The Impact of Family Background and Educational Investment on Students’ Cognitive and Logical Thinking Abilities: Evidence from the China Education Panel Survey

Abstract

This study explores how family background shapes children’s cognitive and logical thinking abilities within the context of contemporary China, using nationally representative data from the 2013–2015 China Education Panel Survey (CEPS). Recognizing the increasing stratification of educational outcomes, this research examines the dual roles of economic and cultural capital in influencing children’s development. Employing multivariate regression models and mediation analysis, we assess both direct effects of family background—measured by household economic status and parental education—and indirect effects through educational investments, including school choice, tutoring participation, academic support, and parental literacy habits. The results reveal that both economic and cultural capital have significant positive effects on students’ cognitive and logical thinking outcomes. However, cultural investment, particularly parental reading and engagement in children’s education, shows a more enduring and pronounced influence. Notably, children from the wealthiest families do not consistently perform better, suggesting that excessive reliance on material resources may crowd out effective parental engagement. In contrast, even the poorest families demonstrate strong educational aspirations, though constrained by limited resources and inadequate guidance. These findings highlight the critical role of cultural capital in mitigating intergenerational inequality and call for policies that support educational involvement across all socioeconomic groups to foster more equitable learning opportunities.

1. Introduction

In modern societies, education plays a dual role: it fosters social mobility and helps reduce inequality, but it also enables the intergenerational transfer of family capital, reinforcing existing social hierarchies. In today’s China, upward mobility through personal effort is still widely accepted and attainable, with education—especially the national college entrance exam (Gaokao)—serving as the primary pathway. As a result, Chinese parents—regardless of their socioeconomic status or education level—consistently place great importance on their children’s education and strive to support them as much as they can.

Educational competition has increasingly become a proxy for disparities in family income and parental education (Goudeau et al., 2025). Because education is seen as the key to breaking class barriers in an unequal society, it has triggered an increasingly intense phenomenon of ‘educational involution’ (Young & Hannum, 2018). In the Chinese educational context, this refers to the intense competition where students and families expend excessive time and resources on academic preparation, often leading to diminishing returns in genuine learning and personal development, rather than fostering creativity or critical thinking. On the one hand, families are allocating increasing amounts of time and money to their children’s education, creating substantial emotional and financial strain. Rising educational expectations and costs are now key factors shaping fertility choices. On the other hand, as income inequality in China widens, disparities in access to both in-school and out-of-school educational resources among students from diverse family backgrounds have become more pronounced (J. Liu et al., 2020). Children from low-income families face structural disadvantages in both educational access and attainment, with declining rates of higher education participation (Schneider et al., 2018).

The rising concern behind the question of whether advantaged families leverage education to consolidate their class advantages—and how this affects broader patterns of social mobility—constitutes a critical empirical question that this study seeks to address.

This study draws on data from the 2013–2015 waves of the China Education Panel Survey (CEPS), a longitudinal survey that tracks a nationally representative sample of students through critical educational transitions. While the timeframe is focused, it enables analyses of short-term mechanisms linking family background to student outcomes. It focuses on two dimensions of family background: economic capital (family income) and cultural capital (parental education), to investigate their effects—and the underlying mechanisms thereof—on students’ cognitive and logical thinking outcomes. Based on the findings, the study offers targeted policy recommendations to foster greater educational equity.

2. Literature Review

As the primary and earliest context for child development, the family plays a foundational role in shaping children’s learning behaviors and academic achievement. Landmark studies—the Coleman Report (U.S.) and the Plowden Report (U.K.) in the 1970s—demonstrated that family background has a more substantial effect on students’ study habits and academic disparities than school or community factors (H. C. Hill, 2017; Clausen, 2013).

A series of empirical studies across various national contexts, including European countries, have demonstrated that family socioeconomic status significantly shapes children’s access to education and attainment levels, exerting a consistently positive influence on academic performance (Kenny & Sirin, 2014). However, the applicability and interaction of these mechanisms within non-Western contexts, such as China, require critical examination. The Maximally Maintained Inequality (MMI) theory posits that educational expansion reduces inequality only after the saturation of privileged groups’ enrollment is achieved (Raftery & Hout, 1993). In China, despite the rapid expansion of higher education, inequality persists as advantaged families continuously adapt to maintain relative advantages, supporting MMI’s core premise but also highlighting its limitations in capturing localized stratification mechanisms, such as the role of the Gaokao examination system and urban–rural disparities. The theory of Effectively Maintained Inequality (EMI) further emphasizes that inequality is sustained not only through quantitative access but also through qualitative differences in school type, curriculum, and academic tracks (Lucas, 2001). Recent studies in China illustrate that EMI manifests via competition for elite urban schools, shadow education (private supplementary tutoring), and tracking within schools—mechanisms that align with EMI predictions but are intensified by China’s unique demographic and institutional structures (Tan & Liu, 2023; Gong et al., 2025; W. Zhang & Bray, 2021).

