parent
communications on private tutoring participation. Second, to understand
the mechanisms of the school process, the author delves into whether the
effects of teacher-parent communications vary by levels of school perfor-
mance pressure. Third, the author examines the cross-level interaction
effect between teacher-parent communications and family income levels
to investigate the heterogeneous effects of school-initiated communica-
tions across family backgrounds. Fourth, the author further discusses and
examines the robustness of the results.

Result
Effects of teacher-parent communications on private tutoring
Table 2 reports the factors influencing private tutoring participation
based on multilevel Poisson regression models. The empty model shows
variation in private tutoring participation at the school level. The
322 Z. PAN

Table 1. The descriptive analysis of variables.

Variable codings and distributions
Outcome variable
Private tutoring A count variable reflecting the number of extracurricular private tutoring
participation classes students have participated in over the past year. The classes include
academic courses aimed at improving exam performance and non-academic
hobby classes designed to enhance cognitive ability and socio-cultural
capital.
Mean = 0.98, SD = 1.27, range = [0, 10]
Key independent
variable
Teacher-parent A Z-score continuous variable, with higher values indicating higher levels of
communications school-initiated teacher-parent communications. Its coding is based on the
CEPS principal report about the frequency of school-initiated activities in
the previous semester, including holding parent-teacher conferences,
providing parents with reports on students’ performance, inviting parents to
attend students’ courses, and inviting parents to talk with teachers (zero
times = 1, once = 2, two to four times = 3, five times and more = 4).
Based on the report, factor analysis is adopted to construct this composite
variable (α = .73).
Mean = −0.01, SD = 1.00, range = [−1.86, 2.39]
Individual-level
covariates
Gender Male = 0 (52.07%); Female = 1 (47.93%)
Migration status Non-migrant = 0 (81.44%)
Migrate within or across provinces = 1 (18.56%)
Hukou Urban household registration = 0 (47.25%)
Rural household registration = 1 (52.75%)
Only child Non-only child = 0 (55.73%); Only child = 1 (44.27%)
Academic A continuous variable in the form of scores, with higher values indicating
performance better parental judgment on children’s academic performance.
Mean = 3.08, SD = 1.05, range = [1, 5]
Parental education A continuous variable reflecting the highest level of parental education (years).
Mean = 10.89, SD = 3.02, range = [0, 19]
Family economic Poor (reference group) =1 (21.09%)
status Medium = 2 (73.15%)
Rich = 3 (5.76%)
Father’s job Government or business leader (reference group) = 1 (12.50%)
Professional worker = 2 (6.16%)
Ordinary worker = 3 (41.64%),
Agricultural worker = 4 (17.22%)
Self-employed = 5 (19.39%)
Unemployed or retired = 6 (3.09%)
Mother’s job Government or business leader (reference group) = 1 (6.54%)
Professional worker = 2 (8.09%)
Ordinary worker = 3 (36.49%)
Agricultural worker = 4 (19.73%)
Self-employed = 5 (16.68%)
Unemployed or retired = 6 (12.46%)
Parental A continuous variable reflecting the maximum number of years of education
educational parents want their children to achieve.
expectations Mean = 17.40, SD = 3.65, range = [8, 23]
Parental A continuous variable, with higher values indicating higher levels of parental
monitoring monitoring of children’s studies. Its coding is based on the student report
about how parents monitor their homework and exams, school
performance, peer interaction, dress up, time spent online, and time spent
on TV (no monitoring = 1, moderate monitoring = 2, strict monitoring = 3).
Based on the report, factor analysis is adopted to construct this composite
variable (α = .71).
Mean = −0.01, SD = 1.00, range = [−3.48, 3.10]

(Continued)
 Chinese Sociological Review 323

Table 1. Continued.
Variable codings and distributions
Classroom climate A continuous variable in the form of scores, with higher values indicating a
better academic climate perceived by students.
Mean = 3.14, SD = 0.89, range = [1, 4]
School-level
covariates
School district Central or western area = 0 (44.09%)
Eastern area = 1 (55.91%)
School location Suburban or rural school = 0 (60.28%)
Urban school = 1 (39.72%)
School ranking Middle or lower rank (reference group) = 1 (18.75%)
Upper middle rank = 2 (60.16%)
Highest rank = 3 (21.09%)
School SES A composite categorical variable summed by the average education level,
average income level, and occupational status of parents from the school.
Low (reference group) = 1 (29.14%)
Medium = 2 (56.50%)
High = 3 (14.37%)
School’s The school average of teachers’ perceived performance pressure, a continuous
performance variable in the form of scores, with higher values indicating higher pressure
pressure imposed on teachers by the schools about students’ high school admission
rates.
Mean = 3.94, SD = 0.69, range = [2, 5]
Note: Given the chronological order, the outcome variable is collected from the Wave 2 survey, and
the key independent variable and most control variables are collected from the Wave 1 survey.

statistically significant likelihood ratio test statistic justifies employing

multilevel analysis.
Model 1 examines the bivariate relationship between key independent
and outcome variables in this study. Without other control variables,
teacher-parent communications exhibit statistically significant positive
effects on families’ private tutoring investment (coef = 0.245, p < 0.001).
The estimate of between-school variance is 0.366. By comparison, this
estimate in the empty model is 0.428, suggesting that teacher-parent com-
munications reduce the between-school variance of private tutoring par-
ticipation by 14.49 percent.
Model 2 includes the individual- and school-level control variables to
obtain the net effect of teacher-parent communications. All other socio-
demographic factors being equal, teacher-parent communications could
still positively impact families’ involvement in shadow education (coef =
0.153, p < 0.001). By converting this coefficient to incidence-rate ratios
(IRR = 1.165), we find that teachers’ communications with parents are
associated with a 16.5% increase in the likelihood of parental invest-
ment in private tutoring, substantiating Hypothesis 1 in this study.
Additionally, the estimate of between-school variance shows that teach-
er-parent communications and other background factors account for
71.03 percent of between-school variance in families’ private tutoring
involvement.
324 Z. PAN

