GENDER INEQUALITY IN EDUCATION IN CHINA: A META...

GENDER INEQUALITY IN EDUCATION IN CHINA: A META-REGRESSIONANALYSIS

JUNXIA ZENG, XIAOPENG PANG, LINXIU ZHANG, ALEXIS MEDINA and SCOTT ROZELLE

Although there is evidence that there was gender inequality in China’s educationsystem in the 1980s, the literature in China has mixed evidence on improvements ingender inequality in educational attainment over the past three decades. Some suggestgender inequality is still severe; others report progress. We seek to understand theprogress China has made (if any) in reducing gender inequality in education sincethe 1980s. To meet this goal, we use a meta-analysis approach which provides anew quantitative review of a relatively large volume of empirical literature on gendereducational differentials. This article analyzes differences across both time and space,and also across different grade levels and ethnicities. Our results indicate that genderinequality in educational attainment still exists, but it has been narrowing over time.Moreover, it varies by area (rural versus urban) and grade level. There is nearlyno significant gender inequality in the case of girls in urban areas or in the case ofthe 9 years of compulsory education (primary school and junior high school). Girls,however, still face inequality in rural areas (although inequality is falling over time)and when they reach high school or beyond. (JEL I24)

I. INTRODUCTION

The 2012 World Development Report (WorldBank 2012) focuses on gender and develop-ment and states explicitly that gender equal-ity is a core development objective in its ownright. According to the Report, gender equal-ity enhances the productivity of the currentgeneration and improves development outcomesfor the next. One of the main mechanisms of

Zeng: School of Agricultural Economics and Rural Devel-opment, Renmin University of China, No. 59, Zhong-guancun Boulevard, Haidian, Beijing 100872, China.Phone 86-13811473256, Fax 86-10-62511064, [email protected]

Pang: School of Agricultural Economics and Rural Devel-opment, Renmin University of China, No. 59, Zhong-guancun Boulevard, Haidian, Beijing 100872, China.Phone 86-10-13371790098, Fax 86-10-62511064,E-mail [email protected]

Zhang: Center for Chinese Agricultural Policy, Institute forGeographical Sciences and Natural Resource Research,Chinese Academy of Sciences, Beijing 100101, China.Phone 86-10-64889440, Fax 86-10-64889440, [email protected]

Medina: Freeman Spogli Institute, Stanford University,Stanford, CA 94305. Phone +1 (650) 724-6402, Fax+1(650) 723-6530, E-mail [email protected]

Rozelle: Freeman Spogli Institute, Stanford University,Stanford, CA 94305. Phone +1 (650) 724-6402, Fax+1(650) 723-6530, E-mail [email protected]; Schoolof Agricultural Economics and Rural Development,Renmin University of China, No. 59, ZhongguancunBoulevard, Haidian, Beijing 100872, China.

development that arises from gender equalityis the improvement in education that moveshand in hand with gender equality. Girls whoreceive more education have more opportunitiesto improve their own vocational opportunities,living conditions, and social status; they alsocontribute more to economic growth (Glewweand Kremer 2006). There are also particularexternalities from female education in terms ofreduced population growth, better child health,and household investments in children, moregenerally (Song and Appleton 2006).

Despite the adverse role that gender inequal-ity can have on development, many develop-ing countries exhibit gender inequality in manydimensions, including education. Almost half ofthe world’s elementary school-aged girls, whoare not in school, live in Sub-Saharan Africa;around a quarter live in South Asia (World Bank2011). In India, the second most populous coun-try in the world, of all the elementary school-aged children who should be in school but arenot, the majority are girls (56%—UNESCO2005). The elementary school drop-out rate of

ABBREVIATIONS

GER: Gross Enrollment RateMDG: Millennium Development Goal

474Contemporary Economic Policy (ISSN 1465-7287)Vol. 32, No. 2, April 2014, 474–491Online Early publication February 25, 2013

doi:10.1111/coep.12006© 2013 Western Economic Association International

ZENG ET AL.: GENDER INEQUALITY IN EDUCATION IN CHINA 475

girls is twice as high as that of boys in Equato-rial Guinea and Grenada. The secondary schooldrop-out rate of girls in many developing coun-tries is also high (UNESCO 2011).

