The Influence of Parental Education on First Marriages in China: The Role of Childhood Family Background

Abstract

China has implemented the universal two-child policy to promote fertility whereas the trends of later marriage and childbearing still prevail. The marriage timing of young adults has received considerable attention. However, studies tailored to the Chinese context remain limited. Using data from the China Labor-force Dynamic Survey 2016, we established bivariate and multinomial logistic regression discrete-time event-history models to examine the influence of parental educational attainment on their children’s first marriages. Not only the effects of parental education on their children’s likelihood of entry to first marriages and marriage postponement but also changes in these effects by sex, region, and across birth cohorts and the individual life course were analyzed. In general, there was a significantly positive relationship between parents’ educational attainment and the likelihood of their children’s entry into first marriages. However, it couldn’t explain variations about the timing of first marriages or marriage postponement. Analyses of subsamples revealed considerable heterogeneity in the effects of parental educational attainment by sex and region. These findings reveal the influence of childhood family background factors on marriage behaviors and provide a basis for predicting future marriage and childbearing trends in China. This study provides meaningful inputs and a rationale for amending the minimum legal marriageable age and for promoting marriage and births within Chinese marriage law.

Keywords

first marriage marriage delay educational attainment family background intergenerational relationship

Introduction

Marriage is considered a milestone in the transition to adulthood (South, 2001) and a key dimension of social networks in most societies (Brauner-Otto et al., 2020). Family is a fundamental social institution, and its formation during an individual’s life course is a key event. In Chinese, chengjia liye means getting married and launching a career. In the Chinese social context, family formation often refers to first marriages and does not include other forms of cohabitation. Hereafter, we use the terms “family formation” and “first marriage” interchangeably. Demographic transitions are closely associated with changes in marriage patterns. A key feature of the second demographic transition observed in developed countries has been a delay in family formation (Sobotka, 2008).

In contemporary China, rapid economic development and social transformations have led to changing attitudes toward family formation and marriage patterns. The crude marriage rate declined from 16.4‰ in 1990 to 9.0‰ in 2015. The mean age at first marriage rose from 22.87 years in 1990 to 26.43 years in 2015, with the mean ages of men and women at first marriage rising from 23.59 years to 27.21 years and from 22.15 years to 25.63 years, respectively (Office of the Leading Group of the State Council for the Seventh National Population Census OLGSCSNPC, 2022) (Figure 1). According to the Marriage Law and Civil Code of the People’s Republic of China promulgated in 1981 and in 2021, respectively, the minimum legal marriageable age is 22 years for men and 20 years for women. A delay in the first marriage by 3 years or more beyond the minimum legal marriageable age is considered a “later marriage” (i.e., no earlier than 25 years for men and 23 years for women). Later marriage is closely associated with later childbearing and featured in the family planning policy aimed at controlling births in 1973 (Zhang, 2017). China’s birth control policies, and especially its one-child policy, have prompted low fertility. The country’s total fertility rate (TFR) dropped from 2.29 in 1990 to 1.18 in 2010 and to 1.05 in 2015, far below the replacement rate. The mean age of first-childbearing mothers rose from 23 years in 1990 to 26.2 years in 2010 (Raymo et al., 2015).

Figure 1.

Age trends at first marriage, 1990–2015.

Although Chinese laws stipulate that children who are illegitimate/born out of wedlock enjoy the same rights as legitimate children and that no organization or individual may harm or discriminate against them, childbearing outside of marriage is discouraged. Compared with children born to married women, those born to single mothers typically have fewer social and financial resources (McLanahan, 1995). Childbearing outside of marriage is frowned upon within the traditional Confucian culture and is therefore uncommon in China. Although births among unmarried women are not recorded/registered in China, their numbers are very low and are thought to be even lower than those in Japan (2%) and Korea (1.5%) (Chu & Yu, 2010; Raymo et al., 2015). Therefore, the overwhelming majority of births are associated with marital fertility. In the Chinese context, marriage largely determines childbearing, and the timing of first marriages has a critical influence on fertility trends and patterns. Later marriage directly affects later childbearing. Therefore, the timing of first marriages has prompted increased concern following the implementation of the universal two-child policy to encourage childbearing at the end of 2015.

Children under the age of 14 years are almost entirely dependent on their families for financial and emotional support. Therefore, their family backgrounds are likely to shape their decisions to enter into first marriages, and the outcomes of these decisions lay the foundation for their subsequent life course (Mooyaart & Liefbroer, 2016). Family backgrounds during childhood strongly influence the timing of first marriages (e.g., Mulder & Smits, 1999; South, 2001; Wiik, 2009). Parents’ resources and educational attainment reportedly affect their children’s marital behavior (Axinn & Thornton, 1992; South, 2001; Wang & Wu, 2013). Educational attainment and occupational status have been used to measure families’ social and cultural resources in China (Lan, 2021; Li, 2003). Studies have found that young adults with highly educated parents generally enter their first marriages at older ages than their counterparts with relatively less well-educated parents (e.g., Mooyaart & Liefbroer, 2016; South, 2001). Following the expansion of college enrollment in China in 1999, the percentage of persons who completed their higher education rose substantially. Specifically, numbers of persons who completed junior college or attained higher degrees rose from 45.71 million in 2000 to 119.6 million in 2010 and 170.9 million in 2015 (National Bureau of Statistics (NBS), 2002, 2012, 2016). The expansion of college enrollment is considered a driving force propelling postponement of first marriages, but the effect of education reportedly drops with increasing age (Liu, 2016). Song and Fan (2017) found that young adults who had completed their higher education were more likely to be unmarried or to enter their first marriage at a later age. Previous studies have shown that parents’ educational attainment impacts positively on their children’ ages when entering their first marriages, whereas the number of siblings impacts negatively on age at first marriage (Wang & Wu, 2013; Yang & Li, 2018). Moreover, the effect of parents’ education on the timing of their children’s first marriages was found to decrease over the life course and across cohorts (South, 2001; Wiik, 2009).