As a result, children from privileged families are more likely to secure access to higher-quality educational resources. The intergenerational transmission of advantage can be conceptualized as a pathway that proceeds from family background to educational investment, then to educational attainment, human capital, labor productivity, and ultimately income and social status. In this chain, family background influences children’s educational investment, which affects long-term outcomes in skill formation, earnings, and mobility (Schneider et al., 2018).

In the wake of China’s economic reforms, social stratification based on income and asset ownership has become increasingly pronounced, and the influence of family economic and social status on individuals’ access to and achievement in education has grown significantly. Despite the universalization of compulsory education and the massification of higher education, disparities in educational opportunity remain persistent (J. Liu et al., 2020; Y. Guo & Li, 2025). In fact, the expansion of higher education has disproportionately benefited students from socioeconomically advantaged groups, particularly urban residents, thereby deepening the urban–rural divide in college access (F. Li & Hong, 2024; Y. Li & Wang, 2023). These unequal opportunities in higher education contribute to divergent labor market outcomes and reinforce stratification in long-term social status. Empirical evidence suggests that inequality in access to higher education is rooted in the cumulative disadvantages embedded in earlier stages of education (X. Zhang & Chen, 2021; Carvalhaes et al., 2023). Students from lower-income or rural families are often tracked out of the academic pathway before senior secondary school, reflecting entrenched structural inequalities in the distribution of educational opportunities across social strata (Kuo et al., 2018). The burden of family education has become one of the most concerning issues in today’s society. Scholars have conducted research on the relationship between economic inequality and parenting, revealing that over the past four decades, there has been a global shift from laissez-faire to intensive parenting styles (Schneider et al., 2018). Among these trends, parents in Asia, particularly East Asia, generally emphasize substantial educational investment in their children, resulting in particularly severe academic pressure for students and significant economic burdens for families.

In today’s knowledge-driven economy, parents from all social backgrounds generally believe in education as a means of upward mobility. However, due to financial constraints, distinct class-based differences emerge in actual educational investment behaviors. There exists a clear socioeconomic gradient: from the poorest to the wealthiest, families demonstrate progressively higher levels of financial investment in their children’s education across all stages. Families with stronger financial backgrounds not only access high-quality educational resources early on but also continue to expand their investments in diverse forms to sustain their educational advantage. For instance, private tutoring institutions in urban settings have exacerbated resource disparities, reinforcing cumulative advantages for students from affluent families (A. Liu et al., 2024). As a result, these trends have undermined efforts to equalize education and deepened structural inequalities.

In July 2021, China’s central government introduced the “Double Reduction” policy—formally titled Opinions on Further Reducing the Burden of Homework and Off-Campus Training for Students in Compulsory Education—which aimed to strengthen the regulation of private tutoring and reinforce the role of schools as the main venue for learning. However, numerous studies indicate that the policy has had limited impact on reducing student burden (Q. Wang & Jiang, 2024; L. Chen & Lin, 2024; Zhou & Fan, 2025). In the context of fierce academic competition, restricting the supply of educational services has proven ineffective in easing pressure, and has instead amplified concerns over educational inequality.

A review of the literature reveals a scholarly consensus that family background significantly influences student achievement (Helin et al., 2023; Wu & Zhang, 2024; Yan & Gao, 2024; J. Zhao & Bodovski, 2020). These disparities—particularly in economic and educational investments—contribute to unequal starting points and outcomes in education, thereby intensifying the risk of intergenerational social immobility. However, few studies have explored the micro-level mechanisms through which family background shapes educational outcomes (Akabayashi et al., 2020). This study addresses this gap by investigating both the magnitude and mechanisms of family background’s effects on children’s cognitive and logical thinking development, with the aim of informing strategies to enhance educational equity and mitigate the risks of educationally driven social immobility.

3. Theoretical Basis

The impact of family background on children’s educational opportunities and attainment is largely mediated through differences in educational investment. Educational investment is generally categorized into two dimensions: economic capital and cultural capital (C. Guo & Min, 2006; Zhu, 2005). High-income families can provide superior and stable learning environments—for instance, by affording school placement fees, purchasing property in desirable school districts, securing sustained access to private tutoring, and ensuring conducive home study conditions. In contrast, low-income families often lack the financial capacity to invest sufficiently in their children’s education, resulting in limited access to learning resources (Qin et al., 2025).

One pathway through which advantage is reproduced is school choice. From the compulsory education stage onward, children from families with greater financial and cultural capital are more likely to attend well-resourced and high-performing schools. These institutions offer superior teaching quality and resources, increasing students’ chances of being placed on favorable academic tracks.