One of the Millennium Development Goals(MDGs) related to education is the eliminationof gender disparity at the primary and secondaryschool levels before 2015. However, despitethe often unsubstantiated reports of progressover the past years (at least unsubstantiated byindependent sources), many countries are stillfar from reaching this goal. Scores of countriesreport that they will not make the goal offull enrollment of girls into school (UNESCO2008). In 2005, only 59 of 181 countries (aboutone-third) with data available had achievedgender parity (i.e., GPIs ranging from 0.97 to1.03) in their gross enrollment rates (GER) forboth primary and secondary education. Mostwere developed countries and most had alreadyachieved parity by 1999. The pace of reducinggender disparity has been much slower both at aglobal level and in those regions with the widestdisparities in 1991 (the Arab States, East Asiaand the Pacific, South and West Asia, and Sub-Saharan Africa—UNESCO 2008).

Despite the considerable body of evidence,the finding of research on gender inequality ineducation in the case of China is still mixedin 2000s. Some researchers say that there arestill significant disparities in access to educa-tion between males and females (Cao and Lei2008; Davis et al. 2007; Hannum, Wang, andAdams 2008; Hong 2010). For example, using a0.95 per thousand micro sample from the 2000China population census, Hannum, Wang, andAdams found that fewer girls were enrolled incompulsory education in 2000. In contrast, otherstudies find that gender inequality in educationhas improved (Liu 2004; Wang 2010; Wu andZhang 2010). For example, using China 1990population census data, Liu found that there wasno significant gender gap in the transition to col-lege in urban areas in 1990. In short, there stillremains great variability among the estimatesof gender inequality in education in China and,according to some, it is still severe.

The goal of this article is to understand theprogress China has made (if any) in reduc-ing gender inequality in education since the1980s. To meet this goal, we have three spe-cific objectives. First, we set out to collect allthe empirical papers that have examined genderinequality in education since the 1980s, focusingspecifically on gender inequality in schooling

attainment. The findings of this literature willbe systematically categorized and turned intoa database which will form the basis of ourstudy and will help us identify the channelsthrough which the gender inequality is occur-ring. In this way, our article can be considered asa meta-analysis. Second, we tabulate the resultsof the studies and document the nature of genderinequality in educational attainment since thereform era (around 1980 to the present). Third,we decompose the overall findings and trackhow gender inequality in education changes overtime, across regions, by grade level and betweenHan and minority children.

Why is China an interesting case study?While the gender inequality in education prob-lem today is almost surely less severe—especially in parts of the country—than else-where in the world, we believe that a study ofChina across the past three decades is certainlyof interest. Most poignantly, China’s economyhas grown extremely fast, rising by more than10 times between 1980 and 2010 in terms ofGDP per capita (NBSC 2011). However, thegrowth has been unequal between rural andurban (NBSC 2011). There also have beenchanges on the supply side of education and therate of change has changed at different paces indifferent levels of schooling (i.e., college, highschool, etc.). Because of this, China more thanany place in the world over the past 30 yearsmight be considered as a laboratory to followthe changes such rapid but differential rates ofchange may have had on gender inequality.

While the goal of our research is ambitious,there is a limitation. We tried to identify allpapers with a set of key words (see SectionIII). We believe that we have done a thoroughand convincing job using this set of papers todemonstrate how gender inequality (a) changesover time; (b) differs between rural and urbanareas; (c) differs by the level of education; and(d) varies between Han and minority popula-tions. However, there is still no guarantee thatour findings and conclusions would not changeif the search criteria were changed.

The remainder of the article is organized asfollows. In Section II, we introduce several ofthe salient features of the education system inChina which may affect the nature of genderinequality (e.g., female schooling in urban andrural areas might be expected to differ). If thereis reason to believe that gender inequality woulddiffer in different dimensions of the educationalsystem (e.g., over time; between urban and rural;

476 CONTEMPORARY ECONOMIC POLICY

by level of schooling; between Han and minorityareas), it means that our empirical analysisshould seek to understand how gender inequalitydiffers by these different aspects. In Section III,we discuss our approach to the meta-analysisand describe the meta-data that are employed inthis article. In Sections IV and V, we presentthe results of the meta-analysis and conclude.

