How does education affect the health level of the elderly in China?

LI Xiang1 ZHAO Xindong2,3

(1.School of Economics and Finance, Huaqiao University, Quanzhou, China 362021)
(2.Institute of Quantitative Economics, Huaqiao University, Xiamen, China 361021)
(3.School of Statistics, Huaqiao University, Xiamen, China 361021)

【Abstract】Based on the CHARLS database, this paper makes a comprehensive and systematic analysis of how education affects the health level of the elderly in China. The results are as follows: First, education has a significant effect on the daily behavior and mental health of the elderly. Although the ability of older people to receive new knowledge has declined, the health effect of education can inspire us to strengthen education from the very beginning, so that the next generation can improve their health in old age. Second, there are gender differences, urban-rural differences and regional differences in the effect of education on the daily behavior and mental health of the elderly. Third, by constructing the structural equation model, it is found that the effect of education on the daily behavior ability of the elderly, in addition to the direct effect, can also be improved through the mediating effect of personal economic situation, while the mediating effect of education on the mental health of the elderly is mainly produced through daily behavior habits. Although education does not have a significantly direct impact on physical health, it has a significantly mediating effect through the habits of life. In addition, the mediating effect of education on the physical health and daily behavior of the elderly is also regulated by happiness. Based on the above conclusions, in addition to developing good health behavior habits and improving the awareness of potential chronic diseases, the government should also increase the spiritual comfort for the “empty nest” elderly, establish and improve the medical and health security system, and increase the participation rate of medical insurance.

【Keywords】 educational level; physical health; daily behavior ability; mental health; mediating effect;


【Funds】 National Natural Science Foundation of China (71973049) National Natural Science Foundation of China (71573093) Postgraduates’ Innovative Fund in Scientific Research of Huaqiao University (17011020003)

Download this article


    [1]. ① The chronic diseases mainly include: hypertension, dyslipidemia, diabetes or hyperglycemia, malignant tumors such as cancer, chronic lung diseases such as chronic bronchitis or emphysema, liver diseases, heart diseases, stroke, hydronephrosis, gastric disorders or digestive disorders, mental or psychiatric disorders, memory-related diseases (such as dementia), arthritis or rheumatic disorders and asthma. [^Back]

    [2]. ① The ten items used to construct variable CES_D. The two positive items: (1) I felt hopeful about the future. (2) I was happy. The eight positive items: (1) I was bothered by things that usually do not bother me. (2) I had trouble keeping my mind on what I was doing. (3) I felt depressed. (4) I felt that everything I did was an effort. (5) I felt fearful. (6) My sleep was restless. (7) I felt lonely. (8) I could not get “going." [^Back]

    [3]. ② The total scores of the depression scale CES_D range from 0–30. The higher the score is, the higher the tendency of depression is. Generally, a score of 0–10 indicates good mental health, whereas a score between 11 and 30 indicates some levels of depression. [^Back]

    [4]. ③ The sample data of education levels consist of a range of answers from the original questionnaire, including no education, traditional Chinese private school, kindergarten, preschool, primary school, junior high school, senior high school, secondary/vocational/skilled worker school, junior/vocational college, undergraduate education and graduate education. Due to the small sample size of traditional Chinese private school, kindergarten and preschool, these samples are regarded as having no education experience in this study, that is, having 0 year of education. [^Back]

    [5]. ④ The respondent is regarded as having a poor level of fitness if one of the following situations was reported: having school suspensions/staying in bed for over one month/being hospitalized for over one month/being hospitalized for over three times in a year for health reasons before the age of 15; otherwise, the fitness level is good. If one of the following situations was reported: staying in bed for over one month/being hospitalized for over one month/being hospitalized for over three times in a year/leaving work for over one month for health reasons after the age of 15, the fitness level is considered poor; otherwise, it is good. [^Back]

    [6]. ⑤ We construct dummy variables indicating the respondent’s experience working for government departments, public institutions, non-profit organizations, enterprises or other organizations. [^Back]

    [7]. ⑥ Social activities include: (1) visiting and socializing with friends; (2) playing mahjong, chess, cards and going to the community activity room; (3) offering help to neighbors and relatives or friends who live away; (4) dancing, exercising, qigong practice, etc. (5) participating in group or organizational activities; (6) doing volunteer or charity work; (7) taking care of patients or disabled persons who live away; (8) attending school or training programs; (9) investing in stocks (funds and other financial securities); (10) going online; and (11) other social activities. Similarly, the answers are assigned corresponding values to obtain a total score. [^Back]