Table 2. Multilevel Poisson regression estimating families’ participation in private

tutoring.
Empty model M1 M2 M3
Teacher-parent 0.245 (0.058)*** 0.153 (0.037)*** −0.277 (0.214)
communications
Eastern area school 0.182 (0.076)* 0.182 (0.074)*
(non-eastern
school = 0)
Urban school 0.420 (0.084)*** 0.436 (0.083)***
(non-urban
school = 0)
School rank
Upper middle −0.025 (0.099) −0.025 (0.097)
Highest 0.003 (0.130) 0.002 (0.127)
School SES
Medium 0.032 (0.087) 0.059 (0.086)
High 0.130 (0.135) 0.161 (0.133)
School’s 0.038 (0.053) 0.021 (0.053)
performance
pressure
Female child (male 0.155 (0.022)*** 0.155 (0.022)***
child = 0)
Only child 0.044 (0.027) 0.044 (0.027)
(non-only
child = 0)
Migrant child −0.107 (0.032)** −0.106 (0.032)**
(non-migrant
child = 0)
Rural child −0.110 (0.029)*** −0.110 (0.029)***
(non-rural
hukou = 0)
Parental education 0.036 (0.005)*** 0.035 (0.005)***
Family economic
level
Medium 0.094 (0.036)** 0.095 (0.036)**
Rich 0.144 (0.053)** 0.145 (0.053)**
Father’s job
Professional workers −0.090 (0.042)* −0.090 (0.042)*
Ordinary workers −0.208 (0.035)*** −0.209 (0.035)***
Agricultural workers −0.318 (0.063)*** −0.319 (0.063)***
Self-employed 0.006 (0.039) 0.006 (0.039)
Retired or −0.191 (0.071)** −0.191 (0.071)**
unemployed
Mother’s job
Professional workers 0.053 (0.043) 0.052 (0.043)
Ordinary workers −0.122 (0.041)** −0.121 (0.041)**
Agricultural workers −0.179 (0.065)** −0.181 (0.065)**
Self-employed −0.056 (0.047) −0.057 (0.047)
Retired or −0.175 (0.048)*** −0.176 (0.048)***
unemployed
Academic 0.035 (0.011)** 0.034 (0.011)**
performance
Parental educational 0.017 (0.003)*** 0.017 (0.003)***
expectations
Parental monitoring 0.113 (0.011)*** 0.113 (0.011)***
Perceived classroom 0.036 (0.013)** 0.036 (0.013)**
climate
TPC × SPP 0.107 (0.053)*
Intercept −0.241 (0.064)*** −0.251 (0.059)*** −1.444 (0.248)*** −1.405 (0.245)***

(Continued)
 Chinese Sociological Review 325

Table 2. Continued.
Empty model M1 M2 M3
Between-school 0.428 (0.062) 0.366 (0.053) 0.124 (0.020) 0.119 (0.020)
variance
AIC 23,695.72 23,681.04 22,927.60 22,925.53
BIC 23,709.97 23,702.41 23,148.35 23,153.41
LR test vs. Poisson 2,710.06*** 2,449.51*** 568.23*** 538.97***
model:
chibar2(01)
Notes: 1. The reference groups for “family economic level,” “father’s/mother’s job,” “school ranking,” and
“school SES” are “poor,” “government or business leader,” “school ranking medium and below,” and
“low SES,” respectively. 2. TPC = Teacher-Parent Communications, SPP = School’s Performance Pressure.
3. +p < 0.1, *p < 0.05, **p < 0.01, ***p < 0.001. Coef (Standard error).

In Model 2, the coefficients for the covariates provide a comprehensive

overview of other influences on private tutoring. At the school level, stu-
dents in eastern-region or urban schools exhibit a higher propensity to
participate in private tutoring. At the individual level, female students,
those with urban hukou, and non-migrant students are more inclined to
engage in private tutoring. Socioeconomic factors, including parental edu-
cation, income level, occupational status, and educational expectations,
positively predict investment in shadow education, supporting the rational
choice perspective. Children’s performance, perceived classroom climate,
and parental monitoring also positively correlate with tutoring participa-
tion, aligning with the social norm perspective.
In Model 3, an interaction term between teacher-parent communica-
tion and schools’ performance pressure is included to investigate the
mechanisms of the school process further. This interaction term has a
significantly positive effect on private tutoring involvement (coef = 0.107,
p < 0.05), supporting Hypothesis 2 that the positive influence of teach-
er-parent communications on families’ tutoring participation is more pro-
nounced when schools prioritize students’ high school admission rates
and impose heightened relevant pressure on teachers. The result suggests
that propelled by the assessment pressure to perform, Chinese schools
and teachers would likely encourage parents to utilize various additional
educational resources to enhance students’ academic performance.
To present the result of Model 3 more clearly, the author calculates the
marginal predicted mean of private tutoring participation based on teach-
er-parent communications and schools’ performance pressure. Low and
high levels of teacher-parent communication are defined as the 10th and
90th percentile of the teacher-parent communication scale. As shown in
Figure 1, for schools with relatively low levels of performance pressure
(pressure scores equal the 10th and 25th percentile), we observe no statis-
tically significant difference between the effects of low and high levels of
teacher-parent communication on tutoring participation. Conversely, when
326 Z. PAN

Figure 1. The marginal predicted mean of private tutoring participation by school’s

performance pressure and teacher-parent communications (with 95% CI). Note:
Predicted mean are calculated from Model 3 of Table 2, with all control variables held
at their sample mean values for calculations. Pressure scores equal 3, 3.5, 4, 4.5, and
5 signify the 10th, 25th, 50th, 75th, and 90th percentile of the pressure score scale,
respectively.

schools impose heightened pressure on teachers (pressure scores equal the

50th, 75th, and 90th percentile), a high level of teacher-parent communi-
cation significantly increases the likelihood of tutoring participation. This
positive effect also expands as the pressure increases. These results indicate
that school performance pressure moderates the effects of school-initiated
interactions with parents.