II. THE NATURE OF EDUCATION IN CHINA ANDEXPECTED SHIFTS IN GENDER INEQUALITIES

The rapid expansion of the economy overthe last three decades will likely have affectedgender inequality since economic changes affectboth the education system (the supply of school-ing) as well as the returns to education andthe ability of parents to send their children—including their daughters—to school (the de-mand for schooling). There are major shiftsin many dimensions. China’s GDP per capitahas grown by nearly 10% per year sincethe late 1970s (NBSC 2010). During thistime, fiscal revenues have also risen con-sistently—especially since the early 1990s.According to Wong and Bird (2008) and NBSC(2011), after rising gradually in absolute realterms between 1980 and the mid-1990s, totalgovernment revenues as a share of GDP morethan doubled between 1995 and 2010. Atthe same time, the economy has been trans-formed from one that was based on the plan-ning (in the 1970s) to one that is now mostlymarket-oriented (Brandt and Rawski 2008).Most employment decisions are now made byindividuals (firm owners and managers) that arehiring and firing in order to make their busi-ness units more profitable and efficient. Indeed,compared to the 1970s, today China is muchwealthier, better able to invest in social services,and its economy is (and firms in the economyare) driven in no small part by market signals.

All these shifts have affected the ability ofthe state to invest in education and increasethe supply of schooling. The capacity to sup-ply educational services at all levels has risen(Hannum et al. 2008). For example, the numberof elementary schools increased monotonicallybetween 1980 and 2000. It was not until after2000 that the number of schools fell, but thefall was mainly due to a large school mergerprogram that built large centralized schoolswith better facilities and more qualified teach-ers (including boarding facilities for studentsthat lived too far from school—Liu et al. 2010).

During this same period, the number of availableslots in secondary school rose—gradually dur-ing the 1980s and accelerating in the 2000s asthe state began to open more vocational edu-cation and training schools (NBSC 2010). Thegreatest expansion—but, the latest in terms oftiming—came in tertiary schooling. Between1998 and 2009, the number of students in col-lege increased by more than six times (NBSC2010). The opportunity to go to school hasclearly increased between the 1980s and 2000s.

Many of the same factors—economic growthand the rise of a market economy (that demandedworkers with greater levels of human capi-tal) as well as the rise in the opportunitiesto go to school (i.e., more slots in schools atall levels)—also has been changing the cal-culus for parents and students. Studies of thereturns to education have shown that betweenthe 1980s and 2000s the returns to rural educa-tion have risen substantially (e.g., deBrauw andRozelle 2008; Zhang, Huang, and Rozelle 2002).Returns to education nearly tripled between1988 and 2003 in urban areas (Zhang andZhao 2007). Other studies have shown signif-icant returns to climbing high within the educa-tion system and attending college. For example,Fleisher et al. (2004) find the return to collegeeducation, in terms of the percentage return peryear of college, increased sharply from 11.85%in 1995 to 23.2% in 2002. The rise in thesupply of schooling opportunities and increaseddemand by families almost certainly increasedthe demand—among other things—for educa-tion for girls both by the parents of girls andby the girls themselves. Hannum et al. (2008)suggest that their research supports the conclu-sion that in order to support the sustainableincreasing demands of skilled labor for eco-nomic growth, educational systems were restruc-tured, and education was expanded by China’sgovernment, a move that provided more accessto educational opportunities for all, includinggirls. From these analyses, for those that arelooking to understand gender inequality in edu-cation, it is important to look at the changes ingender inequality over time.

Other institutional peculiarities, beyondgrowth over time, may also be important inexplaining changes in gender inequality ineducation. One of the most notable struc-tural divides in the population of China isthe urban–rural divide (Naughton 1994). Thehukou system, initiated in the 1950s and 1960s,assigned individuals into one of two groups—


urban or rural. Over time, the system created onewith sharp differences in almost every aspectof life—health, housing, employment, socialsecurity, and many other elements. Because ofthis, in part, there are also sharp differences inincome (NBSC 2010). Between the 1980s and2010, the urban to rural income ratio has fluc-tuated from somewhat greater than 2 to some-what greater than 3. There also have been sharpdifferences between urban and rural areas inthe implementation of China’s family planningpolicies (Yang 2007). Since the late 1970s andearly 1980s urban city officials have almost uni-versally implemented a one-child policy. Ruralfamilies have had more leeway and typicallyare allowed more than one child (except in thecoastal areas).

Over and above the natural biases that giverural households lower demand for education,the systematic differences between urban andrural economies stemming from the hukou sys-tem should almost certainly be expected tochange the supply and demand for schooling.In this way, the rural–urban divide most likelyis an important factor to consider in understand-ing China’s gender inequality in education. Inurban areas, higher incomes and better welfareservices—and perhaps higher returns to educa-tion—have been thought to reduce the genderbias against girls. Rural educational attainmentis positively related to income per capita, andgirls are more likely to drop out due to financialproblems of the family in rural areas (Knightand Li 1996; Brown and Park 2002). More-over, just under 50% of families in urban areashave no choice but to embrace the educationof their daughter, since she is their only child.Higher levels of wealth and greater access totax revenues in China’s industry-biased tax sys-tem (Wong 1991) also mean that there are sharpdifferences in educational opportunities. Urbanspending on elementary and secondary educa-tion was 1.4 times to 2.6 times relatively greateron a per capita basis compared to rural spendingon education in 2000 (NBSC 2001). All thesefactors make it almost certain that there are dif-ferences in gender inequality between rural andurban populations and, as such, urban and ruraldifferences need to be considered in any decom-position analysis.

