The reproduction of working class? Social mobility and the stratification of parenting practice in urban Chinese families

Parenting practice, an embodied cultural capital, is class based. The findings of this study showed that in urban Chinese families, the middle class was inclined to adopt the practice of concerted cultivation while the working class was inclined to adopt the practice of natural growth. However, those who were born in working-class families but were able to achieve upward mobility to the middle class were more likely to adopt similar practices to those who stayed in the middle class; those who were born in middle-class families but moved downward were able to retain some aspects of parenting practices similar to those who stayed in the middle class. The findings of the unbalanced reproduction of parenting practice do not support the argument that class boundaries have solidified in Chinese society.

In recent years, the question of whether the class boundaries in Chinese society have solidified and impeded intergenerational mobility has attracted both public and scholarly attention. There is a fierce debate on whether children from lower social class backgrounds can move upward. Some parents have spent the entire family wealth accumulated over generations in order to purchase property in a good school district due to anxiety that their children may "lose at the starting line". Parenting practice is an embodied cultural capital through which parents transfer their knowledge, habitus, and tastes acquired in school or society to their children (Lareau, 2002;Farkas et al., 1990). Immersed in everyday life, parenting affects children in subtle ways. It creates an "invisible" wall for intergenerational mobility by facilitating the transmission of advantages, perpetuating the intergenerational reproduction of social inequality, and thus stabilizing class boundaries.
There has been a growing body of literature on class-based Chinese parenting practices (Chen and Liu, 2013;Hong and Zhao, 2014;Hou, 2005;Wu and Zhang, 2016;Zhong, 2012), but very few studies have explored the relationship between social mobility and parenting practice. We argue that an examination of the influence of social mobility on parenting practices could reveal the mechanisms of social reproduction. The extent to which a family's class background affects parenting and whether the mobility experience alters parenting practice can inform the current conditions of class boundaries in China.
Using the 2010 Shanghai Household and Family Survey (SHFS 2010), this article seeks to answer two questions. First, is parenting practice class based in urban China? Second, can mobility experience weaken the impact of family class background on parenting practice? We found a pronounced class difference in parenting practice, but mobility experience can alter parenting practice: upward mobility erases the impact of class origin on parenting practice while downward mobility retains some aspects of the parenting practice inherited from class origin. The findings of the asymmetric reproduction of parenting practice do not support the assertion that the class boundaries in China have been solidified.

The stratification of parenting practice
Parenting practice refers to the knowledge, habitus, and tastes that parents pass on to their children, which facilitates children's communication with others and adaptation to the social environment (Lareau, 2002). In the book Unequal Childhood, Lareau (2010) proposes two models of parenting practice: "concerted cultivation" and "accomplishment of natural growth". In concerted cultivation, parents communicate with children, focus on the systematic development of children's social and language skills and problem-solving abilities, and systematically plan for children's extracurricular activities. In contrast, in accomplishment of natural growth, parents give commands to children, pay less attention to children's extracurricular activities, and leave most of the educational responsibilities to schools.
The way in which children are brought up affects the way they interact in school or in other social environments and thus influences their academic performance. The concerted cultivation approach helps children develop a sense of entitlement, know how to defend their own interests, and communicate with teachers and other adults with confidence (Calarco, 2011(Calarco, , 2014. In contrast, the natural growth approach develops a sense of constraints in children, pushing them to obey orders and feel distant from teachers and schools. Later research generally conceptualizes the two modes of parenting practice in three dimensions: parenting values, parent-child relationships, and skill building (see Table 1). In terms of parenting values, a concerted cultivation parent is authoritative while a natural-growth parent is often authoritarian or neglecting (Gerris et al., 1997;Pong et al., 2010;Willis, 1981). In terms of the parent-child relationship, a concerted cultivation parent focuses on frequent and intimate communication with a child while a natural-growth parent usually adopts a commanding tone when interacting with a child and thus has a distant parent-child relationship (Lareau, 2002). Finally, in terms of skill building, a concerted cultivation parent systematically invests in the child's extracurricular activities in order to cultivate noncognitive skills. A natural-growth parent, in contrast, focuses exclusively on cognitive skills. Parents have less interest in investing in extracurricular activities, and when such an investment is made, it is oriented towards cognitive skills (Bennett, et al. 2012). Because cultural capital is an important dimension of class status, parenting practice is class stratified as a means of transmission of cultural capital across generations. The middle class prefers the concerted cultivation approach while the working class is inclined to adopt a natural-growth approach. This is reflected in all three dimensions. In terms of parenting values, middle-class parents take advantage of their relatively high education to learn from the parenting ideas advocated by child development psychologists who advocate taking responsibility for children's growth (Gerris et al., 1997) and acting as authoritative parents (Pong, et al., 2010). Due to the relatively low education of working-class parents, it is often difficult for them to comprehend these new parenting ideas. They often have a sense of powerlessness when parenting (Lan, 2014) and are more likely to be authoritarian or neglecting parents (Willis, 1981). In terms of the parent-child relationship, the middle-class workplace is often creative, emphasizing self-management, collaboration, and articulated communication among colleagues. Following their workplace styles, parents emphasize frequent and intimate parent-child interaction. However, the working-class working environment is often hierarchical and routinized, requiring workers to obey orders. In terms of skill building, middle-class parents are familiar with the qualities needed to succeed in the white-collar workplace and thus emphasize cultivating noncognitive skills in their children (Friedman, 2013). Working-class parents, in contrast, have a relatively limited understanding of the white-collar working style, and they hope their children will move upward through their academic achievements and thus pay attention to academic performance (Lan, 2014).
Thus, this paper proposes the following hypotheses: Hypothesis 1a: In terms of parenting values, middle-class parents are more likely to be authoritative parents than working-class parents.
Hypothesis 1b: In terms of the parent-child relationship, middle-class parents are more likely to be more intimate with their children than working-class parents.
Hypothesis 1c: In terms of skill building, middle-class parents pay more attention to cultivating noncognitive skills in their children than working-class parents.
Due to data limitations, this paper can only test hypothesis 1b and hypothesis 1c. We encourage future studies to examine hypothesis 1a.