Teacher-parent communications moderate the family income effect

Table 3 shows how teacher-parent communications moderate the effect of
family economic backgrounds on private tutoring involvement. Model 4
examines the effect of family income on tutoring participation without
considering teacher-parent communications but with other sociodemo-
graphic factors under control. The result suggests that low-income fami-
lies are less likely to invest in private tutoring. Model 5 adds the
teacher-parent communication variable based on Model 4, showing a
reduction in the effect of low-income families.
In Model 6, the author investigates the effect heterogeneity of family
backgrounds across schools by estimating a random coefficient model and
 Chinese Sociological Review 327

Table 3. Effect heterogeneity of teacher-parent communications on families’ private

tutoring participation across family economic backgrounds.
M4 M5 M6 M7
Individual-level variable
Low-income family −0.099 (0.036)** −0.096 (0.036)** −0.149 (0.049)** −0.149 (0.049)**
School-level variable
Teacher-parent 0.153 (0.037)*** 0.139 (0.037)*** 0.131 (0.036)***
communications
Cross-level interaction
effect
LIF × TPC 0.098 (0.045)*
Intercept −1.486 (0.260)*** −1.352 (0.247)*** −1.379 (0.245)*** −1.371 (0.244)***
Between-school 0.146 (0.024) 0.124 (0.020) 0.116 (0.020) 0.115 (0.020)
variance
LIF slope variance 0.073 (0.032) 0.067 (0.030)
Covariance (intercept, 0.009 (0.018) 0.011 (0.017)
slope)
Notes: 1. LIF = Low-Income Family; TPC = Teacher-Parent Communications. 2. Models 4–7 have con-
trolled for other individual and school-level factors. 3. *p < 0.05, **p < 0.01, ***p < 0.001.

including random intercepts and random slopes for family income. With
individual- and school-level factors controlled, the estimated variance of
the low-income family slopes is 0.073, with a standard error of 0.032.
This result indicates that the effects of family income on private tutoring
participation vary significantly across various school contexts. The 95 per-
cent plausible value range for the low-income family slopes is
( )
−0.149 ± 1.96 × 0.073 = ( −0.679, 0.381), suggesting that low-income fam-
ilies no longer negatively affect tutoring involvement in some schools.
In Model 7, the author further delves into how the school process con-
tributes to variations in the effect of family background by introducing a
cross-level interaction term between low-income family and teacher-par-
ent communication. This interaction term is significant and positive, sug-
gesting that the effects of family income on private tutoring participation
are conditional upon levels of teacher-parent communications. School-
initiated communications might mitigate the negative effect of low-in-
come families, consistent with Hypothesis 4 in this study. With the
cross-level interaction effect taken into account, the residual variance of
the low-income family slopes is 0.067, which, compared to the slope vari-
ance in Model 6 of 0.073, implies a reduction of 8.22 percent. Therefore,
teacher-parent communications help explain 8.22 percent of the variation
in the effects of family income across schools.
The cross-level interaction term also signifies that teacher-parent com-
munications exhibit heterogeneous effects between low-income and
upper-middle-income families. Table 4 illustrates this finding by calculat-
ing the marginal predicted mean of private tutoring participation at dif-
ferent levels of teacher-parent communication and family income. Holding
all other covariates at their sample means, an increase in the level of
328 Z. PAN

Table 4. The marginal predicted mean of private tutoring participation by family

income and teacher-parent communications.
Low-income family Upper-middle-income family
Level of teacher-parent communications
The 10th percentile 0.547 0.731
The 25th percentile 0.647 0.804
The 50th percentile 0.748 0.874
The 75th percentile 0.891 0.966
The 90th percentile 0.980 1.020
Note: The predicted mean is calculated from Model 7 of Table 3, with all control variables held at
their sample means for calculations.

teacher-parent communication from the 10th to the 90th percentile results

in a 39.53% rise in the marginal mean predicted for upper-middle-
income families, shifting from 0.731 to 1.020. In contrast, the predicted
marginal mean for low-income families increases by 79.16% from 0.547
to 0.980, suggesting a more pronounced positive effect of teacher-parent
communications on tutoring participation for low-income families. This
finding indicates that the school process can bolster active educational
behaviors for disadvantaged families, potentially reducing educational
inequality.