Increasing attention has focused on marriage trends and timing following the implementation of the universal two-child policy aimed at promoting fertility and the concurrent trend of later marriage and childbearing. Strategies to address the later marriage and never-married states could alleviate the trend of later and declining childbearing. Marriage in China is strongly influenced by the traditional Confucian culture and by dramatic social changes. Moreover, compared with the timing and strength of first marriages in the West, those in China could be affected by birth control policies, childhood family backgrounds, and especially by parental educational attainment. Nevertheless, limited research has been conducted in the Chinese context, and control variables in existing studies have been inadequate (e.g., South, 2001), or not aligned with the Chinese context (e.g., Mooyaart & Liefbroer, 2016). Moreover, previous studies focusing on the effects of education on marriage have often overemphasized intergenerational similarities and have obscured the heterogeneity of these effects. Several questions arise. Would the influence of parents’ educational attainment on their children’s first marriages change over the adult life course of the latter? To what extent can the postponement of marriage be explained by the level of parental education? And to what extent do childhood family background factors account for the relationship between education and marriage postponement? In this paper, we first describe recent characteristics and trends of young Chinese adults’ marriages, attending especially to how educational differentials among parents affect their children’s decisions on whether and when to enter marriage after controlling for other family background factors and the respondents’ own characteristics. Differing from previous studies, we included more family background factors to examine the variability in the influence of educational effects. We examined not only the effects of parents’ education on their children’s first marriages but also the magnitude of these effects across birth cohorts, regions, and the life course, as well as sex-based variations, thus accounting for social changes.

Data and Methods

Data

We extracted data collected in 2016 for the China Labor-force Dynamic Survey (CLDS). The CLDS is a national, interdisciplinary, longitudinal survey conducted every 2 years across 29 provinces, municipalities under the central government’s direct rule and autonomous regions (excluding Hainan and Tibet) by Sun Yat-sen University. The survey respondents are stratified into the labor force, families, and communities. To facilitate sampling, computer-assisted personal interviewing technology is applied in the CLDS. A total of 399 communities/villages were covered in the survey conducted in 2016, covering 10,610 families and 21,086 individuals. Various issues relating to the labor force were covered, including family, education, work, and health. The samples in this study comprised men and women born between 1960 and 1996, all of whom were subject to the legal marriageable age under the Marriage Law promulgated in 1981. After the data had been pre-processed, the total number of observations in our study was 14,945.

Variables

Education is closely linked to income (Montgomery et al., 2000). Educational attainment has long been used as a proxy for economic resources in studies focusing on marriage issues (e.g., Sweeney, 2002). In China, educational attainment is closely associated with and often determines household income (Wu & Xie, 2003), and its measurement is more reliable and stable compared with measurements of income or reputation (Li, 2008). Generally, well-educated parents provide more cultural capital for their children, compared with the less well-educated counterparts (Hu & Wu, 2019). In this study, we considered parental educational attainment as a proxy for household socioeconomic resources and examined how it affected the likelihood of marriage and the timing of first marriages of offspring. The dependent variables, single or not (coded as 0 = married and 1 = single) and the level of later marriage (operationalized as the difference between the age at first marriage and the minimum legal marriageable age) were used to measure the likelihood of marriage and marriage delays, respectively. Previous studies found that the influence of mothers and fathers on their sons and daughters differed (Harris et al., 1998; Starrels, 1994). Therefore, the effect of education was differentiated by sex, considering possible parent–child dyads (e.g., father–son, father–daughter, mother–son, and mother–daughter). We considered parental educational attainment as a key independent variable for analyzing how childhood family background affected first marriages. Here, parental educational attainment was separated according to the sex of parents. Other controls for variables relating to respondents’ family backgrounds (at ages 0–14 years) were types of organizational employment held by parents; their occupational and marital status; Hukou (household registration); and respondents’ characteristics, including their sex, educational attainment, numbers of siblings, age/birth cohort, and political status.

Eight categories of educational attainment were considered in line with the Chinese education system: illiterate (completion of less than primary school education), completion of primary school, junior high school, senior high school, junior college or a post-secondary certificate, an undergraduate degree, a master’s degree, and a doctoral degree. Educational attainment was measured by the average number of years of education as constructed by Barro and Lee (2001), set at 3, 6, 9, 12, 15, 16, 19, and 22 years. As illiterate individuals may acquire knowledge through social learning (interpersonal communication) and observation, we evaluated this level as 3 rather than 0. We considered parents’ highest educational attainments.

Work units (danwei in Chinese), were categorized as follows: farmers (coded as 1), government, public institution/collective entity, or state-owned enterprise (SOE) (coded as 2), self-employed, and others (coded as 3). Individuals working in government, public institutions/collective entities, or SOEs meant that they held permanent jobs over their lifetimes unless they committed infractions or severely harmed their organizations, described as tie fanwan (or “iron rice bowl”) in Chinese. Occupational status was considered an important indicator for measuring parents’ social class and identity. We adopted the International Standard Classification of Occupation 1988 as the classification tool and generated new standard codes according to the International Socio-Economic Index of Occupational Status formulated by Ganzeboom et al. (1992). We referred to Li (2005) to adapt the scores for this index (84 = heads of governments/state agencies, party and mass organizations, public institutions/collective entities, or enterprises, 71 = professional and technical personnel, 63 = clerks and related personnel, 59 = business and service personnel, 54 = production and transportation equipment operators and related personnel, 45 = production personnel of agriculture, forestry, animal husbandry, fishery and water conservancy, 30 = other personnel). Hukou represented the registered geographic location of a person’s residence as well as their socioeconomic background. There are two types of Hukou in China: rural and urban, coded as 0 and 1, respectively. We also considered political status in line with China’s political system, which was coded as 1 = membership of the Communist Party of China and 0 = other parties/independents or nonparty political status. Marital status was coded according to four categories: first married (coded as 1), divorced, remarried, cohabitated and widowed (coded as 0). Respondents’ gender were coded as 0 = female and 1 = male. Birth cohorts were identified to study groups of people born in the same year(s) sharing common backgrounds and circumstances and experiencing historical events and the effects of social changes on their first marriages. Table 1 provides descriptions of all of dependent and independent variables considered in the study.

Table 1.

Descriptive Statistics.

Variables	Mean	SD	Min	Max	N
Single or not	0.332	0.471	0	1	14,945
The level of later marriage	3.597	3.352	0	34	9986
Father’s educational attainment	6.329	3.399	3	16	14,945
Father’s danwei when respondent at age 14	1.833	0.493	1	3	14,945
Father’s occupational status when respondent at age 14	50.213	12.230	30	84	14,945
Mother’s educational attainment	5.226	2.931	3	16	14,945
Mother’s danwei when respondent at age 14	1.907	0.424	1	3	14,945
Mother’s occupational status when respondent at age 14	47.774	8.492	30	84	14,945
Parental marital status when respondent at age 14	0.938	0.241	0	1	14,945
Sibsize	3.262	2.030	0	14	14,945
Respondent’s Hukou at birth	0.176	0.381	0	1	14,945
Respondent’s educational attainment	8.828	3.680	3	22	14,945
Respondent’s political status	0.085	0.278	0	1	14,945
Respondent’s gender	0.445	0.497	0	1	14,945
Respondent’s birth cohort	1969	12.873	1960	1996	14,945

Analytical Strategy

The main objective of this study was to examine the influence of the educational attainment of parents on the likelihood of their children’s entry into first marriages and marriage timings. We first presented the characteristics of the sample to be analyzed. Next, we established bivariate and multinomial logistic regression discrete-time event-history models to examine the influence of the education of parents on the likelihood of their children’s entry into (first) marriages and their timings. The data on marriage and families were deemed suitable for conducting an event-history analysis (Allison, 1984; Yamaguchi, 1991).

