Favouring Middle-and Upper-Class Students? The Structure and Process of Attending China’s Selective Universities

Research suggests the increasing influence of family socioeconomic status, as measured by parents’ income and occupations, in access to Chinese higher education. Yet, the literature remains inconclusive about the extent to which the social background of rural and urban students is associated with academic and social performance at elite universities. We address this limitation by looking at the academic and social success of representative samples of first-and second-year students enrolled at four Chinese elite universities. Our aim is to understand the characteristics that students from both urban and rural environments bring with them and how those characteristics bear on academic and social performance in university. We found an overrepresentation of students from middle-and upper-class backgrounds in both urban and rural student groups. The fact that the process indicator of cultural capital has a direct association with social success suggests students from urban areas exhibit traits valued in the selective university environment.


Introduction
The number of Chinese higher education institutions has increased from 1,022 in 2008 to 2, 651 in 2017, accompanied by an increase in the gross higher education enrolment rate from less than 9.76% to 45.7% (MOE, 2018). Yet, research suggests students from rural communities, which represent half of the population, are grossly underrepresented, especially in top-tier institutions (Li, 2015). Since the 1990s, researchers saw a gradual decline in the proportion of rural students in China's most selective universities. The proportion of rural students in Peking University, for example, has fallen from around 20% in the end of the 1990s to around 12% in 2012 (Sun, 2017). To offset this problem, the government introduced quota initiatives to increase recruitment of students from designated poor rural communities (Xie, 2015). Yet, a question remains as to whether the policy has brought greater equity in higher education.
Decades of market reform, including a process of de-collectivization and privatization, have stratified both urban and rural society. This has exacerbated differences in wealth, power and prestige among social groups (Bian, 2002;Davis & Feng, 2009). These wealth gaps add complexity to the analysis of access and equity. Research suggests an increasing influence of family socioeconomic status, as measured by the impact of parents' occupations and income on higher education access and equity. Wang (2015), for example, notes that students from the middle and upper social backgrounds, including families of professionals and government cadres, have greater access to the most selective universities. Controlling for the effect of socioeconomic status, Chen (2015) argues that rural students from villages and small towns outperformed their counterparts from county-level cities and county-level towns in gaining access to Project 211 universities 4 . This suggests that rural parents from upper and middle strata families are more able to capitalize on their family resources and to gain an edge over their rural counterparts (Author, 2016). It also suggests that the rural-urban dichotomy in explaining admission to university has shifted more toward income and proximity to urban life.
These findings suggest a mixed influence of factors amidst a newly evolving social structure (Cao, 2019). It also provides sound theoretical reasoning for expecting an underrepresentation of rural students, from more disadvantaged backgrounds, in China's most selective universities. Yet, the literature remains inconclusive about the composition and characteristics of the rural student population in those selective colleges and universities. Studies of students in those elite institutions lack data on family structure, socioeconomic status, and the types of schools they attended before entering colleges and universities. In short, the traits of the rural students are still not well understood. Beyond that, there is a need for a systematic empirical study on how students from different types of rural backgrounds negotiate the academic and social currents in elite university environments, in comparison to their urban peers.
In this paper, we address this limitation by looking at the academic and social success of representative samples of students entering four elite universities in China. We do this to understand the traits that urbanized and ruralized students bring with them into tertiary education, and how these bear on academic and social performance in university.

Chinese Context
Within less than 20 years, Chinese higher education is becoming a high participation system (HPS), already having in 2016 a Gross Tertiary Enrolment Ration (GTER) of 42.7%, and now approaching 50% in 2020 (Sun, 2017). This was the result of the interplay of several social factors. Driven by the 1998 Asian financial crisis, China launched an unprecedented expansion of higher education. This was first an effort to keep more young people out of the labor market until the crisis wound down, and, second, to stimulate the economy by responding to pent up demand from large-savings households by introducing a cost-sharing fee-paying model (Mok, 1999). This was encouraged by human capital-oriented policy makers (Xie, 2016). It also coincided with a knowledge economy discourse advanced by the World Bank (World Bank, 2000). By 2010, further expansion aligned with an economic restructuring move from a labour intensive, low investment, export-oriented system to a high-tech manufacturing and service-based economy that relies more on domestic consumption. World-class universities and mass higher education became essential for making the transition along with the political agenda in building a strong and modern national state (Xie, 2016). A rapidly growing urban middle class sought universities for their children that would confer status culture and sustain their social mobility (Farrell et al., 2006). Credentialism grew in a competitive labour market that had to absorb upwards of eight million college and university graduates each year (Altbach and Umakoshi, 2005;Collins, 1979;Marginson, 2011).