Robustness check
This section deals with the potential endogenous biases in this research,
starting with reverse causality. In theory, individual behavior outside of
school would not directly impact the management system within the
school. If such an impact were to occur, the possible way is that the
increased rates of private tutoring participation would improve students’
average performance, consequently reducing schools’ demand for seeking
interactions and support from families. This scenario does not correspond
to the positive correlation predicted in this study. Additionally, in the
empirical model, the author adopts a lag period for the outcome variable
to address this concern.
The other possible bias is whether a third variable, family background,
influences the correlation. Students from advantaged families are more
likely to participate in private tutoring, and concurrently, they might enter
a better school that highlights teacher-parent communications. Given this
scenario, the author takes several steps to address this concern. First, the
models have controlled for parental education, occupation, and income
level to examine the net effect of teacher-parent communications. Second,
the cross-level interaction effect analysis suggests that the positive effect
of teacher-parent communications is more pronounced in low-income
families, implying that the mechanisms and effects of family background
differ from those of the school process. Third, the author further
 Chinese Sociological Review 329

Table 5. Multilevel Poisson regression estimating families’ participation in private

tutoring by sub-samples.
Low SES school Medium SES school High SES school
Teacher-parent 0.216 (0.076)** 0.145 (0.047)** 0.019 (0.108)
communications
Between-school variance 0.113 (0.037) 0.123 (0.027) 0.094 (0.040)
N 2,665 5,168 1,314
Notes: 1. All other individual and school-level factors have been controlled. 2. **p < 0.01. Coef
(Standard error).

examines whether school SES primarily explains the effect of teacher-par-

ent communications. Table 5 presents the multilevel regression results
estimated respectively based on the sub-samples of low, medium, and
high SES schools. We see that teacher-parent communications have stron-
ger positive effects in low SES schools than in medium SES schools while
showing no significant impacts in high SES schools. All these results sug-
gest that the effect of teacher-parent communications does not stem from
the aggregation of advantaged families. The communication process itself,
associated with schools’ management systems, teachers’ practices, and
organizational climate, can significantly shape parental educational atti-
tudes and behaviors.
Another consideration is whether past tutoring experiences contribute
to the observed correlation. It sounds possible that students involved in
private tutoring previously could outperform in the entrance exam to
enter a better school that values teacher-parent communications, and they
would maintain tutoring later on. However, there is not necessarily a pos-
itive association between students’ enrollment performance and the
school’s active arrangement of teacher-parent communications. Instead, it
may be more imperative for schools with a composition of low-achieving
students to increase communications with families and garner parental
support. Furthermore, the CEPS Wave1 principal questionnaire shows
that in our samples, 90.37% of students (N = 8,266) entered the current
school without passing an entrance exam, suggesting that students’ past
tutoring may not be associated with the characteristics of their current
schools.

Discussion and conclusion

Given the rapid expansion of shadow education in the contemporary
world, empirical research has documented that household resources and
social interactions among families are associated with shadow education
investment (Entrich 2017; Lee and Shouse 2011; Schaub 2010). While
prior studies primarily focus on family-level explanations, the school-level
factor is crucial in understanding this phenomenon. Using data from the
330 Z. PAN

China Education Panel Survey (CEPS), this study investigated how teach-
er-parent communications affect families’ private tutoring involvement
and obtained three main findings.
First, with covariates of family and school characteristics included, the
result from the multilevel model suggests that teacher-parent communica-
tions significantly increase the likelihood of private tutoring involvement,
consistent with prior studies demonstrating the significant relationship
between Chinese shadow education investment and teacher-parent inter-
action chances net of social origin factors (Zhou and Wu 2018). This
result implies that cultural and psychological factors are pivotal in paren-
tal educational involvement. Intensive parenting behavior in China is not
solely influenced by the family’s possession of elite class capital but instead
cultivated and reinforced in the schooling process.
Second, the school-level interaction effect analysis indicates that the
impact of teacher-parent communications is more pronounced when teach-
ers perceive higher performance pressure from their schools. This result
underscores the impact of the school management system on teachers’ goals
and behaviors during their interactions with parents. In China, confronted
with educational assessments initiated by superior educational departments,
many middle schools strive to improve students’ test scores and admission
rates to high school, treating student performance as an essential indicator
to evaluate teachers and determine teachers’ welfare (Zhao 2015). This eval-
uation system motivates teachers to increase communications with parents
and arouse parental concern about children’s studies. Furthermore, when
facing heightened performance pressure, teachers are more inclined to
extend their teaching responsibilities and pressure to families by encourag-
ing investment in extracurricular educational resources.
Influenced by teachers, Chinese families’ involvement in private tutor-
ing transcends the rational pursuit of quality educational resources;
instead, it is propelled by the norms and climate constructed within the
school. In this scenario, private tutoring involvement is not a secret edu-
cational alternative that parallels the mainstream school, nor is it a
“shadow” hidden behind the school. Instead, it emerges as an open and
unwritten rule encouraged and legitimized by formal schooling (W. Zhang
and Bray 2015). Looking further, when schools highlight a competitive
academic climate and shape parental attitudes towards extracurricular
education, part of the school’s duty has subtly shifted to the family and
the private tutoring market. Private tutoring participation in China can
thus be viewed as an extension of the school process.
Third, the cross-level interaction effect analysis further reveals the
heterogeneous effect of teacher-parent communications on tutoring par-
ticipation across different family economic backgrounds. The result sug-
gests that the positive effects of school-initiated communications are
 Chinese Sociological Review 331

more potent in low-income families, implying that the positive parenting

attitudes and elite cultural habitus initially held by advantaged families
are potentially transmitted to disadvantaged families through the school
communication process. This way, teacher-parent communications
encourage low-income parents to follow the social norms of active edu-
cational involvement and investment, thereby mitigating the initially neg-
ative effect of disadvantaged families on private tutoring participation.
Overall, this study provides insight into social stratification studies by
examining the role of the school process. Prior studies have highlighted
a connection between students’ educational attainment and schools’ cul-
tural climate and social interactions (Allen et al. 2013; Ryan and Patrick
2001; Nelson 1972). Following this strand of studies, this research iden-
tifies that school plays a complicated role in social stratification. On the
one hand, teacher-parent communications exhibit a positive general
effect on private tutoring investment, denoting a potential impact in
widening educational inequality. On the other hand, this effect is more
pronounced for low-income families, implying that the school process
might encourage the disadvantaged strata to engage in educational com-
petition and pursue social mobility. Nevertheless, given that the advan-
taged strata have greater access to high-quality tutoring resources, the
positive effect of teacher-parent communications on private tutoring is
more likely to improve advantaged children’s performance, ultimately
promoting social structure reproduction. In this scenario, policy should
provide better guidance to schools and parents to further promote edu-
cational equality.