Given that single or not is measured as dichotomous variables relating to a single destination state, we considered logistic regression appropriate for estimating discrete-time event-history models (Allison, 1982). Hence, in our event-history–based analysis of the effect of parents’ education on the likelihood of their children entering first marriages, all young adults were considered potential candidates for first marriages until this event occurred. Because event-history models take account of right censoring, young adults were still considered single (unmarried) at the onset of the minimum legal marriageable age and were no longer observed (considered ineligible) after they got married. Thus, we overcame age censorship and event truncation problems stemming from cross-sectional survey data (Ryder, 1975). Discrete-time methods as applied to continuous-time processes are nevertheless popular in sociological research (Allison, 1982). Moreover, discrete-time and continuous-time models almost always yield similar results. If the probability of the occurrence of an event within each time interval is small, then the coefficients obtained from the discrete-time specification for most models will be similar to those obtained from the continuous-time specification (Allison, 1984; Petersen, 1993). We investigated potential differences in effects by estimating a series of models at the individual level.

Results

The Effects of Parents’ Education on the Likelihood of Their Children’s Entry Into First Marriages

Table 2 presents the estimation results of the bivariate logistic regression discrete-time event-history models used to examine the likelihood of entry into first marriages. Results are presented as coefficients, with positive coefficients indicating increased likelihood of being single/unmarried. Model 1, the basic model, included only parental educational attainment. Model 2 was established by adding respondents’ own educational attainment to Model 1. Controls included in Model 3 comprised parents’ danwei, occupational status, marital status, and numbers of respondents’ siblings derived from Model 2. Respondents’ political status was considered in Model 4. Parental education was associated with the likelihood of entry into first marriage at the 0.1% level. Higher levels of parental education were associated with a delay of the first marriages of offspring, who remained in a “still single/unmarried” state, which could lead to later marriage or non-marriage. Theoretically, a one-year increase in parental education was associated with a 6.5% increase in the odds of remaining single/unmarried (i.e., (e^0.063–1) *100%). The results were still striking after controlling for other family background variables and the personal characteristics of offspring. In particular, the influence of mothers’ educational attainment was strongly positive and robust, over 30% (34% and 32%, calculated from the ORs respectively) higher than those for the fathers’ or children’s own educational attainments shown in Model 4. Other control variables were also noteworthy. The adult children whose parents worked in government, public institution or SOE had a lower risk of being single/unmarried by 51%–58%, compared with those counterparts whose parents are farmers. Parents’ occupational status had a distinctly negative influence on their children’s first marriage and made them have lower odds of being single/unmarried (OR=0.99 and 0.97, respectively). The effect of parents’ marital status when their children were 14 years old was only evident at the 10% level. The intention to form families earlier was greater (at lower odds of being single/unmarried [OR=0.68]) among respondents with more siblings than among those with no siblings. Moreover, those who were members of the Communist Party of China (CPC) expressed more traditional views on the formation of unions, declaring the intention of early entry into first marriages upon reaching adulthood.

Table 2.

Results of the Regressions on the Influence of Parents’ Education on the Likelihood of Their Children’s Entry Into First Marriages.

	Model 1	Model 2	Model 3	Model 4
Father’s educational attainment	0.063^†	0.022*	0.095^†	0.097^†
Father’s educational attainment	(0.011)	(0.012)	(0.015)	(0.015)
Mother’s educational attainment	0.358^†	0.316^†	0.392^†	0.389^†
Mother’s educational attainment	(0.012)	(0.013)	(0.016)	(0.016)
Respondent’s educational attainment		0.123^†	0.089^†	0.110^†
Respondent’s educational attainment		(0.010)	(0.012)	(0.012)
Father’s danwei when respondent at age 14
Farmer (ref.)
Government, public institution or SOE			−0.734^†	−0.708^†
Government, public institution or SOE			(0.165)	(0.165)
Others			0.149	0.151
Others			(0.150)	(0.150)
Father’s occupational status when respondent at age 14			−0.016^†	−0.015^†
Father’s occupational status when respondent at age 14			(0.004)	(0.004)
Mother’s danwei when respondent at age 14
Farmer (ref.)
Government, public institution or SOE			−0.880^†	−0.869^†
Government, public institution or SOE			(0.182)	(0.182)
Others			1.072^†	1.045^†
Others			(0.156)	(0.156)
Mother’s occupational status when respondent at age 14			−0.033^†	−0.033^†
Mother’s occupational status when respondent at age 14			(0.005)	(0.005)
Parental marital status when respondent at age 14			−0.264*	−0.294*
Parental marital status when respondent at age 14			(0.145)	(0.146)
Sibsize			−0.394^†	−0.387^†
Sibsize			(0.025)	(0.025)
Respondent’s political status				−0.886^†
Respondent’s political status				(0.129)
Constant	−5.062^†	−5.741^†	−2.813^†	−2.978^†
Constant	(0.087)	(0.1100)	(0.282)	(0.285)
Nagelkere R sq	0.306	0.323	0.464	0.470
N	14,945	14,945	14,945	14,945

Notes: a. Dependent variables were “Single or not” and all estimated by the bivariate logistic regression discrete-time event-history analysis. b. Standard errors in parentheses. c. *, **, ***, and ^† represent significance at the 10%, 5%, 1%, and 0.1% level, respectively.