However, the weakening of the hukou systemsince the late 1980s (Cai, Park, and Zhao 2008),the rise of off-farm employment (Zhang, Huang,and Rozelle 2002), and falling fertility in ruralareas (as well as a strengthening of family

planning and successful implementation of theone-child policy in many rural areas—Yang2007) mean that the differences between ruraland urban gender inequality in education maybe changing.

Other factors may also create structuralsources of gender inequality in education. Forexample, there are striking differences in therules covering school attendance and the costof attending school across different levels ofschooling. Since the early 1980s, grades 1–6have been compulsory; since the mid-1990sgrades 7–9 have been so (NBSC 1997). Sincethe early 2000s tuition for grades 1–9 wereeliminated, and there are now subsidies forthe poor to attend school (Hannum, Wang, andAdams 2008). In contrast, attendance in uppersecondary schools (grades 10–12) is not manda-tory and the tuition for rural public high schoolsin China is higher than that of almost all otherdeveloping countries in the world (Liu et al.2009).1 College enrollment slots, despite recentexpansion, are still restricted (Li 2010). Tuitionfees and other costs of sending a child to col-lege can be 20 or more times higher than percapita income of a family in poverty (Liu et al.2011). However, the emergence of scholarships(and educational loans) for the poor may haveoffset the rise of tuition and lowered cost as abarrier to going to college (Wang et al. 2011).While over time, the relative costs, the avail-ability of slots in schools, and the nature of therules governing compulsory education have dif-fered, the differences among the different levelsof schooling almost certainly mean that an anal-ysis seeking to decompose the gender inequalityof education must consider this as an importantfactor.

Finally, ethnicity may also play a role in thedetermination of the gender inequality of edu-cation. Minority groups accounted for 8.5% ofChina’s population in 2011. As most of themreside in sparsely populated, relatively poor, andrural interior regions, minorities almost certainlymake up more than their population share ofthose that are under achieving in China’s educa-tion system. There are differences in the socioe-conomic status and employment structure ofHan and non-Han populations. For example, the

1. Although the cost of going to high school in rural andurban areas is nearly the same, given differences in incomesbetween rural and urban residents, the cost in terms share ofper capita income (i.e., what percent of per capita incomeis needed to pay for high school) for going to high schoolvaries greatly between rural and urban.


average household income of the rural non-Hanpopulation was only 64.1% of that of the Hanpopulation in 1995; only 14.75% non-Han pop-ulation had non-agricultural jobs compared with24.97% Han population (Gustafsson and Li2003). In addition, there are certainly also well-known cultural biases (that may be reinforcedby the same factors that determine the incomeand social gaps between Han and non-Han).The wide use of the Han language in the cur-riculum and the unified examination based onthe Han language (which the minority childrenare relatively less familiar with) are examplesof such cultural biases (Hong 2010). Many ofthese factors may make it so there are differ-ences in the education of Han and non-Hansocieties—especially in the case of girls.

III. METHODOLOGY AND DATA

Meta-analysts employ statistics to describeand explain previously reported statistical anal-yses that examine the same phenomenon. There-fore, in simplest terms, a meta-analysis is astatistical analysis of the survey findings of alarge number of empirical studies. In meta-analysis, papers investigating one particulartopic are collected and each reported empiricalstudy becomes one or more observations. Meta-analyses allow the evaluation of the effect ofdifferent data characteristics and methodologieson the results reported (Stanley 2001).

While a meta-analysis has the same over-all goal as a detailed literature review, thereare inherent differences. When conducting a tra-ditional narrative literature review, it is diffi-cult to provide a full quantitative assessmentof the literature. The author has full controlover his/her essay and interpretation. In most lit-erature review-based reviews many papers arediscarded or not addressed due to an under-standable need to distill ideas and focus effort.Meta-analyses, on the other hand, are supposedto eliminate the discretion of narrative reports.Authors collect papers in a standardized way,a step that is purposively done to eliminatethe author’s personal bias. As a result, meta-analyses are considered by some as a morerigorous alternative to narrative discussions ofresearch or literature reviews (Phillips 1994).