References
Allen, Joseph, Anne Gregory, Amori Mikami, Janetta Lun, Bridget Hamre, and
Robert Pianta. 2013. “Observations of Effective Teacher-Student Interactions in
Secondary School Classrooms: Predicting Student Achievement with the
Classroom Assessment Scoring System-Secondary.” School Psychology Review 42
(1): 76–98. https://doi.org/10.1080/02796015.2013.12087492
Addi-Raccah, Audrey, and Oshra Dana. 2015. “Private Tutoring Intensity in
Schools: A Comparison between High and Low Socio-Economic Schools.”
International Studies in Sociology of Education 25 (3): 183–203. https://doi.org/
10.1080/09620214.2015.1069719
Hawrot, Anna. 2022. “Do School-Related Factors Affect Private Tutoring
Attendance? Predictors of Private Tutoring in Maths and German among
German Tenth-Graders.” Research Papers in Education 39 (1): 1–23. https://doi.
org/10.1080/02671522.2022.2089209
Baker, David P. 2009. “The Educational Transformation of Work: Towards a New
Synthesis.” Journal of Education and Work 22 (3): 163–91. https://doi.org/10.1080/
13639080902957822
332 Z. PAN

Baker, David, and GeraldK. LeTendre. 2005. National Difference, Global Similarities:
World Culture and the Future of Schooling. Palo Alto, CA: Stanford.
Barg, Katherin. 2019. “Why Are Middle-Class Parents More Involved in School
than Working-Class Parents?” Research in Social Stratification and Mobility
59:14–24. https://doi.org/10.1016/j.rssm.2018.12.002
Bianchi, Suzanne M, P. John, and MelissaA. Robinson. 2006. The Changing
Rhythms of American Family Life. New York: Russell Sage Foundation.
Bidwell, Charles E. 2013. “The School as a Formal Organization.” In Handbook of
Organizations (RLE: Organizations), 972–1022. Boca Raton, FL: Routledge.
Bodovski, Katerina. 2010. “Parental Practices and Educational Achievement: Social
Class, Race, and Habitus.” British Journal of Sociology of Education 31 (2): 139–
56. https://doi.org/10.1080/01425690903539024
Boudon, Raymond. 1974. Education, Opportunity, and Social Inequality: Changing
Prospects in Western Society. New York: Wiley.
Bourdieu, Pierre. 1973. Cultural Reproduction and Social Reproduction. London:
Tavistock.
Bowles, Samuel, and Herbert Gintis. 1976. Schooling in Capitalist America. New
York: Basic.
Bray, Mark. 1999. “The Shadow Education System: Private Tutoring and Its
Implications for Planners.” In No.61 Fundamentals of Educational Planning,
Paris: UNESCO International Institute for Educational Planning (IIEP).
Bray, Mark. 2011. The Challenge of Shadow Education: Private Tutoring and Its
Implications for Policymakers in the European Union.” Luxembourg: European
Commission.
Bray, Mark. 2021. “Shadow Education in Europe: Growing Prevalence, Underlying
Forces, and Policy Implications.” ECNU Review of Education 4 (3): 442–75.
https://doi.org/10.1177/2096531119890142
Bray, Mark, and Chad Lykins. 2012. Shadow Education: Private Supplementary
Tutoring and Its Implications for Policymakers in Asia. No. 9. Mandaluyong:
Asian Development Bank. https://doi.org/10.1007/s11159-012-9324-2
Buchmann, Claudia, Dennis J. Condron, and Vincent J. Roscigno. 2010. “Shadow
Education, American Style: Test Preparation, the SAT and College Enrollment.”
Social Forces 89 (2): 435–61. https://doi.org/10.1353/sof.2010.0105
Byun, S., H. J. Chung, and D. P. Baker. 2018. “Global Patterns of the Use of
Shadow Education: Student, Family, and National Influences.” In Research in
the Sociology of Education, edited by Hyunjoon Park and Grace Kao. Bingley:
Emerald Publishing Limited.
Carolan, Brian V., and Sara J. Wasserman. 2015. “Does Parenting Style Matter?
Concerted Cultivation, Educational Expectations, and the Transmission of
Educational Advantage.” Sociological Perspectives 58 (2): 168–86. https://doi.
org/10.1177/0731121414562967
Chan, Tak, Cheung, Zhiding Shu, and Hong Ying Xiao. 2021. “Perception of
Chinese Parents toward School and Family Collaboration.” School Community
Journal 31 (1): 233–58.
Choi, Yool, and Hyunjoon Park. 2016. “Shadow Education and Educational
Inequality in South Korea: Examining Effect Heterogeneity of Shadow Education
on Middle School Seniors’ Achievement Test Scores.” Research in Social
Stratification and Mobility 44:22–32. https://doi.org/10.1016/j.rssm.2016.01.002
Chua, Amy. 2011. Battle Hymn of the Tiger Mother. New York: Penguin Press.
 Chinese Sociological Review 333