Another critical mechanism is the rise of what is known as “shadow education”—a term referring to fee-paying, private supplementary tutoring outside the formal school system. This sector has emerged as a booming industry amid the universalization of urban compulsory education and significantly shapes students’ academic achievement. Studies indicate that participation in shadow education is largely driven by family income. Affluent students not only have higher participation rates but also tend to enroll in higher-quality tutoring programs, thereby exacerbating educational inequality (S. Li, 2023). Even when controlling for other factors, shadow education shows a measurable positive impact on academic performance and broader competencies, ultimately improving students’ prospects in terms of further education and career advancement.

Studies show that family income is the primary determinant of participation in shadow education. Moreover, affluent students tend to enroll in higher-quality tutoring programs, thereby amplifying the inequality in educational outcomes (S. Li, 2023). When controlling for other factors, shadow education has a measurable positive impact on academic performance and broader competencies, ultimately enhancing students’ prospects for further education and career advancement.

Drawing on Bourdieu’s theory of social reproduction, empirical research has shown that cultural capital exerts a stronger influence on students’ academic achievement than economic capital (Bourdieu, 1986, 2018; Luo et al., 2022; Y. Zhao et al., 2023).

A key indicator of this is the robust correlation between parental education levels and the educational attainment of their children (Sayer et al., 2004).

Cultural capital includes not only parental education and knowledge but also aesthetic dispositions, preferences, and an understanding of how to navigate the formal education system. Well-educated families tend to prioritize their children’s academic development from an early age. They often foster effective study habits, encourage intrinsic motivation for learning, and actively participate in school-related activities. Such parents are also more likely to communicate regularly with teachers and other parents, and their own reading habits often serve as implicit role models that support their children’s academic growth (N. E. Hill & Tyson, 2009; Benner et al., 2016; Tárraga et al., 2017; Kim et al., 2024).

Moreover, parental expectations serve as a key mediating mechanism linking cultural capital to higher education access. Families with greater cultural resources tend to set higher educational goals for their children and actively support their realization, thereby transforming aspirations into academic advantages (Pitman, 2013; Low, 2015; S. Wang & Huang, 2021).

4. Research Design

Drawing on the preceding literature and theoretical framework, this study constructs an empirical model to test the effects of family background on children’s cognitive and logical thinking outcomes. Using data from the 2013–2015 wave of the China Education Panel Survey (CEPS), the study employs quantitative methods to analyze how economic and cultural dimensions of family background affect both types of student outcomes, and to explore the mechanisms behind these effects (NSRC, 2013–2015). The ultimate aim is to generate theoretical insights or empirical evidence that supports efforts to improve educational access and outcomes for children from disadvantaged backgrounds, and to advance broader goals of educational equity.

4.1. Data Source

This study utilizes data from the 2013–2015 wave of the China Education Panel Survey (CEPS), designed and implemented by the Institute of Social Science Survey at Renmin University of China. The survey began with two grade cohorts—7th and 9th grades—and used a stratified, nationally representative sampling method to select 112 middle schools and 448 classes, yielding a total sample of approximately 22,000 students.

Given the well-documented disparities in educational investment between urban and rural families, this study focuses specifically on urban households (Y. Li et al., 2020). Accordingly, only students with an urban hukou (a household registration system that ties access to public services to one’s registered location) were included in the analysis, resulting in a final sample of 8526 valid observations (Chan & Zhang, 1999).

4.2. Model Specification

Based on the above analysis, this study constructs a regression model to examine the impact of family background on children’s cognitive and logical thinking abilities (Yang, 2020; Y. Q. Li, 2022; Fang & Hou, 2019). The model aims to capture the direct effects of family socioeconomic status and parental education, while controlling for relevant individual and school-level characteristics.

The basic model can be specified as follows:

$${\mathrm{Ln}(\mathrm{Grade}}_{ijt})={\alpha }_0+{\beta }_1{\mathrm{Wealth}}_{ijt}+{\beta }_2{\mathrm{Edu}}_{ijt}+{\displaystyle \sum _{k=1}^K{\delta }_k{\mathrm{Control}}_{ijt}^k}+{\mu }_i+{\gamma }_t+{\epsilon }_{ijt}$$

(1)

$$\mathrm{Ln}Score-co{g}_{ijt}={\alpha }_0+{\beta }_1{\mathrm{Wealth}}_{ijt}+{\beta }_2{\mathrm{Edu}}_{ijt}+{\displaystyle \sum _{k=1}^K{\delta }_{k}Contro{l}_{ijt}^k}+{\mu }_i+{\gamma }_t+{\epsilon }_{ijt}$$

(2)

Here, the subscripts i, j, and t denote region, student, and time, respectively; β₁ and β₂ represent the estimated coefficients for wealth and educational level, respectively; K refers to the total number of control variables included in the model, with δ_k denoting their corresponding coefficients. Ln(grade_ijt) denotes the natural logarithm of the cognitive ability score of student j in region i at time t. ln(Score_cog_ijt) denotes the natural logarithm of the logical thinking ability score of student j in region i at time t. Wealth_ijt captures the economic status of the household, serving as a proxy for family economic capital. Edu_ijt represents the highest level of parental education, serving as a proxy for family cultural capital. Control_ijt is a set of individual and family-level control variables, such as gender, school quality, only child status and grade level. μ_i represents region-fixed effects, controlling for unobserved regional heterogeneity. γ_t denotes year-fixed effects, controlling for temporal shocks or policy changes. ε_ijt is the error term.