Table 3 shows variations in the effects of parents’ educational attainment on their children’s first marriages by sex and region. The same controls applied in Model 4 (see Table 2) were used. The regression results, by sex (Models 5 and 6) show that the effects of parental education levels were still strongly positive, and made their adult children incline to be single or unmarried. No sex differentials were found for the influence of fathers’ education levels, and the influence of mothers’ educational levels was greater on sons than on daughters. Clear sex differentials were seen for the effect of respondents’ own education. Whereas women’s education levels positively affected their entry into marriage, with higher education levels associated with a greater likelihood of marrying later or remaining single, men’s education levels did not affect their decisions to marry. The regression results differentiated by region (Models 7 and 8) showed that respondents from rural regions were affected not only by their parents’ education levels but also by their own significantly, with higher education levels associated with a higher likelihood of being single/unmarried. Respondents from urban regions were significantly and positively affected by their mothers’ educational levels, but their rural counterparts were mainly affected by their fathers’ and their own educational levels. Further analyses by sex and region (Models 9–12) showed that the influence of mothers’ educational levels on their children was significant at the 0.1% level (OR=1.30–1.73) irrespective of region or sex. Although the effect of male respondents’ educational levels on their own first marriages was not significant, differences between rural and urban regions still existed. Young adults with higher education in rural areas were more likely to choose to marry than their urban counterparts. The influence of fathers’ education levels on their daughters in urban and rural regions differed. Whereas fathers’ influence on their daughters was not significant in urban regions, it was significant at the 0.1% level for their rural counterparts. Women with higher education could postpone their own marriages, but the margin of effect depended on regions of origin. The effect of urban women’s education levels, indicated by OR comparison, was only 82% of that of rural women.

Table 3.

Regression Results by Sex and Region.

	Model 5	Model 6	Model 7	Model 8	Model 9	Model 10	Model 11	Model 12
	Male	Female	Rural	Urban	Rural male	Urban male	Rural female	Urban female
Father’s educational attainment	0.103^†	0.105^†	0.105^†	0.068**	0.094^†	0.077*	0.134^†	0.057
Father’s educational attainment	(0.020)	(0.024)	(0.018)	(0.027)	(0.025)	(0.040)	(0.030)	(0.038)
Mother’s educational attainment	0.459^†	0.323^†	0.447^†	0.269^†	0.548^†	0.260^†	0.358^†	0.276^†
Mother’s educational attainment	(0.023)	(0.024)	(0.020)	(0.028)	(0.029)	(0.039)	(0.031)	(0.038)
Respondent’s educational attainment	0.010	0.265^†	0.117^†	0.058*	−0.001	0.033	0.305^†	0.101**
Respondent’s educational attainment	(0.017)	(0.020)	(0.014)	(0.028)	(0.020)	(0.040)	(0.024)	(0.041)
N	6656	8289	12,313	2632	5515	1141	6798	1491

Notes: a. Dependent variables were “Single or not” and all estimated by the bivariate logistic regression discrete-time event-history analysis. b. Controls were the same as Model 4 in Table 2. c. Standard errors in parentheses. d. *, **, ***, and ^† represent significance at the 10%, 5%, 1%, and 0.1% level, respectively.

We conducted a further analysis, dividing respondents into three age groups: 22–29 years, 30–39 years, and 40 years or above to examine whether educational effects varied during different stages of the life course. Table 4 shows the results of the regression models, indicating that the effect of parents’ educational attainment differed according to stages in their children’s life cycle. Fathers’ education levels mostly influenced their sons aged 22–29 years, but this influence declined with age, even becoming insignificant. The relative log odds of their influence on their daughters assumed an inverted V-shaped curve, first increasing and then decreasing with a peak value for women aged 30–39 years (OR=1.04, 1.20, and 1.06). The influence of mothers’ educational levels on their children was significantly positive at the 0.1% level for all age groups. This result reflects the enduring cultural influence of mothers on their children’s entry into first marriage. Their influence on sons first decreased and then increased, whereas it first increased and then decreased for daughters.

Table 4.

Regression Results by Age Groups.

	22—29	Model 13	40 or elder	20—29	Model 14	40 or elder
		Male			Female
		30—39			30—39
Father’s educational attainment	0.176***	0.062	0.035	0.042	0.186***	0.058
Father’s educational attainment	(0.062)	(0.040)	(0.037)	(0.038)	(0.073)	(0.064)
Mother’s educational attainment	0.471^†	0.218^†	0.328^†	0.300^†	0.353^†	0.295^†
Mother’s educational attainment	(0.064)	(0.040)	(0.043)	(0.039)	(0.072)	(0.069)
Participant’s educational attainment	0.048	−0.093***	−0.314^†	0.274^†	−0.065	−0.132**
Participant’s educational attainment	(0.041)	(0.034)	(0.040)	(0.029)	(0.071)	(0.065)
N	700	838	5118	1146	1356	5787

The Influence of Parents’ Education on the Timing of Their Children’s Entry Into First Marriage

We used multinomial logistic regression discrete-time event-history models to examine the influence of parents’ education on the timings of their children’s entry into first marriage. We subtracted the minimum legal marriageable age from the age at first marriage to measure “the level of later marriage”, that is, the duration of marital delays, which was categorized as 0–2 years (the reference category), 3–5 years, and six or more years. Table 5 shows the results of the regression models. The results for Model 15, which only controlled for parental educational attainment, showed that the effect of parents’ education on their children’s later marriages was strongly positive at the 0.1% level. Respondents’ educational attainments were added to create Model 16. Fathers’ educational attainment had positive and significant effects on delays of 3–5 years and six or more years at the 5% and 1% levels, respectively. The effect of respondents’ own educational attainment was more apparent than that of their fathers and was significant positive at the 0.1% level. When other controls (the same as for Model 3 shown in Table 2) were added to create Model 17, only the effect of respondents’ own educational attainment was striking. Compared with the baseline group, the odds of later marriages by 3–5 years, and 6 years or more were increased by 8–11 % (OR=1.08 and 1.11, respectively) for a one-unit increase in respondents’ own educational level. Parental education could only explain a fraction of variations in marriage delays resulting from the addition of other childhood family background factors. Moreover, parents’ influence decreased with the increasing decision-making autonomy of their children. The inclusion of respondents’ political status (Model 18) hardly affected the results.

Table 5.

The Effects of Parents’ Education on Their Children’s Later Marriages.