Elitism has long been part of Chinese higher education system, even in the socialist era. In 1954 the Chinese government named and generously funded two key universities (Tsinghua University and Harbin Institute of Technology) as a strategy of building a first-class university to train political and military leaders. Different policies followed in assigning a status of key university in the higher education system. One aim was to improve the quality of the higher education sector, and another was to build a few world-class universities. Several national excellence initiatives to identify world-class university followed in 1993, 1998, and 2015. The designated universities received a major government investment which made them substantially outperformed their competitors in the higher education system in terms of the number of faculty with Ph.D. holders and research output (Johnes & Li, 2008).
These exclusive universities only admit a tiny fraction (under 2% on average) of the high school graduates, in China (Cao, 2019). It is unsurprising that rising income inequalities (across regional, urban-rural, income, gender and ethnicity) amid rapid economic growth has led to concerns about patterns of access to elite universities. While rapid expansion has led to increased access to higher education by rural students, structural inequality has become to solidify (Li, 2015). For rural secondary school graduates born in the 1970s, gaining access to higher education was nearly similar to their urban counterparts. This began to change with the cohort born in the 1980s. Compared to their counterparts in urban areas, the rural secondary school graduates have a markedly less chance than their urban counterparts of gaining access to selective colleges and universities (Li, 2010).
In a system still influenced by its Confucian heritage, the public believe higher education should be a great equalizer for a better society. Admission to higher education institutions is almost entirely based on academic criteria, heavily determined by one national examination commonly known as the gaokao. To ensure that rural students have an equal chance to gain a high enough score on the gaokao, reforms were initiated beginning in 2000 to improve the quality of rural schools. A new system of financing rural education (Notice on the Deepening of Reforms on Mechanisms for Guaranteeing Funding for Rural Compulsory Education) was introduced in 2005, which exempted rural students from paying various schools related and textbook fees. Cost sharing by central and provincial governments helped to improve the recruitment of qualified teachers in village and small township schools. Six of the country's leading Normal Universities (that emphasize teacher education) initiated free tuition and special allowances for graduates who would teach in rural schools (Hallinger & Liu, 2016).
Measures were also instituted to decrease the gap between the prosperous eastern coastal regions and the less developed central and western areas of the country. A Collaborative Plan to Improve College Enrolment in Central and Western Regions was introduced in 2008, and universities in the eastern coastal cities were asked to set up quotas (35,000 in 2008 and 210,000 in 2016) of students from 14 provinces in the central and western areas. In 2012, the Chinese central government started a new policy to Revitalize Higher Education in the Central and Western Regions 2012-2020 by investing 10 billion Renminbi (RMB)in 100 higher-education institutions to help repair infrastructure, improve teaching and research capacity, and provide more places for students from the central and western areas. Moreover, a Scheme for Enrolling Students from Poor and Rural Areas was introduced to increase rural students' participation in leading universities. The number of rural freshmen in those first-tier universities was reported to have increased by 10% in 201210% in , 8% in 201310% in , and 11.4% in 201410% in (MOE, 2014.

Literature Review and Theoretical Framework
Without a doubt, the higher education system has pushed back on the increased inequality by admitting more students from diverse social backgrounds, including ethnic minority students from among a minority population of 120 million, whose minority status grants them extra points added to their gaokao scores. While traditionally underrepresented groups have increased their access to higher education, top tier admission has remained unequal (Mok, 1999). In a neoliberalism discourse, individuals and families are assumed to take on a significant amount of responsibility for the cost in higher education (Ball, 2016). Reforms in higher education admissions have highlighted the family's role in preparing their children for the university's independent recruitment process, which gives an advantage to those students who have had intensive parenting in their early years (Liu & Pensiero 2016).