The logarithmic transformation of the dependent variables (grade and cognitive/non-cognitive scores) was applied to normalize the distribution of the outcome measures and improve model linearity, as is common in econometric analyses of educational and behavioral outcomes. All models report robust standard errors clustered at the school level to account for within-school correlations in the error terms and ensure valid statistical inference.

4.3. Variable Description

4.3.1. Dependent Variables: Cognitive and Logical Thinking Abilities

Standardized Test (LnGrade) is measured as the average of midterm exam scores in Chinese, Mathematics, and English for the current academic year, as reported by the schools. The natural logarithm of this average score is used as the proxy for students’ cognitive ability.

Logical thinking ability (LnScore-cog) is based on a standardized, nationally representative, and internationally comparable CEPS assessment designed by grade level to measure logical reasoning and problem-solving skills. The natural logarithm of the test score is used as the proxy for logical thinking ability.

4.3.2. Independent Variables: Family Background

Based on Bourdieu’s theory of capital, which conceptualizes social position as shaped by the distribution of economic and cultural capital (Bourdieu & Passeron, 1990), this study includes two indicators of family background, as follows.

Economic capital (Wealth) is derived from parental self-assessments of household economic conditions and is categorized into five levels: very affluent, relatively affluent, average, relatively disadvantaged, and very disadvantaged.

Cultural capital (Edu) is measured by the years of schooling attained by the more educated parent. It is categorized into nine levels: no formal education, primary school, junior high school, technical/vocational school, vocational high school, academic high school, associate degree, bachelor’s degree, and postgraduate or above.

4.3.3. Mediating Variables

As previously noted, family economic status and parental education do not directly determine students’ academic outcomes, but rather exert their influence through mediating factors such as educational investments. These investments enhance learning environments, influence learning behaviors, and shape students’ academic goals. Accordingly, this study incorporates two categories of mediators to examine the mechanisms through which family background affects student performance, namely, economic capital investment and cultural capital investment.

Economic capital investment is measured using three indicators, as follows:

School choice behavior (SC): Whether the family actively pursued enrollment in a preferred school.

Supplementary education participation (SE): Whether the student received private tutoring.

Supplementary education cost (SEC): The reported cost of tutoring per semester.

Cultural capital investment is assessed via four dimensions, as follows:

Parental homework assistance (AT): The frequency of helping the child with homework over the past month.

Parent–teacher communication (SAT): The number of proactive contacts with the homeroom teacher during the previous semester.

Cultural co-participation (EA): The frequency of engaging in cultural enrichment activities with the child in the previous semester.

Parental reading habits (PRH): The average number of hours per week spent reading during the past six months.

4.3.4. Control Variables

The control variables include the following: gender, categorized as male or female; school quality (Level), a self-reported ranking of the school’s teaching quality and facilities within the county/district, ranging from lowest to highest across five levels; only child status (Brosis), which concerns whether the student is the only child in the household; grade level, which concerns whether the student is in Grade 7 or Grade 8. Please refer to

Table 1. Variable descriptions.

Variable Category		Indicator	Description	Mean	Std. Dev.
Dependent Variables		Standardized Test Scores (LnScore-cog)	Log of standardized test scores for cognitive ability	2.603	0.529
		Logical Thinking Ability Score (LnGrade)	Log of standardized recent midterm exam scores	5.433	0.351
Independent Variables	Family Economic Background	Household Wealth (Wealth)	1 = Very Poor; 2 = Poor; 3 = Average; 4 = Relatively Wealthy; 5 = Wealthy	2.85	0.50
	Family Cultural Background	Parental Education Level (Edu)	No schooling = 0; Primary = 3; Junior High = 9; Vocational/Tech = 11; Senior High = 12; Associate = 15; Bachelor’s = 16; Graduate and above = 19	10.87	3.05
Mediating Variables	Economic Capital Investment	School Choice (SC)	Binary variable: 0 = School choice; 1 = No school choice	0.35	0.41
		Participation in Tutoring (SE)	Binary variable: 0 = No; 1 = Yes	0.47	0.46
		Tutoring Expenditure (SEC)	Total tutoring expenses per semester (range: [0, 80,000])—continuous variable	4476	0.54
	Cultural Capital Investment	Academic Supervision (AT)	Continuous variable, weekly frequency [1, 4]	2.6	0.32
		Communication with Head Teacher (SAT)	Continuous variable, per semester [1, 4]	2.1	0.34
		Enrichment Activities (EA)	Continuous variable, monthly frequency [1, 6]	3.2	0.34
		Parental Reading Habits (PRH)	Continuous variable, weekly reading time [1, 4]	2.1	0.41
Control Variables	Student Characteristics	Grade (Grade)	0 = Grade 1 (Junior); 1 = Grade 2; 2 = Grade 3	0.472	0.499
		Gender (Gender)	Gender (Male = 1, Female = 0)	0.5014	0.5000
		School Tier (Level)	Ranking of school (1 = Low/Mid; 2 = Upper-Mid; 3 = Top)	2.0636	0.6490
		Only Child (Brosis)	Whether the student is an only child (1 = Yes, 0 = No)	0.4434	0.4968

for all variable descriptions.