Because of the almost natural inclination byauthors to use their discretion, it is perhaps notsurprising that the current literature in China thatseeks to summarize and draw conclusions aboutgender inequality in education is conflicted.

Some authors have come to the conclusionthat there is considerable gender inequality ineducation in China (Cao and Lei 2008; Daviset al. 2007; Hannum, Wang, and Adams 2008;Hong 2010). Others state strongly that there islittle gender inequality in education in China(Liu 2004; Wang 2010; Wu and Zhang 2010).For this reason, it seems that a careful, objectivemeta-analysis on the topic may be a welcomecontribution to the literature.

However, a meta-analysis would be an inef-ficient way to study this question if reliable,comprehensive, disaggregated nationwide dataexisted. In other words, if national statisticaldatabases had data available by period, region(e.g., urban and rural), grade level (e.g., ele-mentary school to college), and ethnicity forboth males and females, there would be no needfor a meta-analysis. Unfortunately, China doesnot have and/or does not make such data avail-able to the research community. While nationallevel educational statistics are released everyseveral years, these statistics are highly aggre-gated. Moreover, there is reason to be suspiciousof the quality of data collected by the nationalstatistical system.2

For many of the same reasons, it is eitherimpossible and/or perhaps undesirable to rely onthe national education database. To our knowl-edge, there is no systemic database on atten-dance or on schooling attainment that exists forall levels of schooling, by gender and by ruraland urban. National assessment and standardizeddata in China are almost never made available toindependent research teams. The statistics thatare published are often incomplete and do notallow for systematic decomposition and analy-sis. In addition, there are hints in the literaturethat national statistics on education, includingthose that are used to report certain educationalattainment figures, may be subject to qualityproblems. For example, using micro-data thatthey believe to be high quality, Mo et al. (2011)and Zhao and Glewwe (2010) find the dropoutrates of junior high school students are nearlythree times higher than the officially reportedrates.

2. For example, Rawski (2001), for one, speculates thatthere are even problems with the data that are used bythe China National Bureau of Statistics to calculate GDP.Such discrepancy can lead to arguments about the nature ofconclusions made on the basis of such data. Ma, Huang,and Rozelle (2004) and Crook (1993) show that similarproblems occur in the case of many of the most fundamentalagricultural statistical series.


In order to add a new dimension to the empir-ical gender inequality in education literature,we therefore utilize a meta-analysis approach inthe rest of this article. For all of its strengths,it should be remembered that the relationshipexpressed by a meta-data analysis is an asso-ciation across studies. The relationship betweenthe independent and outcome variable may notreflect a causal relationship. The associationfound in a study between the variables couldbe due to an omitted variable bias. There alsomay be heterogeneity across the studies, whichmeans we are only reporting the average results.The reader needs to keep these considerations inmind when interpreting the results.

A. Meta-Data

In the execution of our meta-analysis, weused an easily accessible but universal researchdatabase as a way to make the construction ofthe database as transparent as possible. Specifi-cally, in September 2011, we searched the Webof Science for any new listing with the phrase(education or enrollment or academic achieve-ment or academic attainment) and (China), withor without the additional search term (gender ordifference or inequality or girl ). Relevant arti-cles from their reference lists were also reviewed(Table A2).

Each study included in this meta-analysis metthe following criteria:

1. The study must have presented an empiri-cal estimate of the gender difference or sufficientinformation to calculate it; that is, a study shouldcontain enough statistical information so thattest statistics, such as those resulting from a ttest, ANOVA, and so on, were either providedin the study or could be determined from themeans and measures of variance listed in thestudy.

2. The study must have been concernedabout the educational attainment (or achieve-ment in the case of the analysis in the sectionbelow) of any level of schooling from grade1 through college (or other tertiary educationalinstitution). We did not include pre-school orkindergarten.

3. The study needed to be set in China andcould have been a published or unpublishedstudy.

Our Web of Science search led to 813 ref-erences. Papers that were both concerned withgender and contained empirical analyses were

examined for whether they used any regressionanalysis or had enough statistical informationto calculate gender differences. A total of 55articles met these requirements.3 This does notmean, however, that we have only 55 observa-tions. Some authors studied gender inequalityover different time periods, areas, or grade levelsin a single paper. In our meta-analysis, we treatthese estimates as independent estimates.4 Eachindependent observation is coded separately. Weidentified a total of 167 different study or sub-study observations on gender inequality in edu-cation.