Davies, Scott. 2024. “Cultural Capital—Field Connections for Three Populations

of Chinese Students: A Theoretical Framework for Empirical Research.” Chinese
Sociological Review 56 (1):1–24. https://doi.org/10.1080/21620555.2024.2307480
Day, Elizabeth, and Aryn M. Dotterer. 2018. “Parental Involvement and Adolescent
Academic Outcomes: Exploring Differences in Beneficial Strategies across
Racial/Ethnic Groups.” Journal of Youth and Adolescence 47 (6): 1332–49.
https://doi.org/10.1007/s10964-018-0853-2
de Castro, Belinda V., and Allan B. de Guzman. 2010. “Push and Pull Factors
Affecting Filipino Students’ Shadow Education (SE) Participation.” KEDI
Journal of Educational Policy 7:43–66. https://doi.org/10.22804/KJEP.2010.7.1.003
Deloitte. 2018. “New Era of Education: Report on Chinese Education Development
in 2018.” https://www2.deloitte.com/cn/zh/pages/technology-media-and-
telecommunications/articles/new-era-of-education-china-education-
development-report-2018.html. (in Chinese).
Domina, Thurston. 2005. “Leveling the Home Advantage: Assessing the Effectiveness
of Parental Involvement in Elementary School.” Sociology of Education 78 (3):
233–49. https://doi.org/10.1177/003804070507800303
Entrich, SteveR. 2017. Shadow Education and Social Inequalities in Japan: Evolving
Patterns and Conceptual Implications. Cham: Springer.
Entrich, Steve R., and Wolfgang Lauterbach. 2019. “Shadow Education in
Germany: Compensatory or Status Attainment Strategy? Findings from the
German Life Study.” International Journal for Research on Extended Education
7 (2-2019): 143–59. https://doi.org/10.3224/ijree.v7i2.04
Entrich, Steve R. 2020. “Worldwide Shadow Education and Social Inequality:
Explaining Differences in the Socioeconomic Gap in Access to Shadow
Education across 63 Societies.” International Journal of Comparative Sociology
61 (6): 441–75. https://doi.org/10.1177/0020715220987861
Epstein, JoyceL. 2019. “Theory to Practice: School and Family Partnerships Lead
to School Improvement and Student Success.” In School, Family, and Community
Interaction, 39–52. Boca Raton, FL: Routledge.
Fingerman, Karen L., Yen-Pi Cheng, Eric D. Wesselmann, Steven Zarit, Frank
Furstenberg, and Kira S. Birditt. 2012. “Helicopter Parents and Landing Pad
Kids: Intense Parental Support of Grown Children.” Journal of Marriage and
the Family 74 (4): 880–96. https://doi.org/10.1111/j.1741-3737.2012.00987.x
Guo, Yan, Xueqin Wu, and Xiaoli Liu. 2019. “Challenges and Opportunities in
Parent-Teacher Relationships in Contemporary China.” Comparative and
International Education 47 (2): 5. https://doi.org/10.5206/cie-eci.v47i2.9331
Heo, Nayoung. 2022. “‘She Has the Japanese Style’: Parenting by Japanese Immigrant
Women in Korea from the Perspective of Their Children.” International Journal
of Environmental Research and Public Health 19 (3): 1494. https://doi.org/10.3390/
ijerph19031494
Hornby, Garry, and Rayleen Lafaele. 2011. “Barriers to Parental Involvement in
Education: An Explanatory Model.” Educational Review 63 (1): 37–52. https://
doi.org/10.1080/00131911.2010.488049
Hoy, WayneK, and CecilG. Miskel. 2013. Educational Administration: Theory,
Research, and Practice. 9th ed. New York: McGraw-Hill.
Ireson, Judith. 2004. “Private Tutoring: How Prevalent and Effective is It?” London
Review of Education 2 (2): 109. https://doi.org/10.1080/1474846042000229458
334 Z. PAN