    [8]. ① Detail regression results are available in the “difference analysis of the effect of education on health” section of the working paper of this study. [^Back]

    [9]. ② Detail regression results are available in the “heterogeneity analysis of the effect of education on the health of the elderly by age group” section of the working paper of this study. [^Back]

    [10]. ① The initial structural equation model is available in the working paper of this study. [^Back]

    [11]. ② First, “0, 1” means that 0 is the residual and 1 is the set parameter; and the parameters are set as 0 for latent variables (personal economic situation and lifestyle habits) and 1 otherwise. In Figure 2, for 0 and 0 in “0, 1,” the former is the set parameter value and the latter is the set residual value. Second, the parameters of the structural equation model can be set as either fixed parameters or free parameters. Fixed parameters are not estimated from data and are usually set to 0 or 1 for model identification purposes. [^Back]

    [12]. ① The regression results are available in the “results of the regression of education level on lifestyle habits and personal economic situations” section of the working paper of this study. [^Back]


    [1] Chang, Y. & He, Z. China Sport Science (体育科学), (2): 96–100 (2002).

    [2] Cheng, L., Zhang, Y. & Shen, K. China Economic Quarterly (经济学(季刊)), (1): 305–330 (2015).

    [3] Li, J. & Liu, B. Population & Economics (人口与经济), (6): 1–11 (2015).

    [4] Liu, C. & Wang, L. Chinese Journal of Population Science (中国人口科学), (5): 40–50 (2017).

    [5] Liu, L., Xu, H. & Song, S. Chinese Journal of Clinical Psychology (中国临床心理学杂志), (5): 636–638 (2009).

    [6] Liu, X., Luo, X., Zhang, L. et al. Journal of Chengdu Sport University (成都体育学院学报), (1): 79–82 (2010).

    [7] Liu, Y. Quality of Aged Care: Quality of Life Assessment and Insurance System for the Elderly Population in China (养老质量测评: 中国老年人口生活质量评价与保障制度). Beijing: The Commercial Press (2007).

    [8] Wen, Z. & Ye, B. Advances in Psychological Science (心理科学进展), (5): 731–745 (2014a).

    [9] Wen, Z. & Ye, B. Acta Psychologica Sinica (心理学报), (5): 714–726 (2014b).

    [10] Wu, H. & Jia, Y. Population and Development (人口与发展), (1): 66–73 (2017).

    [11] Ye, X. & Liang, W. Education & Economy (教育与经济), (3): 68–76, 96 (2017).

    [12] Zhu, C. & Yang, Y. Journal of Agrotechnical Economics (农业技术经济), (12): 76–87 (2017).

    [13] Brunello G, Fort M, Schneeweis N, et al. The causal effect of education on health:what is the role of health behaviors? Health Economics, 2016, 25 (3): 314−336.

    [14] Conti G, Hansman C. Personality and the education-health gradient:A note on “Understanding differences in health behaviors by education”. Journal of Health Economics, 2013, 32 (2): 480−485.

    [15] Cutler D M, Lleras-Muney A. Understanding differences in health behaviors by education. Journal of Health Economics, 2010, 29 (1): 1−28.

    [16] Katz S, Ford A B, Moskowitz R W, et al. Studies of illness in the aged: The index of ADL: A standardized measure of biological and psychosocial function. The Journal of the American Medical Association, 1963, 185 (12): 914−919.

    [17] Leopold L, Engelhartdt H. Education and physical health trajectories in old age: Evidence from the survey of health, ageing and retirement in Europe (SHARE). International Journal of Public Health, 2013, 58 (1): 23−31.

    [18] Muszalik M, Dijkstra A, Kędziora-Kornatowska K, et al. Independence of elderly patients with arterial hypertension in fulfilling their needs, in the aspect of functional assessment and quality of life (QoL). Archives of Gerontology and Geriatrics, 2011, 52 (3): e204−e209.

    [19] Ranchor A V, Bouma J, Sanderman R. Vulnerability and social class: Differential patterns of personality and social support over the social classes. Personality and Individual Differences, 1996, 20 (2): 229−237.

    [20] Smith J P, Goldman D. Can patient self-management explain the health gradient? Goldman and Smith (2002) revisited: A response to Maitra. Social Science & Medicine, 2010, 70 (6): 813−815.

This Article


CN: 31-1012/F

Vol 46, No. 03, Pages 139-153

March 2020


Article Outline


  • 1 Introduction
  • 2 Theoretical analysis and research hypotheses
  • 3 Research design and data sources
  • 4 Empirical analysis
  • 5 Analysis of the mediating effect of education on health
  • 6 Conclusions and implications
  • Footnote