Social mobility and parenting practices
In addition to the parents' family class background, parenting practice may be influenced by parents' mobility experiences. Bourdieu's class trajectory theory (Bourdieu, 1980(Bourdieu, , 1984 highlights the intergenerational inheritance of parenting practices, arguing that families of origin determine parenting practices and that social mobility experience has little impact. However, this paper follows the social mobility theories of Sorokin and Lipset (Lipset & Bendix, 1959/1991Sorokin, 1959) and argues that parenting practice is the result of both class inheritance and reflection on mobility experiences. The mobility experience enhances parents' rational choice of parenting practice. In particular, the mobility direction influences parents' judgment of the efficacy of the parenting practices inherited from their original family, and the inheritance of parenting style can be asymmetric.
According to Bourdieu's class trajectory theory, the habitus stems from the class status of the families of origin and is difficult to change completely through social mobility. Class origin shapes linguistic and cultural tendencies in childhood, which has a profound influence on children's adult lives. Although habitus formation is a dynamic process, pre-existing habitus influences one's interpretation of new experiences. Class experience that fits with pre-existing habitus is more easily accepted and assimilated (Bourdieu, 1980(Bourdieu, , 1984. Thus, adults of the same class tend to retain the habitus of their family's origin to a large extent. Based on Bourdieu's class trajectory theory, some scholars argue that parenting practice is formed mainly through a process of reproduction. Although Lareau does not directly address this issue, she also argues that the parenting practice received in childhood provides "children who become adults later on with a sense of what is comfortable and natural for them" (Lareau, 2010: 272). This "comfortable and natural" approach allows parents to form their own parenting practice, passing on the class habitus to the next generation. Some empirical studies based on European and American societies also show that parents of the same class position do differ in their parenting practice, and this difference is largely due to different class origins (Irwin and Elley, 2011;Roska and Potter, 2011;Streib, 2013). Parents who came from working-class families and experienced upward mobility paid less attention to extracurricular activities than those who came from middle-class families. Upwardly mobile parents accentuate the upbringing in childhood and use this as a basis to criticize the parenting practice in mainstream middle-class families (Streib, 2013).
The literature on the reproduction of parenting practices may be helpful for understanding contemporary urban Chinese families. However, in past decades, Chinese society has experienced high volumes of upward mobility due to economic transformation and large-scale adjustment of the industrial structure (Li and Zhu, 2015). Parents now live in a social environment very different from their childhood. This may weaken the legitimacy of their own upbringing and challenge the class reproduction of parenting practice. Contemporary Chinese parents often adopt a narrative of a "generation gap" in describing their own upbringing. They emphasize that their childhood environment is very different from that of their children, criticize their own upbringing, and deliberately change their parenting practice (Yang and Tian, 2017). Therefore, this paper argues that contemporary Chinese parenting practices cannot entirely be the result of class reproduction. They are also shaped by parents' reflection on mobility experiences. Parents can inherit the parenting practice from their class origin, but they may also deviate from it.
The mobility experience can shape the extent to which parents' parenting practice deviates from their own upbringing. Social mobility is not only a structural movement; it also increases one's reflection on their class habitus (Sorokin, 1959). For parents who remained in the same class position, their class experience was consistent and thus may reinforce the class habitus. However, when parents experience social mobility, the experience of different class environments in childhood and adulthood may lead to a more open view of the different behaviors and values. Parents also have to make efforts to adapt to a new class environment so they may feel more convinced about their individual capability, which may provoke a reflection on their upbringing.
Further, the mobility direction may affect parents' identification with their class origin and consequently their judgment of the parenting practice in their upbringing. The asymmetry hypothesis of class identification argues that people's class identification is usually asymmetrical, adhering to higher class positions (Lipset andBendix, 1959, 1991). That is, upwardly mobile parents are more likely to identify with their class destination and thus to accept and learn the habitus associated with the class destination; conversely, downwardly mobile parents are inclined to identify with their class origin and tend to reject the habitus associated with their class destination, hoping their children will also move upward (Abramson and Brooks, 1971). Following Lipset's asymmetry hypothesis, this paper argues that the class reproduction of parenting practice varies according to the mobility direction in an asymmetric way. Upwardly mobile parents are more likely to identify with their class destination and thus are inclined to have a negative evaluation of the parenting practice in their upbringing. They may consciously rely on the knowledge and skills acquired in the workplace and from professionals to imitate and learn from the parenting practice of their class destination (Lan, 2014). However, parents who experience downward mobility, with the hope of achieving upward mobility for their children, may adhere to the parenting practice in their class origin (Chuang and Lin, 2017) and consciously resist the influences of their class destination (Rokska and Potter, 2011). Therefore, parents who experienced mobility in both directions may adopt the parenting practice of higher class positions.
Hypothesis 2: Parents who came from working-class families and experienced upward mobility are likely to adopt a similar parenting practice as parents who remained in the middle class, which is different from the practice of parents who remained in the working class.
Hypothesis 3: Parents who came from middle-class families and experienced downward mobility are likely to adopt a similar parenting practice as parents who remained in the middle class, which is different from the practice of parents who remained in the working class.