Ishizuka, Patrick. 2019. “Social Class, Gender, and Contemporary Parenting

Standards in the United States: Evidence from a National Survey Experiment.”
Social Forces 98 (1): 31–58. https://doi.org/10.1093/sf/soy107
Islam, Asad. 2019. “Parent-Teacher Meetings and Student Outcomes: Evidence
from a Developing Country.” European Economic Review 111:273–304. https://
doi.org/10.1016/j.euroecorev.2018.09.008
Jansen, D., L. Elffers, and S. Jak. 2021. “A Cross-National Exploration of Shadow
Education Use by High and Low SES Families.” International Studies in Sociology
of Education 32 (3): 653–74. https://doi.org/10.1080/09620214.2021.1880332
Kim, Edward, Joshua Goodman, and MartinR. West. 2021. “Kumon in: The
Recent, Rapid Rise of Private Tutoring Centers.” Brown University Annenberg
EdWorkingPaper 21–367.
Kim, Ji-Ha, and Daekwon Park. 2010. “The Determinants of Demand for Private
Tutoring in South Korea.” Asia Pacific Education Review 11 (3): 411–21. https://
doi.org/10.1007/s12564-009-9067-3
Kim, Taehoon, Hayun Jang, and Jinho Kim. 2022. “Do Peers Affect Private
Tutoring Engagement in Korea? Evidence from a Quasi-Experimental Approach.”
Asia Pacific Education Review 23 (2): 271–83. https://doi.org/10.1007/s12564-
021-09738-1
Kim, Taejong. 2004. “Shadow Education: School Quality and Demand for Private
Tutoring in Korea.” KDI School of Public Policy and Management Paper. https://
doi.org/10.2139/ssrn.635864
Kim, Young Chun, and Jung-Hoon Jung. 2019. “Worldwide Shadow Education
Epidemic: From East Asia to Western Hemisphere.” In Shadow Education as
Worldwide Curriculum Studies, 25–60. Palgrave Macmillan, Cham.
Kobakhidze, Magda Nutsa. 2014. “Corruption Risks of Private Tutoring: Case of
Georgia.” Asia Pacific Journal of Education 34 (4): 455–75. https://doi.org/10.10
80/02188791.2014.963506
Kuan, Ping-Yin. 2018. “Effects of Cram Schooling on Academic Achievement and
Mental Health of Junior High Students in Taiwan.” Chinese Sociological Review
50 (4): 391–422. https://doi.org/10.1080/21620555.2018.1526069
Kumar, Indal, and Indrani Roy Chowdhury. 2021. “Shadow Education in India:
Participation and Socioeconomic Determinants.” Journal of South Asian
Development 16 (2): 244–72. https://doi.org/10.1177/09731741211032472
Lareau, Annette. 2011. Unequal Childhoods: Class, Race, and Family Life. Berkley
and Los Angeles, CA: University of California Press.
Lee, Soojeong, and Roger C. Shouse. 2011. “The Impact of Prestige Orientation
on Shadow Education in South Korea.” Sociology of Education 84 (3): 212–24.
https://doi.org/10.1177/0038040711411278
Lin, Wei, Hongbiao Yin, Jiwei Han, and Jiying Han. 2020. “Teacher-Student
Interaction and Chinese Students’ Mathematics Learning Outcomes: The
Mediation of Mathematics Achievement Emotions.” International Journal of
Environmental Research and Public Health 17 (13): 4742. https://doi.org/10.3390/
ijerph17134742
Liu, Shujie, and Anthony J. Onwuegbuzie. 2012. “Chinese Teachers’ Work Stress
and Their Turnover Intention.” International Journal of Educational Research
53:160–70. https://doi.org/10.1016/j.ijer.2012.03.006
Liu, Yixiao, Quanbao Jiang, and Feinian Chen. 2020. “Children’s Gender and
Parental Educational Strategies in Rural and Urban China: The Moderating
 Chinese Sociological Review 335

Roles of Sibship Size and Family Resources.” Chinese Sociological Review 52 (3):
239–68. https://doi.org/10.1080/21620555.2020.1717942
Lucas, Samuel R., and Delma Byrne. 2017. “Effectively Maintained Inequality in
Education: An Introduction.” American Behavioral Scientist 61 (1): 3–7. https://
doi.org/10.1177/0002764216682992
McLanahan, Sara, and Christine Percheski. 2008. “Family Structure and the
Reproduction of Inequalities.” Annual Review of Sociology 34 (1): 257–76.
https://doi.org/10.1146/annurev.soc.34.040507.134549
Meyer, John W. 1977. “The Effect of Education as an Institution.” American
Journal of Sociology 83 (1): 55–77. https://doi.org/10.1086/226506
Milkie, Melissa A., and Catharine H. Warner. 2014. “Status Safeguarding: Mothers’
Work to Secure Children’s Place in the Social Hierarchy.” In Intensive Mothering:
The Cultural Contradictions of Modern Motherhood, edited by Linda Rose
Ennis, 66–85. Ontario: Demeter Press.
Mori, Izumi, and David Baker. 2011. “The Origin of Universal Shadow Education:
What the Supplemental Education Phenomenon Tells Us about the Postmodern
Institution of Education.” Asia Pacific Education Review 11 (1): 36–48. https://
doi.org/10.1007/s12564-009-9057-5
Nelson, Joel I. 1972. “High School Context and College Plans: The Impact of
Social Structure on Aspirations.” American Sociological Review 37 (2): 143–8.
https://doi.org/10.2307/2094022
Pan, Zheng, Donald Lien, and Hao Wang. 2022. “Peer Effects and Shadow
Education.” Economic Modelling 111:105822. https://doi.org/10.1016/j.econmod.
2022.105822
Park, Hyunjoon, Claudia Buchmann, Jaesung Choi, and Joseph J. Merry. 2016.
“Learning beyond the School Walls: Trends and Implications.” Annual Review
of Sociology 42 (1): 231–52. https://doi.org/10.1146/annurev-soc-081715-074341
Park, Hyunjoon, Soo-yong Byun, and Kyung-keun Kim. 2011. “Parental
Involvement and Students’ Cognitive Outcomes in Korea: Focusing on
Private Tutoring.” Sociology of Education 84 (1): 3–22. https://doi.org/10.1177/
0038040710392719
Pillet-Shore, Danielle. 2015. “Being a “Good Parent” in Parent–Teacher Conferences.”
Journal of Communication 65 (2): 373–95. https://doi.org/10.1111/jcom.12146
Redding, Sam, Janis Langdon, Joseph Meyer, and Pamela Sheley. 2004. “The Effects
of Comprehensive Parent Engagement on Student Learning Outcomes.” American
Educational 14 (4): 3–7. https://www.researchgate.net/publication/228862484
Resh, Nura, and Yechezkel Dar. 2012. “The Rise and Fall of School Integration in
Israel: Research and Policy Analysis.” British Educational Research Journal 38
(6): 929–51. https://doi.org/10.1080/01411926.2011.603034
Ryan, Allison M., and Helen Patrick. 2001. “The Classroom Social Environment
and Changes in Adolescents’ Motivations and Engagement during Middle
School.” American Educational Research Journal 38 (2): 437–60. https://doi.
org/10.3102/00028312038002437
Santiago, Rachel T., S. Andrew Garbacz, Tiffany Beattie, and Christabelle L.
Moore. 2016. “Parent-Teacher Relationships in Elementary School: An
Examination of Parent-Teacher Trust.” Psychology in the Schools 53 (10): 1003–
17. https://doi.org/10.1002/pits.21971
Schaub, Maryellen. 2010. “Parenting for Cognitive Development from 1950 to
2000: The Institutionalization of Mass Education and the Social Construction
336 Z. PAN