As already noted, opportunities for university attendance for the rural cohort born in the 1940s, 1950s, 1960s, and 1970s were found to be similar to their urban counterparts after controlling for the effects of gender and the father's occupation and education level. Yet, for the cohort born in the 1980s and 1990s, urban students outperform their rural counterparts 1.7 times more effectively in gaining admission to higher education (Li, 2015;Wu and Li, 2017). The fact that some studies also argue that gap have been narrowing in the past 30 years has intensified the debate (Wang, 2015). Furthermore, for those arguing for a decreasing gap in terms of access to higher education, they still claim that the gap between those from such traditionally advantaged social groups as the cadre, the professional and the new economic elite, and those disadvantaged social groups are still there (Wu, 2017;, a trend revealed in maximum maintained inequality theory in other social and cultural context (Liu & Pensiero, 2016).
Furthermore, in the soon-to-be HPS, with a continuing elite sub-sector, the binary structure has become ternary with high-value inclusion, low-value inclusion, and exclusion (Marginson, 2011). The world-class university movement tends to promote vertical stratification with middle class families gaining a positional competitive advantage (Marginson, 2016). Does the differentiation of value favor those "born into privilege"? Wang (2015) argues that students from those higher socioeconomic backgrounds enjoy substantial advantages in going to the elite sector of HEIs. Chen (2015) examined the differentiated patterns of access by students from different socioeconomic backgrounds to different higher education sectors and argued that students from cities and those from families of higher socioeconomic status outperformed their peers in selecting and gaining access to the elite HEI sector. Other studies yield similar findings. Liu and his colleague (2015) find a similar pattern when examining the influence of fathers' party membership on the access had to elite selecting universities. They point out that students coming from communist party members' families are more representative in the upper tier (Projects 211 & 985) 5 universities, which suggests an importance of political capital (Liu & Yao, 2015).
The inevitable family-based inequalities in terms of economic, social and cultural resources affect access to selective HEIs. They put those families with prior social advantages in a better position to translate earlier educational achievement into more successful life trajectories. Therefore, an examination of the horizontal inequity in higher education should not only look at the socioeconomic difference in access but also the differentiated pattern of success in and through the higher education experience. Those socioeconomically advantaged families tend to leverage their initial success into further advantages in competitive settings. Sociological studies surmise that young people from lower socioeconomic backgrounds are more likely than those of better standing to encounter barriers to full integration into an elite selecting environment (Granfield, 1991;Thiele et al., 2017). In some countries, they achieve less in learning and spend more time in working offcampus (Jack, 2016). Students from socio-economically disadvantaged backgrounds arrive at the university with more uncertainty. They tend to feel like cultural outsiders (Ostrove & Long, 2007;Lehmann, 2014). Their dispositions, which have been shaped heavily by the social milieu from which they come, leave them with a higher potential to face confusion, conflicts, and struggle in the elite HEIs, a social environment which is unfamiliar to them (Bourdieu, 1977;1984).
To investigate the vertical differences in success in and through higher education, this study looks at the academic and social experience of rural students at China's elite universities. Those theoretical bases mentioned above indicate the possibility of generating specific hypotheses. As rural students have less access to capital in different forms, such as human, economic, and cultural capital, these can lead to lower levels of academic and social success in elite selecting universities. To investigate the patterns in academic and social experience of rural students at elite universities, the study draws from the tradition of sociological studies in which family background effects on educational inequalities are decomposed into "structural" and "process" dimensions (Coleman, 1990). While the structural dimension of family influence refers to those factors related to the jobs, income, and education of the parents, the process dimension of family influence speaks of factors indicating parents' participation in and ability to use resources at their dispersal to improve their children's learning performance (Lareau, 2011;McNeal, 1999). Massey et al. (2011) and his colleagues contend that these two concepts are linked to student integration into the university environment.
Breaking family influences up into structural and process dimensions has important implications. Utilizing data on the characteristics that students carry with them to universities, we might be able to trace those academic, financial, and social challenges that rural students confront back to the source. Therefore, we hypothesize that the equity-enhancing interventions in these challenges put students under intense pressure, forcing them to upset their academic aspirations, thus weakening their commitment to the university institutions they attend. This paper helps to evaluate the equity by introducing a new dimension in horizontal inequality.
Research questions: Hypothesis 1: Rural and urban students in elite universities were advanced individuals in their own group. Hypothesis 2: Rural students were less likely to be academically and socially successful than their urban counterparts.