	Model 15		Model 16		Model 17		Model 18
	3–5 years’ delay	6 years’ or more delay	3–5 years’ delay	6 years’ or more delay	3–5 years' delay	6 years’ or more delay	3–5 years’ delay	6 years’ or more delay
Father’s educational attainment	0.051^†	0.076^†	0.019**	0.032***	0.002	0.009	0.002	0.009
Father’s educational attainment	(0.009)	(0.010)	(0.009)	(0.011)	(0.010)	(0.011)	(0.010)	(0.011)
Mother’s educational attainment	0.0403^†	0.071^†	0.001	0.017	−0.013	−0.004	−0.012	−0.004
Mother’s educational attainment	(0.011)	(0.012)	(0.011)	(0.008)	(0.012)	(0.014)	(0.012)	(0.014)
Respondent’s educational attainment			0.103^†	0.144^†	0.076^†	0.104^†	0.074^†	0.102^†
Respondent’s educational attainment			(0.008)	(0.009)	(0.008)	(0.010)	(0.009)	(0.010)
N	9986		9986		9986		9986

Notes: a. Dependent variables were “the level of later marriage” and all estimated by the multinomial logistic regression discrete-time event-history analysis. b. Controls were the same as Model 1–4 in Table 2, respectively. c. Standard errors in parentheses. d. *, **, ***, and ^† represent significance at the 10%, 5%, 1%, and 0.1% level, respectively.

To examine possible changes in the effects of parents’ education induced by social changes in recent decades, we divided the samples into three groups according to birth cohorts: 1986–1996, 1976–1985, and 1960–1975. The regression results shown in Table 6 reveal insignificant changes in the effect of parental educational attainment with birth cohorts, whereas the effect of respondents’ own educational attainment remained significantly positive, with the strongest effects found in 1976–1985. The odds of later marriages by 3–5 years, and 6 years or more compared with the contrast group were increased by 22% and 35% (OR= 1.22 and 1.35, respectively) for a one-unit increase in educational level.

Table 6.

Regression Results by Birth Cohorts.

	Model 19		Model 20		Model 21
	1986 ‐ 1996		1976 ‐ 1985		1960 ‐ 1975
	3–5 years’ delay	6 years’ or more delay	3–5 years’ delay	6 years’ or more delay	3–5 years’ delay	6 years’ or more delay
Father’s educational attainment	−0.012	0.038	−0.023	−0.031	0.007	0.017
Father’s educational attainment	(0.041)	(0.070)	(0.025)	(0.028)	(0.011)	(0.013)
Mother’s educational attainment	0.005	−0.006	−0.016	−0.011	−0.017	0.001
Mother’s educational attainment	(0.041)	(0.069)	(0.026)	(0.030)	(0.015)	(0.017)
Respondent’s educational attainment	0.165^†	0.208***	0.196^†	0.303^†	0.052^†	0.068^†
Respondent’s educational attainment	(0.041)	(0.069)	(0.025)	(0.028)	(0.010)	(0.012)
N	558		1612		7816

Notes: a. Dependent variables were “the level of later marriage” and all estimated by the multinomial logistic regression discrete-time event-history analysis. b. Controls were the same as Model 4 in Table 2. c. Standard errors in parentheses. d. *, **, ***, and ^† represent significance at the 10%, 5%, 1%, and 0.1% level, respectively.

Further analysis of respondents’ sex and regions confirmed the robustness of the results (Tables 7 and 8) and revealed some interesting findings. For example, the influence of mothers’ educational attainment varied by sex (Table 7). Relative to delays of 0–2 years (the reference category), the effect of mothers’ educational attainment on their daughters was negative at the 10% level for marriage delays of 3–5 years and 6 years or more, with values being similar. The results showed that mothers’ educational levels could hasten their daughters’ marriages on the basis of existing later marriages, and the effect grew stronger with the daughters’ increasing age. The odds of being single/unmarried were decreased by 3% (OR=0.97) for an additional year increase in their mothers’ educational attainments. However, mothers’ educational levels had no significant effect on the later marriages of their sons, whose own educational attainments positively influenced later marriages (a delay of 3–5 years) at the 0.1% level, so that they remained unmarried. However, this effect weakened and became insignificant with a marriage delay of six or more years. The effect of daughters’ own educational attainment was significantly positive in the 0.1% level, as they married later, and it increased with age, with women continuing to delay marriage.

Table 7.

Regression Results by Sex.

	Model 22		Model 23
	Male		Female
	3–5 years’ delay	6 years’ or more delay	3–5 years’ delay	6 years’ or more delay
Father’s educational attainment	−0.004	0.023	0.006	−0.000
Father’s educational attainment	(0.015)	(0.016)	(0.013)	(0.016)
Mother’s educational attainment	0.003	0.030	−0.030*	−0.034*
Mother’s educational attainment	(0.019)	(0.021)	(0.016)	(0.019)
Respondent’s educational attainment	0.052^†	0.016	0.095^†	0.165^†
Respondent’s educational attainment	(0.014)	(0.016)	(0.011)	(0.014)
N	4194		5792

Notes: a. Dependent variables were “the level of later marriage” and all estimated by the multinomial logistic regression discrete-time event-history analysis. b. Controls were the same as Model 4 in Table 2. c. Standard errors in parentheses. d. *, **, ***, and † represent significance at the 10%, 5%, 1%, and 0.1% level, respectively.

Table 8.

Regression Results by Sex and Region.

	Model 24 Rural male		Model 25 Rural female		Model 26 Urban male		Model 27 Urban female
	3–5 years’ delay	6 years’ or more delay	3–5 years’ delay	6 years’ or more delay	3–5 years’ delay	6 years’ or more delay	3–5 years’ delay	6 years’ or more delay
Father’s educational attainment	0.006	0.018	−0.001	−0.012	−0.022	0.029	0.0476	0.050
Father’s educational attainment	(0.017)	(0.020)	(0.014)	(0.019)	(0.032)	(0.033)	(0.032)	(0.033)
Mother’s educational attainment	0.020	0.028	−0.028	−0.040*	−0.031	−0.000	−0.0480	−0.040
Mother’s educational attainment	(0.024)	(0.027)	(0.018)	(0.024)	(0.035)	(0.035)	(0.035)	(0.035)
Respondent’s educational attainment	0.045***	−0.011	0.088^†	0.154^†	0.103***	0.112***	0.130^†	0.174^†
Respondent’s educational attainment	(0.015)	(0.017)	(0.012)	(0.016)	(0.038)	(0.040)	(0.035)	(0.036)
N	3431		4702		763		1090

Notes: a. Dependent variables were “the level of later marriage” and all estimated by the multinomial logistic regression discrete-time event-history analysis. b. Controls were the same as Model 4 in Table 2. c. Standard errors in parentheses. d. *, **, ***, and † represent significance at the 10%, 5%, 1%, and 0.1% level, respectively.