Data
This paper uses the 2010 Shanghai Household and Family Survey to examine class differences in parenting practice and how social mobility experience affects such class differences. The data were collected by Professor Liu Xin in the Department of Sociology of Fudan University, Shanghai, China. The survey used a multistage PPS sampling scheme of the adult population within the outer ring of Shanghai. (It used neighborhood committees as the primary sampling unit and then sampled households within each neighborhood committee. Finally, it randomly picked one household member aged from 18 to 70 in each household. The sample was self-weighted at the household level.) The full sample included 1181 respondents. The analytical sample is limited to respondents who had one child. After excluding missing values, the analytical sample size was 803 for the parent-child relationship question and 756 for the skill building question.

Dependent variables
Parent-child relationship The SHFS 2010 asked respondents to answer the question: "Are you close to your children?" The question was answered using a 7-point scale ranging from "very distant" (1) to "very close" (7). We considered a higher score to indicate a closer relationship and thus a higher level of concerted cultivation; conversely, a lower score indicated a more distant relationship and thus a higher level of natural growth.
Skill building The SHFS 2010 covered nine questions on skill building, and each was answered using a 7-point scale ranging from "strongly disagree" (1) to "strongly agree" (7) (please see Table 2). Factor analysis yielded three factors with eigenvalues greater than one. These three factors reflect the importance that parents place on noncognitive skills. The first factor, which is named the "autonomy" factor in this paper, was closely related to the questions "Children can try new things without fear of making mistakes," "Children should have confidence to express different opinions towards others," and "Children should be polite to people they don't like," reflecting the respondents' emphasis on nurturing their children's autonomy. The second factor, which is named the "emotional expression" factor in this paper, was closely related to the questions "Children should be allowed to express dissatisfaction through venting" and "Children can cry to express their feelings," reflecting the respondents' emphasis on nurturing their children's emotional expression. The third factor, which is named the "overall performance" factor in this paper, was closely related to the questions "Children cannot argue with the parents," "Being obedient is more important than having good academic performance," "It is more important to cultivate children's imagination than their academic performance," and "It is more important to cultivate children's imagination than their sport ability," reflecting the respondents' emphasis on nurturing their children's overall performance. For the three measures, a higher score indicated that parents paid greater attention to developing autonomy, emotional expression, and overall performance in children. That is, parents followed concerted cultivation to a greater degree. A lower score indicated that parents paid less attention to these aspects and thus followed natural growth to a greater extent.