Hypothesis 3: After controlling for rural-urban status, parental involvement, socioeconomic background, and senior secondary school background had limited capacity in predicting academic and social success.

Methods
This study uses data from the Universities Student Experience Study (USES), a longitudinal mixed-method study (survey with follow-up interviews) that provides a portrait of students at two intervals of their university education. The first wave of data collection was in June 2014, and the second wave was in October 2015. This paper uses quantitative data from the survey. A probability sample, proportional to size (PPS) strategy was used to select the participants for USES. We contacted the four universities for full class lists of the freshmen entering the universities in 2013. Then, we separated the population into three broad strata based on disciplines of faculties and schools: arts and humanities, science, and engineering and medical. For each stratum, we listed classes and their population sizes. Then, using the PPS method, we calculated the cumulative sum of the population sizes and determined the number of clusters (classes) that would be sampled. All the individuals in the selected classes were invited to complete the survey. The different stages were combined to ensure that the sample is representative of the four universities. In total, 1,936 students completed the first wave of the survey, which generated a 96.8% response rate, and 1,661 students responded to the second wave of the survey, for a response rate of 83.1%.
USES has two components. The first includes 47 items which measure students' social origins and retrospectively assess students' family status, past school experience, and neighbourhood environment at three important junctures (primary school, middle school, and high school). The second component contains 19 items and concentrates on assessing students' social and learning experiences, as well as their social and academic successes during their freshman and sophomore years. The survey instrument is an adaptation and extension of a pre-established baseline survey, the Survey of College Life and Experience, used in the U.S. (Bowen and Bok, 2016). We used forward-translation and an expert panel of consultations to produce a conceptually equivalent Chinese version of this instrument. A pilot study (n = 50 students) was conducted in January 2014. The Cronbach's alpha reliability coefficient of most of the constructs in the pilot test was above .8 which suggests an appropriate instrument internal consistency. Figure 1. illustrates the main indices introduced in the study, including students' social origins, demographic background, parental involvement, students' socioeconomic background, senior secondary school background, academic success, and social success. Firstly, social origin, a binary variable (rural/urban), was measured by the administrative area of the permanent home address. Permanent home address such as the municipality, provincial capital (or equivalent administrative area), city, or county was considered as urban. In contrast, a permanent home address in the village (xiang) or small town (zhen) was considered as rural. In the official national definition, those from county townships are considered rural. Based on the rapid urbanization since 2010 with a rising standard of living and average incomes across the nation that the official category rural should exclude those living in increasingly urbanized county townships. In short, ongoing urbanization has produced a quite different country context with a county student population that is at a greater advantage than those in villages and small towns.
Secondly, the demographic background indices include ethnicity (Han or non-Han), gender, single-child family status, and household geographic location. Thirdly, parental involvement was conceptualized to indicate selected aspects of human capital cultivation, social capital cultivation, and cultural capital cultivation. Specifically, human capital cultivation is represented by the parents' decision to hire a tutor to help with their children's academic performance. Social capital cultivation is measured by their communication and interaction with their children's teachers and friends.
Cultural capital cultivation refers to visits to museums, zoos, theatres and sports events, as well as leisure travel experiences. Parental involvement is a continuous measure, and the sum score of the corresponding items is calculated to represent the three sub-concepts.
Fourthly, student socioeconomic background was conceptualized into four sub-concepts, namely, father's occupation, father's education status, mother's education status, and family financial status. Father's occupation was further classified into four categories: manager and civil servant; professional; clerk; and working class/ peasant farmer. The level for parental education is indicated by graduation status from college or university (zhuanke/benke); senior secondary school (gaozhong); junior secondary school (chuzhong); and, some or all primary school (xiaoxue). Additionally, family financial status is measured according to annual income; as well as the size and the value of the house or apartment owned by the family. Senior secondary school background refers to the environment of the schools which the student attended before university, be it a key school, a state school, or some other; the general quality of the school; and the academic peer support. The general quality of the school was measured by a 3-point Likert scale (where 1 indicates poor, and 3 indicates excellent), and the peer support of academic effort was measured by a 5-point Likert scale (where 1 indicates no support at all, and 3 indicates fully support).
Fifthly, academic success is measured by the students' grade point average (GPA) in the university during the first year. Finally, social success was measured by the students' voluntary participation in university organizations. It has been shown that student leadership is an indicator of their ability, popularity, and social engagement (Lin, Sun & Yang, 2015). We compare, in the Chinese context, the semi-official student organizations are more exclusive.