We conducted further analyses by sex and region (Table 8). The increase of mothers’ educational levels reduced the likelihood of their daughters’ later marriages, although coefficients are not statistically significant except for rural women for a marriage delay of 6 years or more at the 10% level. In general, children’s increasing education levels were associated with later marriages. Apart from rural men, who delayed first marriages for 6 years or more (with their own educational levels having a negative but not significant effect on later marriage), they were significantly positive. The reason could be that rural men with more education are likely to improve their own socioeconomic status (SES) first before marrying. We also estimated models for heterogeneities of parental education on their children’s later marriages. However, we did not find such differences.

Discussion

Our aim was to examine how parents’ educational attainment influenced the likelihood of children’s entry into first marriages and later marriages and to what extent this influence varied by sex, region, and cohort. Parents’ educational backgrounds significantly influenced the likelihood of their children’s entry into first marriages, even after introducing respondents’ own education levels. Differing from Mooyaart and Liefbroer (2016), who reported small sex-based variations for the influence of parents’ education on children, we found significant variations. Fathers’ effects were not differentiated by sex, whereas mothers’ influence was greater for sons than for daughters. Moreover, the influence of mothers’ educational backgrounds on their children was significant and was not regionally differentiated, whereas fathers’ influence on their daughters varied by region. The influence of fathers’ educational levels was not significant on urban women but it was significantly positive for rural women at the 0.1% level.

Fathers’ influence varied according to their children’s ages, being greatest on sons aged 22–29 years and daughters aged 30–39 years. One possible reason is that young adults were less responsive to normative pressures exerted by their families (South, 2001), gradually forming their own worldviews based on independent reflection (Arnett, 2000), although effect curves declined in different shapes. Mothers’ influence was significant at all ages. Some childhood family background characteristics significantly affected the likelihood of respondents’ entry into first marriages, including parental educational attainment, danwei, parents’ occupational status, and marital status when respondents were 14 years old, numbers of siblings, and Hukou. We observed intergenerational patterns of family formation. Harmonious parental relationships, measured by parents remaining in first marriages, appeared to encourage their children’s transition to first marriages (Brauner-Otto et al., 2020). However, while the effect was greater than that reported by Cunningham and Thornton (2006), it was less than those reported by Kobrin and Waite (1984) and South (2001).

Parents’ educational backgrounds significantly influenced their children’s later marriages but became insignificant after controlling for other childhood family background factors and children’s individual characteristics. Fathers’ educational backgrounds did not significantly influence their children’s later marriages, whereas mothers’ had a negative influence relating to marriage delays of 3–5 years and 6 years or more only at the 10% level, but they had no significant influence on their sons’ later marriages. Further analysis revealed that maternal educational backgrounds were significantly negative for marriage delays of 6 years or more only for rural women at the 10% level. Previous studies found that the effect of parental education on the timing of family formation decreased over the life course and across cohorts (South, 2001; Wiik, 2009). Parents’ influence decreased with an increase in their children’s ages and autonomy, possibly because of ineffective parent–child communication or children’s increasingly independent reflections and viewpoints. Young adults with better educated parents were more likely to delay their first marriages than their counterparts with parents whose education levels were lower. This finding is consistent with that reported for Western countries (Axinn & Thornton, 1992; South, 2001). Children with well-educated parents often have abundant resources and cultural advantages. Whereas they may lack the motivation to leave home and form their own families, they may also be influenced by their family backgrounds, leading to delayed marriages until they attain their accustomed living standards. However, considerable discrepancies occurred under different conditions.

The effects of parental educational attainment varied, albeit insignificantly, among birth cohorts. As Mooyaart and Liefbroer (2016) observed, the impact of parents’ educational levels on their children’s union formation changes over the course of the children’s life cycle. Additionally, we examined intergenerational transmission of education by multiplying respondents’ educational level with those of their fathers and mothers, respectively. The results of the regression (Table 9) showed that parents’ education levels had a significant positive impact on their children’s education (at the 0.1% level), indicating intergenerational educational mobility. Well-educated parents are likely to spend more resources on their children than their less well-educated counterparts. Children with well-educated parents are more likely to demonstrate educational attainment, which contributes to later marriages.

Table 9.

Intergenerational transmission of education.

	Model 28		Model 29
	3–5 years’ delay	6 years’ or more delay	3–5 years’ delay	6 years’ or more delay
Father’s educational attainment × Respondent’s educational attainment	0.009^†	0.015^†
	(0.002)	(0.002)
Mother’s educational attainment × Respondent’s educational attainment			0.013^†	0.019^†
			(0.003)	(0.003)

Notes: a. Dependent variables were “the level of later marriage” and all estimated by the multinomial logistic regression discrete-time event-history analysis. b. Controls were the same as Model 4 in Table 2. c. Standard errors in parentheses. d. *, **, ***, and ^† represent significance at the 10%, 5%, 1%, and 0.1% level, respectively.

The robust result of the positive significance of mothers’ educational attainment on the likelihood of their children marrying confirms previously reported findings that mothers have a greater influence on their children within the family domain than do fathers (Aldous & Hill, 1965). A possible reason could be that mothers invest considerably more time and attention to childrearing compared with fathers (Mooyaart & Liefbroer, 2016). They would therefore be more concerned about their children’s marriage, and their education levels could influence their daughters to marry earlier on the basis of existing later marriages. According to socialization theory, parents’ influence on their children is differentiated by gender roles. Thus, the mother’s role in daughters’ socialization is more important than that of the father, whose role is more important in sons’ socialization (Aldous & Hill, 1965; Goslin & Aldous, 1969). Accordingly, not only mothers’ educational attainment but also their other characteristics would exert a profound influence on their adult children’s family formation. Therefore, we further examined the influence of the types of mothers’ danwei on their children’s later marriages (Table 10) and entry into first marriages.

Table 10.

The Influence of Mothers’ danwei on Children’s Later Marriages.

	Model 30		Model 31		Model 32		Model 33
	All samples		1986 ‐ 1996		1976 ‐ 1985		1960 ‐ 1975
	3–5 years’ delay	6 years’ or more delay	3–5 years’ delay	6 years’ or more delay	3–5 years’ delay	6 years’ or more delay	3–5 years’ delay	6 years’ or more delay
Mother’s danwei
Farmer (ref.)
Government, public institution or SOE	1.105^†	1.397^†	0.622	0.071	0.466	0.939**	1.205^†	1.493^†
Government, public institution or SOE	(0.135)	(0.147)	(0.699)	(1.042)	(0.387)	(0.399)	(0.151)	(0.165)
Others	0.435**	0.658***	0.130	−0.366	−0.523	0.424	0.922^†	1.001^†
Others	(0.184)	(0.207)	(0.451)	(0.808)	(0.372)	(0.380)	(0.271)	(0.296)
N	9986		558		1612		7816

Notes: a. Dependent variables were “the level of later marriage” and all estimated by the multinomial logistic regression discrete-time event-history analysis. b. Controls were the same as Model 4 in Table 2. c. Standard errors in parentheses. d. *, **, ***, and † represent significance at the 10%, 5%, 1%, and 0.1% level, respectively.

