Children's education and parents' dietary nutrient intake: an empirical study based on rural China

The study examines whether children’s educational attainment has an upward spillover effect on their parents’ dietary nutrient intake in rural China. While nutrition is fundamental to health and education is known to improve dietary behaviors, middle-aged and elderly rural residents cannot feasibly raise their own education levels, and large-scale public nutrition education is costly in sparsely populated rural areas. Given China’s rapid expansion of education and large intergenerational differences in schooling, the authors hypothesize that better-educated children can improve parents’ dietary behaviors and nutrition through knowledge diffusion and support. The study addresses potential endogeneity between children’s education and parents’ diet by employing an instrumental variable approach using the staggered implementation of China’s Compulsory Education Law (CEL).

Prior research in developed countries shows education is positively associated with diet quality and healthier consumption patterns even after controlling for income. Higher education correlates with greater variety, more fruits and vegetables, and healthier food choices, while lower education predicts poorer dietary patterns. In China, education (especially among women) is linked to healthier dietary indices and improved nutrient intake. Education also exhibits a downward spillover: parents’ higher education improves children’s diets (e.g., greater intake of milk, fruits, vegetables, and key micronutrients). However, little is known about upward spillovers—whether children’s education affects parents’ dietary intake. Given that dietary knowledge can spread within families, the study investigates whether children’s higher education in rural China influences parents’ dietary nutrient intake.

Data: The study uses 2011 China Health and Nutrition Survey (CHNS) data covering provinces that reflect varied geography and development (Beijing, Hebei, Shanghai, Chongqing, Guangxi, Guizhou, Heilongjiang, Henan, Hunan, Jiangsu, Liaoning, Shandong). Individual dietary data are based on 24-hour recall. The sample focuses on rural parents aged 45+ to ensure children are old enough for their education to influence parents. After excluding missing data, N=1194. Measures: Food consumption (g/day) for nine categories—cereals; beans and bean products; vegetables; fruit; livestock meat; poultry; dairy products; eggs; fish/shrimp and other aquatic products—follows China Food Composition Table 2004 and Chinese Dietary Guidelines (CDG) 2007 classifications. Nutrient intake is derived from reported foods using China Food Composition Table 2004, focusing on energy (kcal) and macronutrients: carbohydrates, protein, fat (conversion coefficients: carbs 4 kcal/g, protein 4 kcal/g, fat 9 kcal/g). Diet quality is measured by the Dietary Balance Index (DBI-2007) with three indices: High Bound Score (HBS, 0–32, excessive intake), Low Bound Score (LBS, 0–72, inadequate intake), and Diet Quality Distance (DQD, 0–72, overall imbalance). Given data limits, condiments, edible oil, and drinking water are excluded. Controls: Demographics and socioeconomic variables—age, gender, marital status, educational level, family income, BMI—plus weekly hours of light, moderate, and heavy physical labor to capture energy needs. Empirical strategy: Baseline linear model regresses parents’ food consumption, nutrient intake, and diet quality on children’s educational level with controls. To address endogeneity (omitted variables and reverse causality), an instrumental variables (IV) approach using two-stage least squares (2SLS) is adopted. Instrument: exposure to the Compulsory Education Law (CEL) in rural China. Exposure varies by (a) child’s age at provincial CEL implementation (values 9 to 1 for ages 7 and below to 15, 0 if age ≥16 at implementation), and (b) provincial pre-policy educational development (Infl), measured as the proportion with junior secondary education in 1986 (higher values imply less policy impact). The first stage includes Expose, Infl, and their interaction (Expose*Infl) to predict children’s education. The instrument relevance is assessed via first-stage F-statistics (Cragg-Donald Wald F > 10 indicates no weak IV problem).