The data were analysed with SPSS (version 24.0). The listwise deletion was used to remove the missing values. Descriptive and t-test analyses are presented to examine whether there is a statistically significant difference between urban and rural students with respect to academic success. Likewise, the authors used a chisquare test to detect statistically significant differences between urban and rural students with respect to social success. Linear regression and logistic regression analyses are adopted to test whether or not parental involvement, socioeconomic background, and senior secondary school background predict academic success, as well as social success. For all tests, the alpha level for statistical significance was set at 0.05.

Descriptive Analyses
The proportion of rural students represented in our sample varies by institution (see in These percentages indicate that rural students are disadvantaged because the official rural population proportion of the country is 50%. The discrepancy in the representation of urban and rural students is even greater if we rank our four sample universities by national prestige, as university [D] is ranked slightly behind the other three universities. Table 2. presents the demographics information separately for urban and rural students.   Table 3. shows the students' socioeconomic background. Overall, our data suggest an underrepresentation of students, whether urban or rural, from lower-SES backgrounds. For example, for urban students, only 3.0% of their fathers were peasant farmers or in the working class. For rural students, less than 20.6% of their fathers were peasant farmers or in the working class. Yet, China's 2010 National Population Census suggests that in the year of 2010, over 70% of people are in these two categories (Lu, 2010). This suggests that children from the middle and upper class in both urban and rural China are favoured in the university admission process. With respect to parents' education status (both father's and mother's), the patterns are similar. Urban students have parents with higher educational attainment. For example, 59.3% of urban students' fathers but only 11.4% of rural students' fathers completed college or university degrees. Likewise, 49.2% of urban students' mothers but only 6.5% of rural students' mothers completed college or university degrees. Yet, urban and rural students' parents have a level of educational attainment that is far above the national average in their own social groups. According to the data of the sixth national census (National Bureau of Statistics of China, 2010), there are around 84% have an education that is at (including and above) junior secondary school (equivalent to middle school degrees) for the urban residents. Yet, in our dataset, about 96.6% of fathers (59.3% college/university, 24.7% senior secondary, and 12.6% junior secondary), and 93.7% of mothers of urban students have an education at junior secondary school. At the national level, around 55% of rural residents have an education that is at junior secondary school. In our data, about 79% of fathers and 62.1% of mothers of rural students have an education from junior secondary school.
Regarding family financial status, both urban and rural students in elite universities are more likely to come from upper-income families in their own regions. The findings show that the average annual family income for the urban students in this study in 2013 is 162, 000 RMB, which is higher (top 20% in urban incomes) than the national average annual family income in the urban area of 83,800 RMB in 2013 (Chen and Ni, 2018). For rural students in this study, the average annual family income in 2013 is 59,100 RMB, which is higher (top 20% in rural incomes) than the national average annual family income in the rural area of 38,200 RMB (Chen & Ni, 2018). This suggests an advantage enjoyed by those middle-and upper-class families in both urban and rural areas.