Independent variables
The core independent variables of this study are social class and mobility experience. Social class was measured by parents' current occupations, which were classified into two categories: the working class and the middle class. Following Hong and Zhao (2014), the working class included service, sales, skilled, and unskilled workers, and the middle class included managers, professionals, technicians, and clerical staff. Social mobility experience was defined by a comparison of the parents' and their fathers' social class, which was classified into four types, including working class immobile (parents were working class and their fathers were working class), downwardly mobile (parents were working class but their fathers were middle class), upwardly mobile (parents were middle class but their fathers were working class), and middle class immobile (parents were middle class and their fathers were middle class).

Control variables
The control variables in this study included gender, age, having one child, education level (junior high school or lower, senior high or equivalent, college and above), and whether one had a Shanghai hukou.
The descriptive statistics of dependent, independent, and control variables are presented in Table 3.

Methods
The parent-child relationship was an ordinal variable, so ordinal logistic regressions were used. The factors of skill building were continuous variables and thus were analyzed using multivariate linear regressions. Table 4 reports the class differences in the parent-child relationship and skill building. In terms of the parent-child relationship, middle-class parents were 46% (e 0.380 ) more intimate with their children than working-class parents, which was statistically significant after controlling for other variables. Therefore, hypothesis 1b was supported.

Class difference in parenting practice
In terms of skill building, middle-class parents and working-class parents differ significantly in emotional expression and overall performance. Middle-class parents scored 0.252 higher in emotional expression and 0.171 point higher in overall performance than working-class parents. The results suggested a pronounced class difference in skill building. Middle-class parents paid more attention to noncognitive skills than workingclass parents. Hypothesis 1c was supported. An insignificant class difference in autonomy development was also found in other studies. Studies on Asian-American parents also showed that they were more authoritarian than white parents. In their parenting practice, Asian-American parents emphasized obedience and tended not to share decision-making processes with children (Chao, 2001;Kao, 2004). Table 5 reports the association between mobility experience and parenting practice. Models 1 and 2 show the results for the parent-child relationship, and models 3 to 8 report the results on skill building. The comparison between working class immobile and middle class immobile further supported hypotheses 1b and 1c. Except for autonomy development, parents who stayed in the middle class scored significantly higher than parents who stayed in the working class in the parent-child relationship, emotional expression development, and overall performance development. These findings further reinforced that parenting practice was class based.

Social mobility and parenting practice
Upwardly mobile parents were similar to parents who stayed in the middle class and were significantly different from parents who stayed in the working class. Without The standard errors are in parentheses + p<0.1,*p<0.05, **p<0.01, and ***p<0.001 (two-tailed test) Reference groups: a working-class parents, b female, c parents with multiple children, and d junior high or lower. e The model of the parent-child relationship shows the pseudo-R squared control variables, upwardly mobile parents scored 50% higher in the parent-child relationship than parents who stayed in the working class (model 1). Although the coefficient was no longer statistically significant after including control variables, it was still positive. This provided tentative support for hypothesis 2. In terms of skill building, the upwardly mobile parents scored 0.242 higher (p<0.05) in emotional expression and 0.213 points higher (p<0.05) in overall development than parents who stayed in the working class. Thus, parents who experienced upward mobility did not adhere to the parenting practice from their class origin but adopted the parenting practice in their class destination. These findings supported hypothesis 2. For parents who have experienced downward mobility, their parenting practice was not simply the result of class reproduction, as hypothesis 3 suggested, but rather it actually fell between the two classes. Their parent-child relationship was not significantly different from those who stayed in the working class or from those who stayed in the The standard errors are in parentheses + p<0.1,*p<0.05, **p<0.01, and ***p<0.001 (two-tailed test) Reference groups: a working-class parents, b female, c parents with multiple children, and d junior high or lower. e The model of the parent-child relationship shows the pseudo-R squared middle class; furthermore, their perceptions of skill building in their children were not significantly different from those who stayed in the working class or from those who stayed in the middle class 1 . This suggested that a downwardly mobile parent followed the parenting practice from their class origin to some extent. Therefore, hypothesis 3 had partial support. In summary, upwardly mobile parents had a similar parent-child relationship and skill building as parents who stayed in the middle class, but the relationship was significantly different from parents who stayed in the working class. Thus, parenting practice is not a process of reproduction and can be learned later. The parenting practice of downwardly mobile parents, however, was somewhere between the class origin and class destination. They were able to maintain some part of the practice from their class origin.