As Table 10 shows, mothers who were government, public institution, or SOE employees exerted much more influence on their children’s later marriages than those who were farmers or engaged in other types of employment. Female farmers had a preference for their children’s early marriages, whereas mothers working in governmental organizations, public institutions, or SOEs could have a more tolerant and rational attitude toward their children’s later marriages. It is noteworthy that mothers’ influence, on the basis of their organizational employment, on children within different birth cohorts varied greatly. For the 1960–1975 cohort, mothers with nonagricultural jobs, especially those working in government, public institutions, or SOEs, had significantly positive influences on their children’s marriage delays of 3–5 years and six or more years, and this effect increased with extended delays in the marriage timing. Compared with non-working mothers, working mothers have greater political status in China (Jin, 2013). This superior status and its possible SES advantages may have induced greater tolerance of their children’s later marriages, enabling the latter to spend more time searching the marriage market. For the 1976–1985 cohort, mothers’ employment in governmental organizations, public institutions, or SOEs only affected marriage delays of six or more years at the 5% level. Up to the 1986–1996 cohort, the effect progressively weakened to insignificant levels. Our cohort analysis revealed a dramatic change in the influence of mothers’ working characteristics on their children’s first marriages, with differences between social roles inside and outside the Chinese system gradually diminishing and leading to convergent influence.

Conclusions

Previous studies have not explored how family background factors, especially parents’ educational attainment, affect their adult children’s first marriages in China. We investigated how parents’ educational backgrounds affect their children’s first marriages and, further, how sex, region (urban/rural origins), birth cohorts, and life courses influenced this effect. Our study’s contribution to the existing literature on the effects of parents’ educational attainment on their children’s first marriages in the Chinese social context is twofold. First, we found that the influence of parents’ educational attainment was generally significantly positive regarding the likelihood of their children’s entry into first marriages, but it did not explain variations in the timings of their children’s first marriages or marriage postponement, which were explained by the children’s own educational attainment. Second, our subsample analyses revealed considerable variations in the effects of parents’ educational attainment by sex and region and some (insignificant) variations across birth cohorts. We operationalized the likelihood of entry into first marriage and the level of later marriage, and innovatively applied multiple models to analyze heterogeneity in the effects of parental education on family behavior.

Our findings have wider policy implications. They reveal the influence of childhood family background factors on the current situation of later marriages and childbearing, providing decision makers with a deeper understanding of the mechanism of intergenerational influence on marriage behaviors. Moreover, they enable the prediction of future marriage and childbearing trends in China. They offer constructive and meaningful inputs for future amendments of Chinese marriage law based on the possible adjustment of the minimum legal marriageable age and the promotion of marriage and childbirth via family environments. They can also guide the expansion of intergenerational and multidimensional research in the fields of education and marriage. With the increase of “leftover” women and men in China, Chinese marriage patterns need much more concern.

Limitations and Directions for Future Research

Our study had several limitations. First, although robustness checks of models designed from different perspectives support our conclusions regarding the influence of parents’ education on their adult children’s first marriages as well as variations by sex and region, the control variables within the models explain more correlations than causal links. Our results may be inadequate for drawing causal inferences. There are some unobserved factors/unmeasured characteristics that may contribute to the association between parents’ educational attainment and their adult children’s first marriages. Second, household incomes in CLDS were concealed, misreported, or underreported because of privacy concerns and were not an accurate or reliable indicator based on our preliminary data sorting. We only used educational attainment as a proxy for economic resources. As our study was limited to CLDS data, more registered data (national or census data) and further research on family formation would provide more precise analyses. CLDS 2016 provides a large cross-sectional data sample. However, multivariate causal analyses of the determinants and consequences of family formation require more longitudinal or retrospective data as there is no one-size-fits-all explanation notwithstanding our results. Third, we found that increasing educational attainment only explained a fraction of the later marriage trend. Differences in economic and cultural contexts may explain some of the variations in union formation and timing. Moreover, family formation is also affected by workplace and acquaintance networks, as colleagues, relatives, and friends form bonds through communication and interactive channels. Our future research will focus on the influences of Chinese traditional culture and social networks and their pathways of influence. A study on how the traditional Confucian culture and contemporary changes in this culture affect young adults’ decisions on marriage and childbearing also offers a promising research direction.

Footnotes

Acknowledgements

The author(s) would like to thank Sun Yat-sen University for data collection, as well as Mr. Jian Song, the anonymous referees and the editors for their constructive comments on earlier drafts.

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: This work was supported by the National Office for Philosophy and Social Sciences (17BRK028).

ORCID iD

Manyu Lan

References

Aldous

Hill

(1965). Social cohesion, lineage type, and intergenerational transmission. Social Forces, 43(4), 471–482, https://doi.org/10.2307/2574453

Allison

P. D.

(1982). Discrete-time methods for the analysis of event histories. Sociological Methodology, 13(1982), 61–98, https://doi.org/10.2307/270718

Allison

P. D.

(1984). Event history analysis: Regression for longitudinal event data. Beverly Hills: Sage.

Arnett

J. J.

(2000). Emerging adulthood. A theory of development from the late teens through the twenties. American Psychologist, 55(5), 469–480. https://doi.org/10.1037/0003-066x.55.5.469

Axinn

W. G.

Thornton

(1992). The influence of parental resources on the timing of the transition to marriage. Social Science Research, 21(3), 261–285, https://doi.org/10.1016/0049-089x(92)90008-5

Barro

Lee

J. W.

(2001). International data on educational attainment: Updates and implications. Oxford Economic Papers, 53(3), 541–563. https://doi.org/10.1093/oep/53.3.541

Brauner-Otto

S. R.

Axinn

W. G.

Ghimire

D. J.

(2020). Parents’ marital quality and children’s transition to adulthood. Demography, 57(1), 195–220. https://doi.org/10.1007/s13524-019-00851-w

Chu

C. Y. C.

R. R.

(2010). Understanding Chinese families. Oxford, UK: Oxford University Press.

Cunningham

Thornton

(2006). The influence of parents’ marital quality on adult children’s attitudes toward marriage and its alternatives: Main and moderating effects. Demography, 43(4), 659–672. https://doi.org/10.1353/dem.2006.0031

10.

