Abstract
The recent increase in mental health problems and private tutoring (also called shadow education) among adolescents has been widely documented. This study relates involvement in private tutoring to depressive symptoms using data from a nationally representative sample of middle school students (grades 7–9, Mage = 13.90, N = 25,862) in China. Results show that greater involvement in tutoring is associated with increased depressive symptoms, and a decrease in nighttime sleep duration plays a mediating role. Grade-specific subsample analyses further show that the mediating role of sleep reduction becomes more pronounced for students in higher grades. These findings highlight the essential role of private tutoring for adolescent sleep and mental health and its variation across the schooling process.
Introduction
After several decades of stability (Jane Costello et al., 2006), adolescent mental health has deteriorated markedly in the last one to two decades (Bor et al., 2014; Cosma et al., 2020; Keyes et al., 2019; Marquez & Long, 2021; Mojtabai et al., 2016; Twenge et al., 2018). The prevalence of major depressive episodes among US adolescents increased from 8.7% to 12.5% during the period 2005 to 2015 (Mojtabai et al., 2016). Similar findings are also found in European countries (Bersia et al., 2022; Patalay & Gage, 2019; van Vuuren et al., 2018). The trend is particularly concerning given the well-established links between mental health problems during adolescence and multiple adverse outcomes, including increased risk of obesity (Merikangas et al., 2012; Roberts & Duong, 2013), mental health issues in adulthood (Al Mamun et al., 2007; Melvin et al., 2013), and suicide (Bridge et al., 2006; Weissman et al., 1999).
In parallel with the recent decline in mental health, a notable expansion of private tutoring practices among adolescent students has been widely documented. Private tutoring is an umbrella item that encompasses various out-of-school learning practices intended to help students catch up, keep up, or get ahead of their peers in educational performance (Bray et al., 2014). Also referred to as shadow education, it mimics the formal education sector by offering similar curricula, such as math, science, and English (Bray, 2009; Park et al., 2016). Private tutoring has long been embraced in East Asian societies such as Japan and South Korea (Park et al., 2011; Stevenson & Baker, 1992), but it is now also commonplace in Western societies (Bray, 2021; Buchmann et al., 2010). Indeed, data from the Programme for International Student Assessment (PISA) demonstrate a global upward trend in tutoring practices from 2003 to 2012 (Park et al., 2016). In 2012, more than 30% of 15-year-old students in Germany, Hong Kong, Latvia, Russia, Spain, Turkey, and Uruguay spent some time each week in private tutoring, and almost half of the students in Brazil, Greece, South Korea, and Thailand did so (Park et al., 2016).
Despite the documented increases in both mental health issues and private tutoring practices among adolescents in many societies, few systematic attempts have been made to link the two together, asking whether and how involvement in private tutoring can be related to adolescent mental health. Utilizing data from a nationally representative sample of middle school students in China, this study asks three questions: (1) whether greater involvement in private tutoring is associated with higher levels of depressive symptoms, (2) whether and to what extent the reduction in sleep time due to private tutoring may mediate the association, and (3) whether the pattern of association may vary with the advancement of grade levels. In doing so, we aim to associate private tutoring with sleep and mental health among Chinese adolescents in the middle school setting.
Previous research investigating the correlates of adolescent mental health has largely neglected the potential role of private tutoring, whereas previous research investigating the consequences of private tutoring has largely been limited to academic outcomes, with inadequate attention paid to non-academic aspects such as mental health. This study represents a contribution to both strands of research. In terms of policy relevance, private tutoring practices have been closely monitored and carefully regulated in China and elsewhere (Choi & Choi, 2016; Wang et al., 2022), due to implicit concerns that too much private tutoring may crowd out extracurricular and leisure activities, thereby undermining students’ health. Findings in this study may have important implications for societies that are witnessing the expansion of the private tutoring sector and are concerned about its impact on students’ health and wellbeing.