Robustness checks
The above analyses assumed a stable parenting practice or at least a stable class difference in parenting practice across generations. The assumptions did not hold in China. Chinese families have gradually shifted from the discipline of "learning to be human" to the "child-centered" parenting practice advocated by child development psychology, focusing on early education, emotional communication, and noncognitive skills (Xiong, 2008;Chen and Yang, 2011). For example, the traditional Chinese parent-child relationship was hierarchical-physical punishments such as palm striking and kneeling were allowed and even encouraged (Lin and Wang, 1995). However, contemporary Chinese parents tend to view physical punishment as problematic and irresponsible and prefer to use verbal communication and to refrain from using prescriptive words with children (Chen et al., 2000;Xu et al., 2005;Wu et al., 2002).
Since the data did not include data on the parenting practice of grandparents, the intergenerational stability of parenting practice cannot be directly tested. Therefore, we adopted an indirect approach with a cohort comparison before and after the Chinese economic reform. Since young adulthood (approximately 20 years old) is the most important period for the formation of personal values (Arnett, 2000), we compared the parenting practices of parents who were born before and after 1970 (see Table 6). Table 6 shows no significant differences between pre-70 and post-70 parents' skill building perceptions (i.e., autonomy, emotional expression, and overall performance). However, post-70 parents were closer to children than pre-70 parents. By further differentiating the sample into the working-class sample and the middle-class sample, we found that only middle-class parents showed a cohort difference in the parent-child relationship. Over time, middle-class parents have become more intimate with their children while working-class parents' relationships with children have not changed. Thus, the class difference in the parent-child relationship has increased since the reform, so it would be more difficult for upwardly mobile parents to acquire the practice of middleclass parents. This helps to explain why upwardly mobile parents still differ from those stayed in the middle class in terms of the parent-child relationship, as shown in model 2 of Table 5.