5. Empirical Analysis

5.1. Baseline Regression Analysis

Stepwise regressions based on Equations (1) and (2) were employed to assess the impact of family background on students’ cognitive and logical thinking abilities.

Table 2. Influence of family background on children’s cognitive and logical thinking abilities.

Variable	Standardized Test Score (1)	Standardized Test Score (2)	Logical Thinking Ability Score (3)	Logical Thinking Ability Score (4)
Household Wealth	0.036 ***		0.029 ***
	(8.013)		(7.013)
Parental Education Level		0.054 ***		0.058 ***
		(0.013)		(0.013)
Grade	0.215	0.173	0.164	0.156
	(0.017)	(0.017)	(0.017)	(0.017)
Gender	−0.003 ***	−0.000 ***	−0.129 ***	−0.129 ***
	(0.163)	(0.015)	(−2.32)	(−2.33)
School Tier	0.532 ***	0.388 ***	0.204 ***	0.172 ***
	(0.008)	(0.000)	(0.001)	(0.002)
Only Child	0.030 **	0.028 **	0.084 **	0.067 ***
	(8.645)	(6.255)	(11.585)	(8.847)
Constant	0.316 ***	0.498 ***	0.529 ***	0.632 ***
	(0.008)	(0.002)	(0.000)	(0.006)
Year/District	Yes	Yes	Yes	Yes
N	8526	8526	8526	8526
R2	0.569	0.577	0.276	0.278

Notes: “**”, and “***” denote statistical significance at the 5%, and 1% levels, respectively. Robust standard errors clustered at the city level are reported in parentheses (Pustejovsky & Tipton, 2018); the same applies to the following tables.

presents the estimated effects of family economic status and parental educational attainment on students’ cognitive and logical thinking outcomes. Specifically, Columns (1) and (2) report the effects on cognitive ability, while Columns (3) and (4) present the corresponding effects on logical thinking ability.

The results demonstrate that both family economic status and parental educational attainment have significant and positive associations with students’ cognitive and logical thinking outcomes. A one-level increase in family economic status is associated with a 3.6% rise in cognitive ability and a 2.9% rise in logical thinking ability. Each additional level of parental educational attainment corresponds to a 5.4% increase in cognitive ability and a 5.8% increase in logical thinking ability. This substantial influence suggests that the intergenerational transmission of human capital plays a critical role in shaping educational inequalities, as higher parental education often provides children with not only greater material resources but also enhanced cultural capital and academic expectations.

Regarding the control variables, school quality shows a significantly positive effect on both cognitive and logical thinking abilities. Higher-ranked schools are associated with better student outcomes, particularly in cognitive domains. This suggests a correlation where students in top-tier schools tend to exhibit not only stronger academic performance but also enhanced skills in reasoning and reading.

Gender has a significantly negative effect, indicating that female students tend to outperform male students in both cognitive and logical thinking competencies.

Only child status shows a significantly positive association, possibly due to the concentration of family resources, which facilitates both cognitive and logical thinking development. This supports the “quantity–quality trade-off” hypothesis widely discussed in the context of Chinese families (Lee, 2008; Zhong, 2017).

Grade level does not have a statistically significant effect on cognitive ability.

5.2. Robustness Checks

To verify the reliability and robustness of the empirical findings, several robustness checks are implemented.

5.2.1. Replacing Explanatory Variables

Existing research has demonstrated that educational attainment is closely associated with occupational choice and income stratification. Accordingly, this study replaces parental education and household economic status with parental occupation as an alternative measure of family background. Parental occupations are coded as follows: unemployed/laid-off = 1; self-employed/farmers = 2; general/skilled workers = 3; public institution staff or corporate executives = 4; professionals such as teachers, engineers, doctors, or lawyers = 5. The higher value of the two parents is used. As shown in Column (1) of

Table 3. Robustness test results.