Literature Review
Private Tutoring and Mental Health
The correlates of mental health issues and the consequences of private tutoring practices among adolescents have been addressed separately in two broad strands of research. One line of research, mainly in psychology and public health, has paid much attention to the potential links of adolescent mental health with recently emerging risk factors, such as the overuse of digital devices (Twenge et al., 2018), the increase in schoolwork pressure (Cosma et al., 2020), and the experience of bullying victimization (Kessel Schneider et al., 2015), among others. In addition, a recent study calls for more attention to risk factors related to the country-specific social and educational circumstances in which adolescents grow up (Cosma et al., 2020). However, no study in this line of research, has considered the potential role of private tutoring practices for the psychological well-being of adolescents. The other line of research, mainly in education and sociology, has focused heavily on the impact of private tutoring on students’ academic performance (Park et al., 2011; Y. Zhang & Xie, 2016), cognitive ability (S. Hu & Mu, 2020; Y. Zhang, 2018), and educational transition (Buchmann et al., 2010; Stevenson & Baker, 1992). Unfortunately, the question of whether the consequences of private tutoring can spill over to mental health has been largely marginalized and narrowly confined to East Asian societies such as South Korea (Carr & Wang, 2018; Noh et al., 2020) and China (e.g., Kuan, 2018; Zheng et al., 2020).
Notably, concerns have recently been raised about private tutoring in Taiwan and mainland China for its potentially detrimental effects on students’ health behaviors and outcomes, but empirical findings have been mixed. In Taiwan, the average level of psychological well-being is lower among students who received tutoring than among those who did not (Chen & Lu, 2009; Kuan, 2018). However, in mainland China, participation in private tutoring is found to be associated with either lower, similar, or higher levels of emotional problems (Li & Liu, 2022; Sun et al., 2020; Zheng et al., 2020). Furthermore, going beyond a simple between-group comparison between the tutored and non-tutored students, more intensive involvement in private tutoring, either measured in terms of money or time spent, is significantly associated with more psychological symptoms (S. Hu & Mu, 2020; Q. Zhang & Gao, 2022; Zheng et al., 2020). As regards analytical samples, most of the previous studies have been limited to students in a single grade, and thereby, a systematic assessment of the psychological consequences of private tutoring at a certain educational level remains lacking.
How is it likely that private tutoring practices can influence adolescent mental health? Despite the lack of a direct discussion, a careful inspection would lead to two potential mechanisms alluded to in prior literature: a stress perspective and a displacement perspective. The stress perspective views private tutoring itself as a unique stressor that can lead to adverse mental health. Like too much homework, private tutoring represents a concrete component of academic burden and stress (Kuan, 2018). Many studies have confirmed that greater academic stress is associated with higher levels of depressive symptoms (Liu & Lu, 2012; Tang et al., 2020). It is, therefore, reasonable to expect that greater involvement in private tutoring would lead to greater psychological distress. From the displacement perspective, private tutoring practices would displace or crowd out other activities (e.g., sleep) that may benefit adolescent mental health (Kuan, 2018). In nature, private tutoring represents not only a human capital input but also a cost in terms of time and healthy activities. Anecdotal evidence has also suggested that private tutoring may help academically, but at the expense of leisure time and subjective well-being (Bray, 2013). It was partly on this basis that government regulations restricting private tutoring activities had been implemented (Morgan & Jan, 2022; Wang et al., 2022). Accordingly, our first hypothesis is:
H1: Greater involvement in private tutoring is associated with increased depressive symptoms.
The Mediating Role of Sleep Deprivation
The decline in sleep time due to private tutoring may serve as an important channel through which private tutoring is linked to adolescent mental health. Sufficient sleep plays a crucial role in the health and wellbeing of individuals. In adolescents in particular, longer sleep duration leads to better emotional regulation and higher quality of life (Chaput et al., 2016). Yet sleep deprivation is particularly likely to occur during the developmental stage of adolescence (Chaput et al., 2016), and time trend analyses have documented a significant decline in nighttime sleep duration in many societies over the past decades (Keyes et al., 2019; Twenge et al., 2018).