Discussion and conclusion
This paper examines the class differences in parenting practices and how mobility experience influences parenting practices in urban China. Parenting practice is an important mechanism for the intergenerational reproduction of social inequality, and class differences in parenting practice may prevent intergenerational mobility. However, if parenting practice does not entirely originate from the class origin and can be changed through social mobility experience, it may help upwardly mobile parents pass on their achieved advantages to their children, thus weakening class boundaries. The study had two important findings. First, there are significant class differences in the parenting practices of urban Chinese families in China. Middle-class parents adopt the parenting practice of concerted cultivation, but working-class parents adopt the parenting practice of natural growth. Compared to working-class parents, middle-class parents pay more attention to children's emotional expression and overall performance development, and they are closer to their children. Second, upwardly mobile parents learned their parenting practices in their class destinations. Upwardly mobile parents raised children in a way similar to those who stayed in the middle class but different from those who stayed in the working class. Third, downwardly mobile parents' parenting practice is somewhere between those of working-class parents and middle-class parents. They were able to maintain the practice from their class origin to some extent. It is possible that downwardly mobile parents, although preferring to maintain the parenting practices of their class origin, have difficulty actualizing it due to their limited family resources. These findings reveal that the intergenerational reproduction of parenting practices is asymmetrical.
This study has important implications for our understanding of the social class boundaries in China. The parenting practice of upwardly mobile parents suggests that once people achieve upward mobility, they can pass this advantage on to their offspring. Parents who experienced downward mobility, however, are able to maintain some parts of the parenting practice from their class origin, which may allow their children to move upward again in the future. The parenting practices of both upwardly mobile and downwardly mobile parents imply continuous intergenerational mobility in contemporary Chinese society. The class boundaries are dynamic rather than stabilized. As such, we do not agree that China's class boundaries have solidified (Sun, 2002(Sun, , 2008. Our results of class-based parenting practice pose an interesting contrast to those of Hong and Zhao (2014). Although both studies argue that the class boundaries in urban The standard errors are in parentheses. All models included control variables of sex, Shanghai hukou, having one child, and education level. Age was omitted due to its high correlation with cohort + p<0.1, *p<0.05, **p<0.01, and ***p<0.001 (two-tailed test) China have not yet solidified, their studies find no class difference in parenting values, whereas we find significant class differences in the parent-child relationship and skill building. We consider this discrepancy to be a result of different measurements. Hong and Zhao (2014) used the question "We should consult with our children on matters related to them, no matter how big or small," which is a question about parenting values. We measured parenting practice using the parent-child relationship and skill building. In fact, the two findings share some commonalities. The finding that middleclass parents were reluctant to communicate with their children or make collective decisions is very similar to our finding regarding no class difference in autonomy development. We likewise suggest that Chinese parenting values are more authoritarian than authoritative, regardless of social class. A study by Pong et al. (2010) also suggests that in the USA, college-educated Asian parents were less likely to share decisions with their children than parents of European descent. Scholars have proposed that the authoritarian nature of Chinese parenting originates from Confucius. Chinese families see upbringing as very important to passing the "Five Luns" and filial piety on children. Parenting means teaching children to "learn to behave" in a Confucian way, to honor their ancestors and to carry lineage responsibility. Discipline is the core to teach sons to obey fathers (Chao 2001;Lin and Wang, 1995;Xiong, 2008). However, discipline has gradually been replaced by "love education," a child-centered parenting practice that intensively invests time, money, and affection in children's development (Tao, 2018;Yang, 2018). The results of robustness checks, which shows that middle-class parents are closer to their children over time, also support this claim. Future research can examine the extent of the stratification and intergenerational changes in parenting practice in China with all three dimensions listed in Table 1.
Though the focus of this study is on occupational environments, we do not intend to undermine the impacts of other contexts on parenting. Schools are also critical in shaping parenting practices. Schools shape students' values, language habitus, and other characteristics (Jack, 2014;Khan, 2010;Horvat & Davis, 2011), which in turn influence the parenting practice of these students when they become parents. In addition, education is important for class mobility, which can indirectly affect parenting practice. Our results also highlight the importance of education in parenting practice. Compared to lower-educated parents, college-educated parents were closer to children and paid more attention to children's autonomy and emotional expression. However, the association between education and parenting practices weakens when mobility experience was included in the model, which suggests that education also affects parenting practice indirectly through class mobility. The expansion of higher education in recent years implies that education's impact on class mobility may vary by the credentials and the prestige of the school and students' major. It is fruitful to consider the heterogeneity of college education when analyzing parenting practices. Unfortunately, we cannot make further distinctions among the college-educated group due to data limitations. We encourage future studies to address this issue when data become available.
Marriage is another important context for parenting practice. First, marital status affects parenting practice. Divorced or single-parent families may find it difficult to adopt the parenting practice of concerted cultivation due to financial and time constraints (Berger and Mclanahan, 2015;Mclanahan and Pecheski, 2008). Second, a spouse's class background also influences how children are brought up. In particular, the mother's education and occupation affect family attitudes on extracurricular activities, cognitive development, and health (Munrane, 1981;Chen and Li, 2009;Potter and Roska, 2013). Finally, marital relationships also affect parenting practice. Higher marital quality is often associated with a closer relationship with children and less physical punishment, which leads to better cognitive development (Berger and Mclanahan, 2015). Due to data limitations, we cannot examine marital relationships and thus encourage future research to examine how marriage affects parenting practice in a more detailed way.
There are other limitations in this paper. First, the SHFS 2010 is representative of the Shanghai population, which cannot reflect the situation in other places in China. However, as the most populous and economically developed city in the country, class differences in parenting practices have implications for other regions in China as well. Second, due to data limitations, we cannot investigate people's parenting practices in all three dimensions and encourage future research to design studies to examine parenting practices in a more comprehensive way. Third, without information on grandparents' parenting practice, this paper only examines the class reproduction of parenting practice in an indirect way. Future surveys may cover both parents' and grandparents' parenting practices to provide a direct test of the class reproduction or transmission of parenting practices.