The ownership of houses/apartments is another indicator of a family's financial status. For urban students in this study, the estimated average value of their families' houses/apartments is 1.1 million (RMB). On average, rural students estimated the value of their families' houses/apartments to be 0.35 million (RMB). Although the self-reported prices were not free of measurement errors, the authors consider that these values were not significantly higher or lower than the average house price on the market in 2015. Note that there was no specific information available in the national report or previous publication regarding the estimated average value of families' houses/apartments for urban and rural residents, and the average price for 37 major cities in China was 8,823 RMB per square meter in 2015 (National Bureau of Statistics of China, 2016). Both 0.35 and 1.1 million RMB across China (including major/non-major cities, and non-cities) suggests that the university students in this study are from more affluent backgrounds within their own social groups (rural or urban). A notable feature of the selective university students documented by researchers was the overrepresentation of graduates from the so-called "No.1 Middle School" or "Key School", the best public key school in each province, city, and county. National Bureau of Statistics of China (2016) indicated that in total there were 2.7% secondary schools nationwide were key schools (at provincial, city, or county level). Wu & Huang (2017) suggested that key schools are more exclusive of students from rural and low socioeconomic backgrounds. The dataset for this study shows that 92.5% of urban and 89.5% of rural students graduated from key schools. Specifically, 55.2% of the urban and 31.3% of the rural students were from a province-level key school (see Table 4.). Nearly 24.5% of both urban and rural students were from a city-level key school. About 12.8% of urban but 33.7% of rural students were from county-level key school. In this study, a small proportion of students (5.9% of urban, and 8.1% of rural students) were from ordinary (non-key) schools. Few (6.6% urban and 5.6% rural) students were from private (minban) schools, which were considered a less advantaged school type in China. Urban and rural students in this study highly rated the quality of their high schools, including the facilities, teaching, as well as school reputation, which suggest a particularly good school climate. They (urban and rural) also reported getting sufficient support from their peers in secondary schools.  Table 5. documents the process dimension of parental involvement. Human capital cultivation, social capital cultivation, and cultural capital cultivation are three indicators of parental involvement. The average human capital cultivation score for urban students is 0.69 which is higher than 0.34 for rural students, where the larger values suggest the parents were more likely to hire a tutor to help with their children's academic performance. The average social capital cultivation score for urban students is 2.69 which is higher than 2.42 for rural students, where the larger values conclude that the parents know and frequently communicate/interact with their children's teachers and friends. The average cultural capital cultivation score for urban students is 2.35 which is higher than 1.54 for rural students, where the larger values convey the idea that the parents were more likely to frequently take their children to visit museums, zoos, and theatres. Generally, urban parents provide more cultivation than those from rural areas.

Group Differences Analyses
For testing hypothesis 1, academic success is measured by students' GPA scores. Our findings show that the average GPA of urban students is 3.57, and the average GPA of rural students is 3.62. The t-test shows no statistically significant difference between rural and urban students in academic success (t = 1.263, df = 1570, p = .207). Social success is measured as a binary categorical variable indicating whether or not a student is a leader of student organizations . There are similar proportions of urban student leaders (53%, n = 592) and rural student leaders (49%, n = 236) participating in activities of student organizations. The chisquare test shows no statistically significant difference between rural and urban students as a leader in student organizations (χ2 = 2.632, df = 1, p = .105).
Considering the different nature of student bodies in terms of importance for social and cultural capital accumulation as well as criteria for admission, we further classified student organizations into two types, namely, official and non-official. The official student organization refers to those connected with the university administration or party apparatus, for example, the student union and Youth League Committee, which have higher criteria for recruitment, are important for profile building, and predict employment upon graduation. In this study, 36% of urban students (n = 396), while 23% of rural students (n = 110) are leaders in official student organizations. The chi-square test shows a statistically significant difference between rural and urban students as being a leader in official student organization (χ2 = 28.55, df = 1, p = <.001). The non-official student organizations include less formal and self-regulated organizations such as football clubs, comic and animation clubs, and so on so forth. The recruitment of non-official student organizations was more based on personal interests and less important for profile building. In our sample, there were 241 urban student leaders (21.9%), while 124 rural student leaders (25.8%) were in non-official student organizations. The chi-square test shows no statistically significant difference between rural and urban students as a leader in non-official student organizations (χ2 = 3.252, df = 1, p = 0.714). It is worth noting that some students were leaders in both official and non-official student organizations. Therefore, the number of leaders in official (i.e., 396 urban students) plus in non-official (i.e., 241 for urban students) student organizations were larger than the number of leaders in any student organizations (i.e., 592 for urban students).

Regression Analyses
This study used multiple linear regression to examine whether parental involvement, socioeconomic background, and senior secondary school background predict academic success or not. In order to mitigate post-treatment bias (Montgomery et al., 2018), several covariates, especially demographic background variables were included in the analysis. The key/non-key school indicator was removed due to the low variance. Geographical regions of family location and mother's education were removed due to the violation of the collinearity assumption of regression. The influence of different variables is examined by adding them to the model one at a time, and the model satisfied all regression assumptions.