The change in sleep duration should be related to both mental health and private tutoring if it could serve as a mediating mechanism. On the one hand, many studies have linked sleep deprivation to adolescent mental health (Chaput et al., 2016; Roberts & Duong, 2014; J. Zhang et al., 2017). Using data from a nationally representative survey of over 10,000 US adolescents aged 13 to 18 years, one study shows that shorter weeknight sleep duration is associated with increased odds of a range of mental disorders (J. Zhang et al., 2017). A recent review based on a database of 141 studies involving over 590,000 participants from 40 different countries confirms that shorter sleep duration is associated with adverse physical and mental health outcomes in school-aged children and youth (Chaput et al., 2016). Indeed, the impact of sleep loss on mental health has a solid neuroscientific basis (Goldstein & Walker, 2014). Prior research reveals that adolescents with insufficient sleep suffer from increased levels of fatigue, decreased levels of energy, and other physical symptoms such as headaches, stomachaches, and backaches (Roberts & Duong, 2014; Roberts et al., 2001), many of which overlap with mental health problems. On the other hand, sleep time may be crowded out or compromised by private tutoring practices. Some ethnographic research speculates that students receiving private tutoring postpones the timing of dinner, bathing, and sleep (Bray, 2013). A recent study of 2,149 primary students in China finds that private tutoring significantly reduces students’ sleep time (Zhao & Xue, 2018). Another study middle school students in China finds that more time spent on private tutoring leads to worse sleep quality (Q. Zhang & Gao, 2022).
In framing the association between private tutoring and poor mental health, a recent study specifically argues that sleep deprivation may act as an important mediating factor for the association between private tutoring and mental health (Kuan, 2018). Unfortunately, the study did not provide empirical results to support this. Taken together, despite the documented associations of sleep time separately with private tutoring and adolescent mental health, the potential role of sleep time in mediating the association between private tutoring and mental health remains to be clarified. Therefore, our second hypothesis is:
H2: Greater involvement in private tutoring will lead to a decrease in sleep time, which will be further related to adolescent depressive symptoms.
The Influence of Grade Level
Little research has discussed the association between private tutoring and mental health about its potential variation over the schooling years. In particular, more evidence is needed with respect to whether and how the association of private tutoring with sleep time and mental health may differ among adolescent students in different grade levels. This study uses a nationally representative dataset of Chinese middle school students (grades 7–9). The data are suitable for investigating this issue by comparing students at different grade levels in the middle school years.
In China, middle school often spans three grades, from the 7th to the 9th, and is the last stage of compulsory education. The main aim of compulsory education is to provide all students with free and equal educational resources. Upon the completion of middle school, almost all students who expect to continue their studies must take a standardized, entrance exam to compete for a slot in senior high school, which marks the beginning of non-compulsory education and precedes college education. The universal access to compulsory education and expansion of college education in China have combined to create a bottleneck for continuing on senior high school (Lin & Zhang, 2006; Wu & Zhang, 2010). Most importantly, the senior high-school stage has long been characterized by an essential tracking process. In particular, the Chinese government has stressed the distinction between vocational and academic high schools, with the former being generally linked to vocational colleges and designed to train skilled workers. Despite being a central component of the educational system, vocational education in China is hardly recognized in the Chinese culture and ends up being the last study chance for students with the lowest entrance exam scores. On the other hand, among the academic high schools, the so-called “key point” high schools significantly excel others in terms of the threshold admission scores, quality of schooling, and the likelihood of entering tertiary education (Ye, 2015).
Therefore, although Chinese middle school students are still in the compulsory phase, as they progress from grade 7 to grade 9, they are getting closer to the entrance examination for academic and even key-point high schools. In this situation, Chinese parents and students turn to private tutoring to improve their chances of performing well in the upcoming entrance examinations (Y. Zhang & Xie, 2016). Existing studies reveal that higher grades are associated with more private tutoring practices (Y. Zhang, 2018), shorter sleep duration (Zhao & Xue, 2018), and more psychological symptoms (Tang et al., 2019). Thus, it is possible that the role of sleep in mediating the relationship between private tutoring and mental health may be more pronounced among students at higher grade levels. Accordingly, our third hypothesis is:
H3: The mediating role of sleep reduction becomes more pronounced as students move from lower to higher grades in middle school.