of Parenting in the United States.” Sociology of Education 83 (1): 46–66. https://

doi.org/10.1177/0038040709356566
Scott, W. Richard. 2013. Institutions and Organizations: Ideas, Interests, and Identities.
Thousand Oaks, CA: Sage Publications.
Shi, Jing, and Zexuan Yin. 2020. “Survey on the Educational Needs and Mentality
Characteristics of Chinese Parents under the New Situation.” Governance 277
(13): 11–7. (in Chinese).
Stevenson, David Lee, and David P. Baker. 1992. “Shadow Education and
Allocation in Formal Schooling: Transition to University in Japan.” American
Journal of Sociology 97 (6): 1639–57. https://doi.org/10.1086/229942
Vincent, Carol, and Stephen J. Ball. 2007. “Making Up’ the Middle-Class Child:
Families, Activities and Class Dispositions.” Sociology 41 (6): 1061–77. https://
doi.org/10.1177/0038038507082315
Walker, Allan, Haiyan Qian, and Shuang Zhang. 2011. “Secondary School
Principals in Curriculum Reform: Victims or Accomplices?” Frontiers of
Education in China 6 (3): 388–403. https://doi.org/10.1007/s11516-011-0138-y
Woo, Hongjoo, and Nancy N. Hodges. 2015. “Education Fever: Exploring Private
Education Consumption Motivations among Korean Parents of Preschool
Children.” Family and Consumer Sciences Research Journal 44 (2): 127–42. https://
doi.org/10.1111/fcsr.12131
Yang, Hyunwoo. 2017. “The Role of Social Capital at Home and in School in
Academic Achievement: The Case of South Korea.” Asia Pacific Education
Review 18 (3): 373–84. https://doi.org/10.1007/s12564-017-9492-7
Ye, Hua. 2015. “Key-Point Schools and Entry into Tertiary Education in China.”
Chinese Sociological Review 47 (2): 128–53. https://doi.org/10.1080/21620555.20
14.990321
You, Yongheng. 2007. “A Deep Reflection on the “Key School System” in Basic
Education in China.” Frontiers of Education in China 2 (2): 229–39. https://doi.
org/10.1007/s11516-007-0019-6
Yu, Hui. 2020. “The Making of ‘Incompetent Parents’: Intersectional Identity,
Habitus, and Chinese Rural Migrant’s Parental Educational Involvement.” The
Australian Educational Researcher 47 (4): 555–70. https://doi.org/10.1007/
s13384-019-00361-z
Yulianti, Kartika, Eddie Denessen, Mienke Droop, and Gert-Jan Veerman. 2022.
“School Efforts to Promote Parental Involvement: The Contributions of School
Leaders and Teachers.” Educational Studies 48 (1): 98–113. https://doi.org/10.10
80/03055698.2020.1740978
Zhang, Wei. 2014. “The Demand for Shadow Education in China: Mainstream
Teachers and Power Relations.” Asia Pacific Journal of Education 34 (4): 436–
54. https://doi.org/10.1080/02188791.2014.960798
Zhang, Wei, and Mark Bray. 2015. “Shadow Education in Chongqing, China:
Factors Underlying Demand and Policy Implications.” KEDI Journal of
Educational Policy 12 (1): 83–106.
Zhang, Wei, and Mark Bray. 2018. “Equalising Schooling, Unequalising Private
Supplementary Tutoring: Access and Tracking through Shadow Education in
China.” Oxford Review of Education 44 (2): 221–38. https://doi.org/10.1080/030
54985.2017.1389710
Zhang, Wei, and Mark Bray. 2020. “Comparative Research on Shadow Education:
Achievements, Challenges, and the Agenda Ahead.” European Journal of
Education 55 (3): 322–41. https://doi.org/10.1111/ejed.12413
 Chinese Sociological Review 337

Zhang, Wei, and Yoko Yamato. 2018. “Shadow Education in East Asia: Entrenched
but Evolving Private Supplementary Tutoring.” In Routledge International
Handbook of Schools and Schooling in Asia, edited by Kerry J. Kennedy, and
John Chi-Kin Lee, 323–32. Boca Raton, FL: Routledge.
Zhang, Yueyun, and Yu Xie. 2016. “Family Background, Private Tutoring, and
Children’s Educational Performance in Contemporary China.” Chinese Sociological
Review 48 (1): 64–82. https://doi.org/10.1080/21620555.2015.1096193
Zhao, Xu. 2015. Competition and Compassion in Chinese Secondary Education.
New York: Palgrave Macmillan.
Zhou, Dong-yang, and Yu-xiao Wu. 2018. “Educational Competition and Reference
Group: A Sociological Explanation of the Popularity of Private Tutoring in
China.” Journal of Nanjing Normal University (Social Science Edition) 5:84–97.
in Chinese).

About the author

Zhidi Pan is a Ph.D. Candidate at the Department of Sociology, Zhejiang
University, Hangzhou, China. Her research interests include the sociology of edu-
cation, social stratification, higher education, and youth studies.
Copyright of Chinese Sociological Review is the property of Taylor & Francis Ltd and its
content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder's express written permission. However, users may print, download, or email
articles for individual use.