In general, 10.3% of the variance of students' GPAs is explained by rural/urban origins, demographic background, parental involvement, SES, and senior secondary school background (see in table 6.). Ethnicity, gender, parental involvement, socioeconomic background, senior secondary school background are statistically significant predictors of the university GPA, after controlling for other variables. Specifically, Han Chinese (ethnic majority group) students have higher GPAs than ethnic minority students (b = .158, p = .027). Likewise, male students have lower GPAs than female students (b = -.344, p < .001). Regarding parental involvement, all three indicators, human capital cultivation (b = -0.54, p=.038), social capital cultivation (b = -.058, p = .010), and cultural capital cultivation (b = .069, p = .018) can predict the academic success. In other words, if the students received more human or social capital cultivation, they had lower GPAs. The results suggest a unique pattern of school involvement by Chinese parents. Parents may choose to intervene in their children's learning process when the children experience learning difficulties.
The result suggests that if the students received more cultural capital cultivation, they had higher GPAs. Regarding SES, father's occupation and father's education predict academic success. For students whose fathers were managers and civil servants, professionals, and clerks, there were no statistically significant differences. For students whose fathers were workers and peasant farmers, their GPAs were relatively lower than (b = -.1804, p = .042) for students whose fathers were managers and civil servants. Students whose fathers graduated from primary school (b = -.196, p = .017) or senior secondary school (b = -.121, p = .013) had lower GPAs than students whose fathers graduated from college. The students whose fathers graduated from junior secondary school had similar GPAs (b = -.068, p = .251) compared with the students whose fathers graduated from college. In terms of the senior secondary school background, for the students who attended province-level key school, city-level key school or country-level key school, their GPAs are very much alike. However, students who attended ordinary schools had lower GPAs, compared to students who attended province-level key schools. The overall quality of the school (b = .020, p = .532) and peer support for academic effort (b = .029, p = .388) were not good predictors of academic success. In this study, we used two logistic regression analyses to examine whether or not parental involvement, socioeconomic background, and senior secondary school background are predictors of social success, while controlling for demographic background. In the first logistic regression model (Model 1 in Table 7.), the outcome variable is whether or not a student is a leader in either an official or non-official student organization (1 = being a leader, and 0 = not being a leader). The results show that all variables (ethnicity, gender, single child or not, rural or urban, parental involvement, socioeconomic background, and senior secondary school background) together explain 3.3% of the variance of being a student leader. In this model, ethnicity and parental involvement are statistically significant predictors, after controlling other variables. Specifically, Han Chinese majority students have a greater chance to be a leader than ethnic minority students (Exp(B) = .532, Wald(1) = 2.418, p = .016) in student organizations. Regarding parental involvement, only social capital cultivation is a predictor of student leadership (Exp(B) = .159, Wald(1) = 2.304, p = .022). In other words, the students who received more social capital cultivation were more likely to receive leadership roles in student organizations.
In the second logistic regression model (Model 2 in Table 7.), the outcome variable is whether a student is a leader in an official student organization (1 = being a leader, and 0 = not being a leader). The results show that all variables together explain 8.5% of the variance of being a leader in an official student organization. After controlling for other variables, ethnicity, gender, parental involvement, and socioeconomic background are predictors of being a student leader in official student organizations. Particularly, Han Chinese majority students have a greater chance of being put in leadership roles than can ethnic minority students (Exp(B) = .773, Wald(1) = 2.386, p = .017) in official student organizations. Male students have a lower chance of becoming a leader than female students (Exp(B) = -.333, Wald(1) = 2.431, p = .015). Regarding parental involvement, only cultural capital cultivation is a predictor of student leadership in official student organizations (Exp(B) = .321, Wald(1) = 3.178, p = .001). In other words, the students who received more culture capital cultivation were more likely to be a leader in official student organizations.
Concerning socioeconomic background, all three indicators, father's occupation, father's education, and family financial status predict being a leader in official student organizations. For students whose fathers were managers, civil servants, or clerks, there is no statistically significant difference in terms of the chance of being a leader in official student organizations (Exp(B) = .053, Wald(1) = 2.789, p = .779). For students whose fathers were professionals (Exp(B) = .464, Wald(1) = 2.320, p = .021) or working class and peasant farmers (Exp(B) = .849, p = .031), the chance of being a leader in official student organizations was lower than students whose fathers were managers and civil servants. The students whose fathers graduated from senior secondary school (Exp(B) = -1.013, Wald(1) = 2.666, p = .008) were less likely to be a leader in official student organizations than students whose fathers graduated from college. Nevertheless, there is no statistically significant difference regarding being a leader in official student organizations among students whose fathers graduated from college, junior secondary school (Exp(B) = -.101, Wald(1) = 0.419, p = .638), or primary school (Exp(B) = -.128, Wald(1) = 0.739, p = .459).