Test Method	Replacing Explanatory Variable (1)	Replacing Explained Variable (2)	Excluding Top 10% Households (3)	Winsorization 1% (4)	Winsorization 5% (5)
Household Wealth	0.009 ***	0.016 **	0.008 **	0.010 ***	0.009 ***
	(0.003)	(0.007)	(0.003)	(0.003)	(0.002)
Parental Education Level	0.009 **	0.012 **	0.002 *	0.015 ***	0.011 ***
	(0.012)	(0.009)	(0.001)	(0.002)	(0.001)
Control	Yes	Yes	Yes	Yes	Yes
Year/District FE	Yes	Yes	Yes	Yes	Yes
N	8526	8526	8526	8526	8526
R2	0.7034	0.8459	0.8255	0.8726	0.7034

Notes: “*”, “**”, and “***” denote statistical significance at the 10%, 5%, and 1% levels, respectively. Numbers in parentheses are robust standard errors; see the following table for details.

, the direction and statistical significance of the coefficients remain unchanged, supporting the robustness of the main findings.

5.2.2. Replacing Dependent Variables

The students’ cognitive and logical thinking ability scores are recalculated using indicators derived from a three-parameter Item Response Theory (IRT) model. Column (2) of Table 3 shows that the coefficients remain positive and statistically significant, further validating the robustness of the baseline estimates.

5.2.3. Excluding Cities with Highest GDP Levels

To account for potential biases arising from regional disparities in economic development and educational resources, this study excludes cities in the top 10% in terms of per capita GDP—such as Beijing, Shanghai, Tianjin, Wuxi, and Guangzhou. Column (3) in Table 3 shows a slight reduction in the coefficients, but the effects remain statistically significant, indicating robustness to regional economic variation.

5.2.4. Adjusting for Outliers (Winsorization)

To address the potential impact of outliers caused by data entry errors, extreme values, or specific school contexts, the student test scores are Winsorized at both the 5–95% and 1–99% levels. As reported in Columns (4) and (5) of Table 3, the main results remain consistent, suggesting that outliers exert limited influence on the estimates.

6. Mechanism Analysis

Family background may not exert a direct effect on children’s cognitive and logical thinking abilities; instead, its influence is potentially mediated through various forms of family educational investment. It is therefore essential to examine the underlying mechanisms through which family background affects children’s academic and developmental outcomes via family-level investments.

This study adopts school choice behavior, participation in supplementary tutoring, tutoring expenditures, parental involvement in homework, frequency of communication with teachers, engagement in enrichment activities, and parents’ reading habits as mediating variables to test this pathway.

Following established research frameworks, the analysis proceeds in two steps. First, it assesses the impact of family background on educational investment; second, it evaluates how such investment, in turn, influences children’s cognitive and logical thinking performance.

6.1. Influence of Family Background on Educational Investment

First, we analyze how family background influences investments in economic educational capital. As shown in Column (1) of

Table 4. Analysis of the impact of family background on household education investment.

Variable	Family Background
	Household Wealth	Parental Education Level
School Choice	0.3530 *** (0.149)	0.2249 *** (0.143)
Participation in Tutoring	0.4471 *** (0.176)	0.2050 ** (0.153)
Tutoring Expenditure	0.1904 *** (0.173)	0.1214 * (0.145)
Academic Supervision	0.520 ** (0.146)	0.254 *** (0.154)
Communication with Head Teacher	0.0495 (0.146)	0.1810 *** (0.144)
Enrichment Activities	0.3472 *** (0.081)	0.4212 *** (0.100)
Parental Reading Habits	0.5501 (0.148)	0.5181 *** (0.150)
Control	Yes	Yes
Year/District	Yes	Yes
N	8526	8526
R2	0.005	0.007

Notes: “*”, “**”, and “***” denote statistical significance at the 10%, 5%, and 1% levels, respectively. Numbers in parentheses are robust standard errors; see the following table for details.

, each one-level increase in family economic status is associated with a 35% higher probability of school selection; similarly, a one-level increase in the highest parental education level raises this probability by 22%. Compared with parental education, family economic status is the dominant factor influencing participation in extracurricular tutoring and related expenditures. Specifically, each level increase in family economic status increases the likelihood of tutoring participation by 40% and raises tutoring expenditures by 19%. Although parental education also has a significant positive effect, its magnitude is comparatively smaller.

Second, we examine how family background influences cultural educational investment. Column (2) of Table 4 shows that parental education level significantly influences all four dimensions of cultural capital investment. A one-level increase leads to a 25% rise in weekly academic support time and an 18% increase in parent–teacher communication frequency. Both economic status and educational background significantly promote children’s participation in enrichment activities. Parental education level is also strongly correlated with reading habits; each additional level is associated with a 52% increase in weekly reading hours. In contrast, economic status shows no statistically significant effect on tutoring involvement, communication with teachers, or parental reading habits.

In summary, both family economic and educational backgrounds are positively associated with children’s educational investments. Compared with parental education, economic background has a stronger effect on school choice and tutoring participation, while educational background more strongly predicts cultural investment patterns such as parental engagement, communication, and reading behaviors.