- Food consumption (IV-2SLS): Children’s higher educational level significantly reduces parents’ intake of cereals (coef −251.021, p<0.01) and vegetables (−622.201, p<0.01). It significantly increases consumption of animal-source foods: livestock meat (+240.017, p<0.01), poultry (+54.070, p<0.05), dairy products (+122.874, p<0.05), eggs (+66.087, p<0.01), and aquatic products (+30.103, p<0.10). Fruit and beans show no significant changes. First-stage F-statistic ≈21.946, suggesting strong instruments. - Total nutrient intake (IV-2SLS): Children’s higher education is associated with lower total energy (−722.858 kcal, p<0.05) and carbohydrate intake (−60.565 g, p<0.01), and higher protein (+6.176 g, p<0.10) and fat intake (+24.351 g, p<0.10). - Nutrient intake structure (IV-2SLS): Plant-based energy (−956.965 kcal, p<0.01) and plant protein (−3.159 g, p<0.01) decrease, while plant fat increases (+3.823 g, p<0.01). Animal-based energy (+225.699 kcal, p<0.01), animal protein (+9.637 g, p<0.01), and animal fat (+19.286 g, p<0.01) all increase, indicating a shift toward higher-quality protein sources. - Diet quality (DBI-2007, IV): Children’s higher education reduces excessive intake (HBS −1.463, p<0.10), inadequate intake (LBS −0.242, p<0.05), and overall imbalance (DQD −1.221, p<0.05), indicating improved diet quality. - Heterogeneity by child’s gender: Sons’ higher education has stronger and more pervasive positive effects on parents’ diet than daughters’. Sons significantly reduce cereals and vegetables and increase animal-source foods; they significantly reduce parents’ energy and carbohydrate intake and increase protein and fat; and they significantly lower HBS, LBS, and DQD. Daughters’ education shows smaller and fewer significant effects. - Heterogeneity by parent: Mothers’ diets are more responsive than fathers’. Children’s higher education significantly improves mothers’ food consumption (less cereals and vegetables; more animal-source foods), lowers energy and carbohydrate intake, raises protein and fat intake, and significantly reduces mothers’ HBS, LBS, and DQD. For fathers, effects are limited (notably reduced carbohydrate intake and some animal foods) and diet quality indices are not significantly affected. - Mechanism (IV): Children’s higher education significantly enhances parents’ dietary knowledge. Parents show greater agreement with correct statements regarding the benefits of beans, dairy, diverse diets, and high-quality protein from eggs/milk, and greater disagreement with incorrect statements (e.g., vegetables containing more starch than staples), consistent with knowledge diffusion improving dietary behaviors.

The findings support an upward intergenerational spillover of education on parental dietary behavior in rural China. Better-educated children likely transmit health and nutrition knowledge to parents (social network and knowledge diffusion mechanisms), improving parents’ perceptions and choices. Additionally, children with higher education may provide financial or material support, facilitating access to higher-quality foods. The results show meaningful shifts away from excessive carbohydrate dependence toward increased intake of protein and fats, especially from animal sources, and improved overall diet quality. Gendered cultural norms in rural China may explain stronger effects from sons and greater receptivity among mothers than fathers. These results align with theories of social support and health behavior change, and extend prior work by documenting diet-specific pathways in a developing-country context.

This study demonstrates that children’s higher educational attainment significantly improves rural Chinese parents’ food consumption patterns, nutrient intake, and overall diet quality, evidencing an upward educational spillover on dietary nutrition. Policy implications include: (1) relying solely on income growth is insufficient to correct dietary imbalances rooted in limited nutrition knowledge; children’s education can effectively raise parents’ dietary knowledge and improve diet structure; (2) strengthening rural basic education and strictly enforcing compulsory education can yield positive externalities for health and nutrition; (3) facilitating urban school access for rural children, consolidating compulsory education achievements, considering extension of compulsory education to include high school, and integrating comprehensive health and nutrition education into the national curriculum. Future research could explore longer-run health outcomes of parents, causal channels in more detail (knowledge vs. financial transfers), and replication across different cohorts or regions.

Self-reported 24-hour dietary recalls in CHNS are subject to recall and reporting biases (memory, knowledge, interview context), and not all foods (e.g., condiments, edible oils, water) are captured, which may affect precise measurement of intake and diet quality indices. These data limitations warrant cautious interpretation of effect magnitudes.