Data and Methods
Data
Data are from the first two waves of the China Education Panel Survey (CEPS), an ongoing nationwide longitudinal survey of a cohort of middle school students. The baseline survey was conducted during the 2013 to 2014 school year, and respondents were selected using a stratified, multistage probability proportional to size sampling. In practice, county-level districts served as the primary sampling unit (PSU), middle schools as the secondary sampling unit (SSU), and classes as the final tertiary sampling unit (TSU). Specifically, 28 county-level districts were first selected, then four middle schools were selected within each district, and then four classes-two in grade 7 and two in grade 9-were selected within each middle school. By design, the baseline survey did not collect data from grade 8.
All students in the selected classes were invited to participate in the survey. A total of 10,279 7th graders and 9,208 9th graders participated in the baseline survey. A follow-up survey was conducted 1 year later, during the 2014 to 2015 academic year, in which the 7th graders (n = 9,449) were followed up and the 9th graders were not. It should be noted that the 7th graders in the baseline survey were in 8th grade at the time of the follow-up. In this study, the two waves of the survey are treated as cross-sectional data and pooled, resulting in a sample of students (n = 28,936) covering the entire middle school period, that is, grades 7 to 9. Observations were excluded if they had missing responses to depressive symptoms (n = 818), private tutoring (n = 705), sleep duration (n = 390), and other covariates (n = 1,165). Therefore, our analysis includes a total of 25,862 students, including 8,987 7th graders, 8,659 8th graders, and 8,215 9th graders. The CEPS has several merits that fit our purposes. In addition to the detailed questions on students’ private tutoring practices, which allow for a comprehensive coverage of private tutoring participation, the large sample size of the CEPS allows for more nuanced patterns to be revealed through cross-grade comparisons.
Measures
Depressive Symptoms
Depressive symptoms are assessed using an adapted version of the Center for Epidemiologic Studies-Depression (CES-D) questionnaire (Radloff, 1977). It consists of five items centering on negative emotions, involving aspects such as depression, anxiety, unhappiness, meaninglessness, and sadness. On a Likert scale ranging from 1 (never) to 5 (always), respondents were asked to rate how often they had felt each emotion in the past week. The 5-item scale has been widely used in previous research on Chinese adolescents (Ge, 2020; Jiang et al., 2018; Xu et al., 2018; Young & Hannum, 2018). Most prior research has used simple summation to produce a total score, with higher scores indicating greater depressive symptoms. However, the scale has not been systematically evaluated for its psychometric properties in Chinese adolescents. To be conservative, the five items in this study were treated as representing one latent variable, depressive symptoms.
Private Tutoring
Involvement in tutoring has often been measured in terms of time and money (Chen & Lu, 2009; S. Hu & Mu, 2020; Zheng et al., 2020). This study uses three items asking about tutoring time and expenditure. The first two items assess the average amount of time spent on tutoring per day, one for weekdays and one for weekends. We used tutoring hours on weekdays and tutoring hours on weekends in our analysis. The third item measures the amount of money spent on tutoring: “During this semester, how much (in Chinese dollars) did you spend on your child’s private tutoring?” To reduce the potential leverage effect of a few extremely large values, we winsorized the values at the 99th percentile. The three tutoring items in this study were treated as representing one latent variable, private tutoring.
Sleep Time
Sleep time per night is the mediator of interest through which tutoring is expected to exert its effect on depressive symptoms. It was self-reported using the question: “On average, how much time do you sleep each night?” The variable is measured in hours and included in the analysis as a continuous variable.