6.2. The Impact of Family Educational Investment on Children’s Cognitive and Logical Thinking Abilities

Multiple linear regression (OLS) is employed for continuous dependent variables and Logit/Probit models are used for binary variables to examine the impact of economic and cultural capital investment on children’s cognitive and logical thinking outcomes.

The results in Columns (1)–(3) of

Table 5. Analysis of the impact of family education investment on children’s cognitive and logical thinking abilities.

Influence Mechanism	Economic Capital Investment			Cultural Education Investment
Variable	School Choice (1)	Participation in Tutoring (2)	Tutoring Expenditure (3)	Academic Supervision (4)	Communication with Head Teacher (5)	Enrichment Activities (6)	Parental Reading Habits (7)
Standardized Test Score	0.017 ** (0.013)	1.607 *** (0.095)	0.010 *** (0.011)	0.326 *** (0.009)	0.178 ** (0.116)	0.422 *** (0.033)	0.729 *** (0.001)
Logical Thinking Ability Score	0.027 (0.013)	1.607 *** (0.005)	0.010 *** (0.022)	0.226 *** (0.032)	0.378 ** (0.123)	0.242 *** (0.033)	0.722 *** (0.005)
Control	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year/District FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	8526	8526	8526	8526	8526	8526	8526
R2	0.005	0.007	0.630	0.268	0.145	0.821	0.905

Notes: “**”, and “***” denote statistical significance at the 5%, and 1% levels, respectively. Numbers in parentheses are robust standard errors; see the following table for details.

show that economic capital investment has a significant impact on both cognitive and logical thinking abilities. School choice shows a marginally significant effect on cognitive performance but no meaningful effect on logical thinking outcomes. Given the intrinsic relationship between participation in private tutoring and its associated expenditure, we analyzed their composite effect. The results show that participation exerts a fundamental positive effect on both competency dimensions. The regression coefficients indicate that while both participation and expenditure exhibit positive coefficients, the effect of participation is more substantial. This pattern suggests that the primary driver of the observed benefits is the decision to participate in tutoring itself, whereas the specific amount spent demonstrates a relatively minor influence on the outcomes.

Columns (4)–(7) report the effects of cultural capital investment. Parental academic support time, frequency of parent–teacher interaction, extracurricular participation, and parents’ reading habits all exert significant positive effects on both cognitive and logical thinking outcomes. Notably, parents’ reading habits have the strongest positive association among all cultural capital factors.

Synthesizing both sets of regression analyses, we can conclude that family background influences children’s cognitive and logical thinking development primarily through educational investment. Economic background promotes school choice and tutoring, while educational background contributes through a broader range of channels, including academic support, enrichment activities, parental engagement, and reading habits.

Thus, educational background appears to have a more pronounced relationship than economic status in shaping children’s ability development. Several mechanisms may help explain this difference.

First, highly educated parents are better equipped to support academic content due to their knowledge base and are more likely to identify learning difficulties early. This advantage can be understood through Bourdieu’s concept of cultural capital, where parental knowledge represents embodied cultural capital that is transmitted to children, fostering academic readiness.

Second, their extensive educational experience enhances their understanding of school norms, encouraging more proactive engagement with teachers and educational processes. This fosters synergy between home and school, strengthening the child’s academic identity. This aligns with Bourdieu’s notion of institutionalized cultural capital and social capital, as parents leverage their familiarity with the system to build networks and navigate institutions, thereby reinforcing home–school synergy and strengthening the child’s academic identity.

Third, such parents often maintain lifelong learning habits (e.g., reading, critical thinking), which influence household values and subtly shape children’s attitudes, motivation, and learning behaviors. These habits reflect the habitus associated with cultural capital, creating a psychologically supportive environment that perpetuates academic advantages across generations, in line with Bourdieu’s theory of reproduction. These mechanisms create a supportive environment conducive to long-term academic success.

This once again validates Bourdieu’s theory of social capital, indicating that privileged groups with abundant cultural capital find it easier to invest in and maintain valuable social networks. They can convert this social capital into other forms of capital, thereby consolidating and transmitting their advantageous status.

7. Further Discussion

7.1. The Situation of the Wealthiest Families

As established earlier, family economic capital generally shows a significant positive effect on children’s academic cognitive abilities and logical reasoning skills. However, CEPS data reveal a non-linear trend; while children’s standardized scores tend to rise with increasing family economic status, there is a pronounced decline in both cognitive and logical thinking performance among students from the very wealthiest families.

This finding not only aligns with but also extends existing theories of cultural and social reproduction (Bourdieu, 1986), suggesting that extreme economic advantage may lead to a decoupling of economic and cultural capital. Our results provide empirical support for the notion that cultural and educational investment—rather than economic resources alone—is a decisive mechanism influencing children’s academic outcomes. It further suggests that parents in the wealthiest households may rely heavily on material investment while neglecting active cultural or educational engagement.

Data show that despite high levels of financial investment, the wealthiest families demonstrate below-average engagement in key educational activities, including homework assistance, parent–teacher communication, and school participation. Notably, 36.4% of parents in the wealthiest group reported not having read a single book in the past year—the highest share across all socioeconomic categories.