Conclusion
While policies to prepare rural students for the national college entrance examination and other preferential policies for rural and ethnic minority students have been helpful in gaining access to selective universities, the debate continues about the extent of access and equity in China's HPS (Liu, 2012;Mok, 1999). As Chinese society becomes increasingly stratified coupled with a tradition of elitism, there is an intensified competition for social mobility and the admission to top-tier universities. This provides theoretical reasoning for expecting under representations of rural students from more disadvantaged backgrounds gaining admission and succeeding in top-tier universities. It is also reasonable to expect that their integration into an elite university environment may be more problematic.
The USES survey of the freshmen cohort of four elite universities in the autumn of 2013 cast a light on who most benefits from the recent policies aimed at enhancing access and equity in higher education. Our analysis provides an empirical foundation for discussion. The structural dimensions of family influence confirm a disadvantaged position for rural students in access to the most selective universities in China. It also confirms an advantaged position that upper-and middle-class backgrounds students from rural areas have access to selective universities. The data emphasize the role of parenting among the expanding Chinese middle class, a parenting style increasingly characterized by intensive investment in cultural capital. For example, rural students are less likely than their urban counterparts to succeed socially when measured by their memberships of official student bodies, something that is predictable by parental involvement in the creation of cultural capital.
Contrary to studies in Western industrialized countries, our data suggest that rural students perform academically as well as their counterparts from urban areas (Gao, Liu & Fang, 2015;Li, Hou & Wen, 2015). Two reasons explain. First, the national college and university entrance examination (gaokao) is the single, often only, determinant of admission to a top-tier university. To a great extent, it excludes other admission criteria such as high school achievement, other talents and hobbies, family donations, and legacy admission. Second, an increasing number of the more well-off rural parents have the resources and parenting perspectives to position their children into an advantageous position for admission to selective universities. Once on campus, rural students earn grades comparable with their urban counterparts (Author, 2018). This paper highlights the analysis and importance of the process dimension of family influence in access and equity to China's elite universities. Early parental involvement translated to cultural capital for both academic and social success in top-tier universities. This is especially central to understanding the structuring feature of the post-socialist state. As a new structure is arising, the ways those privileged social groups pass on their advantages to their children are becoming more sophisticated.
Upper-and middle-class parents' practices in cultural capital investment are the results of the intensive status competitions among households who are affected by a neo-liberal discourse on individual responsibilities for their own success (Ball, 2003;. Although there is no clear pattern of a dominant culture, many anxious upper-and middle-class urban parents tend to invest in anything that they think might bring privileges to their children. And, their cultural capital strategies are successfully translated into their children's social success in elite universities (Marginson, 2016). Therefore, the concept of cultural capital has the potential to disenchant the privileges in education that the urban and middle-class have. It helps in revealing the traits of these students from upper-and middle-class families in both urban and rural families that are valued by the elite universities.
As Marginson (2016) argues a double competition plays out in the field of higher education: between institutions which compete for status, and between students and their families who compete to obtain credentials to achieve social positions.
Chinese higher education is also becoming increasingly stratified; an examination of the vertical differences in success in access to and within higher education adds a dimension to evaluating policy initiatives to raise the number of rural students from underserved communities in Chinese top-tier universities. To broaden rural student participation also remains an aim of policy. Yet, policies generally fail to consider family influences and the differentiated patterns of success in admission to and success within elite universities by socioeconomic differences. Policies to improve the access and equity in higher education should be recalibrated for those students from underserved rural communities. Further, the stratification of higher education is both supply sideand demand side-driven and it leads to fiercer competition among both universities and families. Students who are better ascribed with cultural, economic and social capital are inevitably in an advantaged position while their counterparts self-exclude or underinvest (Chesters & Watson, 2013;Zha, 2017). More effective reforms should involve the supply side of higher education and reverse the trend towards further polarization of higher education institutions in terms of prestige and resource allocation and geographical location. Moreover, this study applied regression analyses as an explanatory model, and future research can examine how latent variable modes (e.g., structural equation modeling) can be incorporated into representing the concepts such as parental involvement and students' socioeconomic background, and evaluating the relationship between academic and social success.