Control Variables
Control variables include respondents’ social demographic characteristics and educational background. Social demographic characteristics include age in years, gender (0 = male and 1 = female), ethnicity (0 = Han and 1 = ethnic minority), parental education, family economic condition (1–5), and sibling information (0 = having siblings and 1 = only child). Parental education was assessed by taking the highest years of schooling reported for either mother or father. Educational background was assessed by grade level and academic performance. Academic performance is quantified by calculating the mean of the last midterm exam scores in Chinese, mathematics, and English. The participating schools provided the raw midterm scores. As the scores may vary across schools in terms of the upper limit and may not be comparable across grades, the data collectors standardized the raw scores by school and grade, making them to have a mean of 70 and a standard deviation of 10. As a result, the average academic performance included in the analysis also has a mean of 70. Table 1 presents descriptive statistics for the full sample and the grade-specific subsamples.
Descriptive Statistics: Full Sample and by Grade.
Note: Data are presented as mean (SD) for continuous measures, and % (n) for categorical measures. p-value refers to the test for differences across grade levels.
Analysis Strategy
We first use Ordinary least squares (OLS) regression analysis to assess the effects of private tutoring indicators on depressive symptoms indicators and the corresponding changes with the additional inclusion of sleep time. We then use Structural equation modeling (SEM) to analyze the relationship between tutoring involvement and depressive symptoms, and the extent to which sleep time may play a mediating role. This study relies on multiple indicators to represent the two broad constructs of private tutoring and depressive symptoms, respectively. SEM is thus very appropriate because it can account for potential measurement errors of observed variables and latent constructs. In addition to the full sample analysis, we conduct SEM analyses for the three grade-specific subsamples: 7th, 8th, and 9th graders. Throughout the modeling process, robust standard errors are obtained to account for the clustering nature of the dataset.
Results
OLS Regression Results
Table 2 presents the results from OLS regressions of each depression item on each tutoring item, first excluding and then including sleep time. When sleep time is excluded, the regression coefficients for all tutoring items are significantly positive, comprehensively indicating that greater involvement in private tutoring is associated with higher levels of depressive symptoms. This finding supports our first hypothesis (H1). After including sleep time, all of the positive coefficients for tutoring items shrink noticeably in size, suggesting that sleep time plays a mediating role. This finding supports our second hypothesis (H2). Overall, the findings provide a preliminary basis for suggesting that greater involvement in tutoring, represented by either more tutoring hours or monetary expenditure, may be associated with a decrease in sleep time, which may further translate into greater depressive symptoms. It should be noted that the separate depression items and tutoring items, although informative, may be subject to measurement error and cannot be described with certainty in terms of the independent association between each tutoring item and each depression item. Therefore, the OLS results here can be treated as a preliminary step before we use SEM approach to better capture the association between private tutoring and depressive symptoms.
Ordinary Least Square Regression of Each Depression Item on Each Tutoring Item.
Note. The items from dep-1 to dep-5 refer to “Feeling blue,” “Depressed,” “Unhappy,” “Not enjoying life,” and “Sad,” respectively. All models included a set of covariates including gender, age, ethnicity, parental education, family economic condition, only child, average academic performance, and grade. In parentheses are robust standard errors.
p < .05. **p < .01. ***p < .001.
SEM Results
To specifically examine whether and to what extent sleep duration can mediate the relationship between tutoring and depressive symptoms, we now turn to SEM analysis. Figure 1 plots the path coefficients for the full sample. The standardized coefficient on the path from private tutoring to sleep duration is β = −.134, p < .001. The negative sign of this coefficient is consistent with our expectation that greater involvement in private tutoring could be related to a loss of nightly sleep. In addition, the standardized coefficient on the path from sleep duration to depressive symptoms is β = −.152, p < .001, indicating a significant role of sleep duration in reducing depressive symptoms. Taking into account of the mediating role of sleep duration, the direct effect of tutoring on depressive symptoms is significantly positive (β = .047, p < .01). Table 3 summarizes the total, direct, and indirect effects of tutoring on depressive symptoms. The statistical significance of the indirect effect is reported based on the Sobel test (Sobel, 1982). As shown in the first column of Table 3, the indirect path via nighttime sleep duration explained 30.2% of the total effect of private tutoring on depressive symptoms, and the indirect effect was statistically significant (p < .001). Thus, we observe a clear indirect effect of private tutoring on depressive symptoms via sleep duration among Chinese middle school students. Here again, our second hypothesis (H2) is supported.