Ganzeboom

Graaf

Treiman

(1992). A standard international socio-economic index of occupational status. Social Science Research, 21(1), 1–56. https://doi.org/10.1016/0049-089x(92)90017-b

11.

Goslin

D. A.

Aldous

(1969). Handbook of socialization theory and research. Chicago: Rand McNally.

12.

Harris

K. M.

Furstenberg

F. F.

Marmer

J. K.

(1998). Paternal involvement with adolescents in intact families: The influence of fathers over the life course. Demography, 35(2), 201–216. https://doi.org/10.2307/3004052

13.

(2019). Science or liberal arts? Cultural capital and college major choice in China. The British Journal of Sociology, 70(1), 190–213. https://doi.org/10.1111/1468-4446.12342

14.

Jin

Y. H.

(2013). Chinese working mothers in social changes. Academia Bimestrie, 2, 56–63. (in Chinese). https://doi.org/10.3969/j.issn.1001-9790.2013.02.007.

15.

Kobrin

F. E.

Waite

L. J.

(1984). Effects of childhood family structure on the transition to marriage. Journal of Marriage & Family, 46(4), 807–816. https://doi.org/10.2307/352528

16.

Lan

M. Y.

(2021). Relationship between Chinese women’s childhood family background and their fertility intentions under different policy conditions. Journal of Child and Family Studies, 30(12), 3016-3028. https://doi.org/10.1007/s10826-021-02095-y.

17.

C. L.

(2003). Socio-political changes and inequality of educational opportunity. Social Sciences in China, 3, 86–98. (in Chinese). http://www.en.cnki.com.cn/Article_en/CJFDTOTAL-ZSHK200303007.htm.

18.

C. L.

(2005). Cleavage and fragment: An empirical analysis on the social stratification of the contemporary China. Beijing: Social Sciences Academic Press (China). (in Chinese).

19.

(2008). Trends in educational assortative marriage: 1949–2000. Chinese Journal of Population Science, 3, 73–96. (in Chinese). https://doi.org/10.3969/j.issn.1000-7881.2008.03.009.

20.

Liu

(2016). The influence of college enrollment expansion on the first marriage age in China: Analysis based on census data. Population and Economics, 214(1), 19–28. (in Chinese). https://doi.org/10.3969/j.issn.1000-4149.2016.01.003.

21.

McLanahan

(1995). The consequences of nonmarital childbearing for women, children, and society. In, National Center for Health Statistics. Report to Congress on Out-of-Wedlock Childbearing. Hyattsville, MD: National Center for Health Statistics.

22.

Montgomery

M. R.

Gragnolati

Burke

K. A.

Paredes

(2000). Measuring living standards with proxy variables. Demography, 37(2), 155–174. https://doi.org/10.2307/2648118

23.

Mooyaart

J. E.

Liefbroer

A. C.

(2016). The influence of parental education on timing and type of union formation: Changes over the life course and over time in The Netherlands. Demography, 53(4), 885–919. https://doi.org/10.1007/s13524-016-0473-y

24.

Mulder

C. H.

Smits

(1999). First-time home-ownership of couples the effect of inter-generational transmission. European Sociological Review, 15(3), 323–337. https://doi.org/10.1093/oxfordjournals.esr.a018266

25.

National Bureau of Statistics (NBS)

China.

(2002). Tabulation on the 2000 population census of the People’s Republic of China. China Statistics Press.

26.

National Bureau of Statistics (NBS)

China.

(2012). Tabulation on the 2010 population census of the People’s Republic of China. China Statistics Press.

27.

National Bureau of Statistics (NBS)

China.

(2016). The national 1% population sample survey data in 2015. China Statistics Press.

28.

Office of the Leading Group of the State Council for the Seventh National Population Census (OLGSCSNPC) (2022). China population census yearbook 2020. China Statistics Press.

29.

Petersen

(1993). Recent advances in longitudinal methodology. Annual Review of Sociology, 19(1), 425–454. https://doi.org/10.1146/annurev.so.19.080193.002233

30.

Raymo

J. M.

Park

Xie

Yeung

W. J. J.

(2015). Marriage and family in East Asia: Continuity and change. Annual Review of Sociology, 41(1), 471–492. https://doi.org/10.1146/annurev-soc-073014-112428

31.

Ryder

(1975). Fertility measurement through cross-sectional surveys. Social Force, 54(1), 7–35. https://doi.org/10.2307/2576075

32.

Sobotka

(2008). Overview chapter 6: The diverse faces of the second demographic transition in Europe. Demographic Research, 19, 171–224. https://doi.org/10.4054/demres.2008.19.8

33.

Song

Fan

W. T.

(2017). The influence of higher education on youth’s first marriage and its gender difference. Youth Studies, 412(1), 1–8.(in Chinese). http://jtp.cnki.net/bilingual/detail/html/QNYJ201701001.

34.

South

S. J.

(2001). The variable effects of family background on the timing of first marriage: United States, 1969–1993. Social Science Research, 30(4), 606–626. https://doi.org/10.1006/ssre.2001.0714

35.

Starrels

M. E.

(1994). Gender differences in parent-child relations. Journal of Family Issues, 15(1), 148–165 https://doi.org/10.1177/019251394015001007

36.

Sweeney

M. M.

(2002). Two decades of family change: The shifting economic foundations of marriage. American Sociological Review, 67(1), 132–147. https://doi.org/10.2307/3088937

37.

Wang

Y. X.

(2013). Determinants of the age of first marriage: A study based on CGSS2006. Chinese Journal of Sociology, 33(3), 89–110. (in Chinese). https://doi.org/10.1186/scrt274.

38.

Wiik

K. A.

(2009). “You’d better wait!”—Socio-economic background and timing of first marriage versus first cohabitation. European Sociological Review, 25(2), 139–153 https://doi.org/10.1093/esr/jcn045

39.

Xie

(2003). Does the market pay off? Earnings returns to education in urban China. American Sociological Review, 68(3), 425–442. https://doi.org/10.2307/1519731

40.

Yamaguchi

(1991). Event history analysis. Newbury Park: Sage.

41.

Yang

K. W.

G. Q.

(2018). The influence of education acquisition on the age of first marriage. Population Journal, 40(6), 5–19. (in Chinese). https://doi.org/10.16405/j.cnki.1004-129X.2018.06.001.

42.

Zhang

(2017). The evolution of china’s one-child policy and its effects on family outcomes. Journal of Economic Perspectives, 31(1), 141–159. https://doi.org/10.1257/jep.31.1.141