This apparent disengagement underscores the novel contribution of our study, namely, we identify and empirically demonstrate a “high-income dip” pattern in educational engagement and child development within the Chinese context, a phenomenon that is underexplored in prior literature. It may help explain why children from the highest-income families perform worse in both cognitive and logical thinking domains, despite their material advantages.

7.2. The Situation of the Poorest Families

Analysis of the survey data indicates that the most economically disadvantaged families have not disengaged entirely from their children’s education. Specifically, their rate of school choice—actively selecting schools for their children—is higher than the overall sample average. Furthermore, their expenditures on tutoring classes are not the lowest among all income groups. These behaviors reflect strategic adaptive practices described in theories of marginalized families and directly address our research question regarding how low-SES families navigate structural constraints through educational agency.

However, in other domains of educational involvement—particularly in-home academic support, parent–teacher communication, participation in enrichment activities, and parents’ own reading habits—the lowest-income households consistently rank at the bottom. This is consistent with the theoretical argument that non-economic capital, such as cultural and social resources, mediates educational success (Bourdieu, 1986; Y. Zhao et al., 2023). These families face dual barriers: not only financial constraints but also limited access to effective parenting strategies and educational resources.

These findings suggest that while poor families maintain strong aspirations for their children’s upward mobility and view education as a critical pathway for mobility, they face dual barriers—not only financial constraints but also limited access to effective parenting strategies and educational resources. This reinforces our hypothesis that non-economic forms of capital play a differentiating role even among economically disadvantaged groups.

Overall, the analysis offers a significant scientific upgrade by integrating quantitative patterns with theoretical mechanisms and contributes novel insights into the non-linear relationships between family socioeconomic status and child development in China.

8. Conclusions

First, both household economic conditions and parental education levels have a significantly positive effect on children’s cognitive and logical thinking development.

Students from more privileged family backgrounds are better positioned to access superior educational resources both in and out of school—through mechanisms such as school choice, supplementary tutoring, academic guidance, and participation in extracurricular enrichment. These advantages help them perform better both cognitively and logically, placing them in a favorable position within the mainstream education system.

In contrast, children from low-income and low-education households typically experience limited financial investment and insufficient educational or cultural support at home, which often results in lower performance in both cognitive and logical thinking domains. This underscores the failure of the current public education system to adequately address the educational inequalities rooted in family background.

Although academic achievement during junior high school is not the sole determinant of one’s future career or life trajectory, under China’s nine-year compulsory education framework, students who underperform are at a much higher risk of being excluded during the transition to senior high school, thereby losing opportunities to pursue higher education.

Therefore, the growing alignment between student outcomes and family background may point to a future in which social stratification becomes more pronounced and rigid, posing the risk of further entrenching class-based inequalities.

Secondly, cultural education investment exerts a more substantial influence on student outcomes than economic capital investment. Although household wealth and income disparities in China are growing, cultural stratification remains relatively fluid and has not yet hardened into insurmountable class boundaries. Unlike in many Western societies, Chinese families across all socioeconomic backgrounds maintain similarly high educational expectations for their children.

Therefore, effective cultural educational investment—such as parental involvement, reading habits, and academic engagement—can significantly improve the developmental outcomes of children from disadvantaged backgrounds and serve as a buffer against intergenerational class reproduction. The continued emergence of “successful children from poor families” in China suggests that social mobility remains viable, and the educational system retains a degree of openness.

To mitigate the intergenerational impact of family background and educational disparities—and to safeguard educational equity—public policy should adopt a multifaceted approach.

At the family level, initiatives such as parenting workshops and literacy programs should be implemented to raise parental cultural capital and foster an emotionally supportive learning environment at home, thereby mitigating the disadvantages associated with insufficient economic capital.

At the school level, reducing disparities between elite and non-elite schools is essential. Priority should be given to improving instructional quality, enhancing teachers’ professional development, and cultivating students’ intrinsic motivation and learning habits through differentiated instruction.

Regarding after-school tutoring, policy should shift the focus from selective enrichment to remedial support, enforce pricing regulations, and prevent excessive market competition that places a disproportionate financial burden on low-income families.

Ultimately, strengthening the educational capacity of disadvantaged households is a necessary pathway to achieving substantive social equity and reversing the trend of structural class immobility.

It is important to acknowledge several limitations of this study. First, the reliance on CEPS data, while nationally representative, restricts our analysis to observed variables and may not capture all nuanced aspects of family dynamics and educational processes. Second, the cross-sectional nature of part of the data limits causal inference. Future research could employ longitudinal designs to track developmental trajectories over time and incorporate mixed methods to explore qualitative mechanisms behind the observed patterns. Additionally, further investigation is needed into the role of school-level policies and teacher practices in moderating or amplifying family effects.