Standardized coefficients for paths linking private tutoring, sleep time, and depressive symptoms.
SEM Results: Tutoring Effects on Depressive Symptoms and Model Fit Statistics.
Note. CFI = comparative fit index; TLI = Tucker–Lewis index; SRMR = standardized root mean square residual; RMSEA = root mean square error of approximation.
Columns 2 to 4 of Table 3 present results from grade-specific subsamples. It is clear that at all three grade levels, greater involvement in tutoring is indirectly related to depressive symptoms through nighttime sleep duration. Comparatively, sleep duration explains the largest proportion (38.6%) of the tutoring effect on depressive symptoms for 9th graders, the second largest proportion (31.8%) of the effect for 8th graders, and the smallest proportion (21.3%) of the effect for 7th graders. Thus, the role of sleep time in channeling the link between private tutoring and depressive symptoms tends to become more pronounced as students proceed from early to later years in the middle school stage. As such, our third hypothesis (H3) is supported.
Conclusion
Despite the recent rapid deterioration of mental health and the remarkable expansion of private tutoring among adolescents worldwide, the relationship between private tutoring and adolescent mental health has received little attention from either health or education researchers. This study brings these two lines of inquiry together by linking private tutoring to depressive symptoms in a representative sample of Chinese middle school students (grades 7–9). This study additionally asks whether nighttime sleep duration can be an effective mechanism and whether the pattern of association between tutoring, sleep duration, and depressive symptoms varies across grade levels.
The results show that, overall, greater involvement in tutoring is associated with higher levels of depressive symptoms, ceteris paribus (Tables 2 and 3). This finding is consistent with several recent studies of middle school students in mainland China. For example, two recent studies found that more time spent on private tutoring was associated with significant declines in confidence in the future and psychological well-being (S. Hu & Mu, 2020; Q. Zhang & Gao, 2022). Another study found that higher levels of expenditure on private tutoring were related to higher levels of depressive symptoms (Zheng et al., 2020). On the other hand, many previous studies have documented the positive association between private tutoring and academic achievement and cognitive ability (e.g., Y. Zhang, 2018; Y. Zhang & Xie, 2016). Therefore, private tutoring acts as a double-edged sword, with beneficial effects on educational performance and negative effects on the psychological well-being of adolescent students (S. Hu & Mu, 2020; Kuan, 2018; Q. Zhang & Gao, 2022).
This study further considers sleep time as a possible mechanism in mediating the relationship between tutoring and depressive symptoms: greater involvement in tutoring leads to a decrease in nighttime sleep time, which further leads to an increase in depressive symptoms (Figure 1 and Column 1 in Table 3). Specifically, sleep time explains over 30% of the tutoring effect on depressive symptoms. Although the potential mechanism of sleep deprivation in explaining the association between tutoring and mental health has been discussed in several studies (Kuan, 2018; Q. Zhang & Gao, 2022), this study is among the first to specifically establish and test the link between tutoring, sleep time, and adolescent mental health.
Finally, this study highlights that the mediating role of sleep time becomes more pronounced during the middle school years as students progress from 7th to 9th grade (Columns 2–4 in Table 3). Most of the previous studies have predominantly focused on students from a single grade (Kuan, 2018; Li & Liu, 2022; Sun et al., 2020), which precludes the possibility of elaborating the potential variation in the pattern across grades or even educational levels. Findings in this study indicate that the mediating role of sleep time would be more pronounced as students approach an upcoming entrance exam for learning opportunities at a higher educational level. Future studies, hopefully with better data and design, could extend the observation to other educational contexts or levels, such as elementary or senior high school.
These findings have important implications for both health studies and education studies. For health studies, this study draws attention to an increasingly pervasive social behavior, private tutoring engagement, which, as clearly shown in our analysis, can significantly shape the psychological wellbeing of adolescents. For education studies, the impact of private tutoring on mental health outcomes warrants more attention, because mental health during the adolescent period could have profound effects on the likelihood of entering higher levels of education as well as the likelihood of obtaining degrees from elite universities. As a result, a better understanding of the psychological outcomes of private tutoring would ultimately contribute to the understanding of the formation of educational inequality. In particular, the grade-specific analyses reveal that the mediating role of sleep reduction in depressive symptoms intensifies in higher grades, particularly in 9th grade. This underscores the need for educational policies to prioritize interventions that promote healthy sleep habits, especially for older middle school students. Implementing targeted sleep education programs and stress management workshops could effectively mitigate depressive symptoms in adolescents, particularly those engaged in private tutoring.
The present study has several limitations. First, the data are cross-sectional, which means that it is not possible to establish a causal relationship between private tutoring, sleep deprivation, and depressive symptoms. Without longitudinal data from a complete cohort of middle school students, we have to use the cross-sectional differences across grade-specific subsamples to approximate the dynamic change from grade 7 to grade 9 for the same group of students. Another limitation is that the sample used in this study, although derived from a nationwide survey, covers a small age range and includes only students in middle school. Private tutoring is also common in elementary and senior high schools (Xue & Ding, 2009). To the extent that elementary or senior high school students may have different patterns in terms of private tutoring, sleep duration, and depressive symptoms, our findings based on middle school students cannot be readily generalized to these particular populations. Third, our key independent variable, private tutoring, despite our efforts to use tutoring information as much as possible, remains limited in capturing tutoring activities and quality of tutoring. Finally, the data analyzed in the current study were collected approximately 10 years ago (i.e., 2013–2015), which may raise concerns about their ability to reflect more recent trends in private tutoring and its psychological consequences among Chinese adolescent students. While acknowledging this limitation, it is important to emphasize that private tutoring has been a long-standing and deeply ingrained feature in China and other East Asian societies, driven by cultural and structural factors that have remained largely unchanged over time. Recent evidence further underscores the persistence of this phenomenon. For example, a study by Du (2024) found that 26.04% of primary and junior high school students in China received private tutoring in 2018, and this figure rose to nearly 40% in 2020 during the first wave of COVID-19. Given the lack of more recent publicly available data, our findings still provide valuable insights into the persistent dynamics of private tutoring and its impact on adolescent well-being, offering a meaningful foundation for future research.
In conclusion, this study highlights the importance of paying attention to the expansion of private tutoring, which may lead to more psychological problems among adolescents in many societies. Using data from a nationally representative sample of Chinese adolescents in middle school, we show that greater involvement in private tutoring is associated with greater depressive symptoms of adolescents, and in particular that decrease of sleep time explains a substantial part of the association. Notably, sleep time accounts for a larger proportion of the association between private tutoring and depressive symptoms among the 9th graders than among the 7th graders, probably because the former group was under greater pressure to prepare for the upcoming merit-based entrance examinations for a higher education level. Given the prominent prevalence of sleep problems and mental health problems and also the expansion of private tutoring among children and adolescents in recent decades, the findings of the current study provide empirical support for educational policies in regulating the development of the private tutoring at the societal level, and may also shed new light on tailored policies to alleviate sleep and mental health problems among adolescents. Future research, hopefully with longitudinal samples of students in a wider grade range, can shed more light on these association patterns.
Footnotes
Author’s Contribution
Yueyun Zhang and Jiankun Liu designed the research. All authors performed data analysis and literature review. Yueyun Zhang wrote the manuscript. All authors reviewed the manuscript.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research was supported by the National Social Science Fund of China (19CSH026). The founder had no role in the study design, data collection and analysis, writing of the paper and the decision to submit the paper for publication.
Ethical Approval
Ethical approval was waived using the publicly available data for the current study.
Informed Consent
Informed consent was obtained from all participants in the survey project.
Research Involving Human Participants and/or Animals
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.