According to the results of our search oneducational attainment (and gender inequality),we counted studies that focused on a numberof different elements of educational attainment.For example, we included studies that exam-ined enrollment rates, drop-out rates, graduationrates, transition rates, and years of schooling.Although these concepts are all somewhat dif-ferent, all of them can be converted into anexpression that measures educational attainment.

Following the discussion in the previoussection, we were able to code our data in a waythat will allow us to decompose (through regres-sion analysis—see next subsection for a com-plete discussion) into several key dimensions.In each study, we know the time period beinganalyzed: the 1980s, the 1990s, or the 2000s.Therefore, in the case of each observation, wecreate a matrix of dummy variables called Time.We also know if the data were from a rural pop-ulation (in which case a variable Rural is codedas 1) or urban (Urban) population or one thatincluded both rural and urban (henceforth, callednationwide). The dummy variables for rural,urban, and nationwide are collectively calledArea. We also code the data by the grade levelthat was being analyzed (elementary school ;lower secondary ; upper secondary ; and tertiary,which as a group forms a matrix that is calledGrade level ). Finally, studies could be catego-rized by the ethnicity of the cohort being studied,either Han or minority (non-Han). These finaltwo variables are formed into a matrix calledEthnicity.

Table 1 summarizes the data employed inthis meta-analysis over Time, Area, Grade

3. Nine percent of studies are in Chinese.4. Because we might worry that a single study could

have undue influence, in the results section below we seekto control for this and eliminate any bias that would resultfrom our decision to use multiple findings from a singlestudy.


TABLE 1Descriptive Statistics about the Data for the

Meta-Analysis Study of Gender Inequality inEducational Attainment

Numberof Studies Percentage

Time 1980s 42 251990s 73 442000s 52 31

Area Rural 53 32Urban 33 20Nationwide 81 48

Grade level Elementary 33 20Lower secondary 61 36Upper secondary 45 27Tertiary 28 17

Ethnicity Han 151 90Minority 16 10Total 167 100

level, and Ethnicity. Within the Time matrix,approximately 25% of studies use data from the1980s, 44% use data from the 1990s, and 31%present data from the 2000s. Within the Areamatrix, more studies use data from rural areas(32%) than from urban areas rural (20% urban);most studies (48%) use data for the whole coun-try. Within the Grade level matrix, just over halfof the studies (around 56%) focus on the years ofcompulsory education (20% on primary school;36% on middle school). Twenty seven percentfocus on high school. Only 17% focus on tertiaryeducation. Within the Ethnicity matrix, about90% of the studies use data on Han students.In contrast, only 10% focus on minorities.

B. Meta-Regression Approach

For the purposes of the current study, thedependent variable of interest, y, is a dummyvariable that refers to whether the study foundgender inequality against girls in terms of enroll-ment or educational attainment. If y = 1, thestudy found that girls suffered a statisticallysignificant disadvantage in terms of educationalattainment. A statistical cutoff at the 10% levelis employed here. If girls were not statisticallydisadvantaged relative to boys, the variable wasequal to 0. Three percent of the studies foundthat boys have significantly lower educationalattainment than girls. In those cases, we stillcoded the dependent variable as 0.

In order to estimate the trend and patternof gender inequality in educational attainment,we also want to control for the variables that

might influence the estimated inequality againstgirls when we run the meta-regression model.Following the discussion above, we include foursets of independent variables, Time, Area, Gradelevel, and Ethnicity.

Given these definitions, the following simpli-fied “marginal model” is specified:

y = a0 + a1 × Time + a2 × Area + a3(1)

× Grade level + a4 × Ethnicity + e

where y is a dummy variable equal to 1,if there is gender inequality against girls. InEquation (1), Time is a matrix that includesthree dummy variables (1980s, 1990s, and2000s) and is included to examine how gen-der inequality in educational attainment changesover time. The matrix Area includes three vari-ables (rural, urban, and nationwide) and isincluded to examine whether there is a differ-ence in gender inequality between rural andurban areas. Grade level is a matrix that includesfour variables (elementary schools, lower sec-ondary schools, upper secondary schools, andtertiary schools). Ethnicity is a dummy variableequal to 1 when the study population is non-Han. It is included to examine whether thereis a difference in gender inequality betweenHan and Minority groups. The other terms inEquation (1) are defined as: e is an error termand a0, a1, a2, a3 and a4 are parameters to beestimated.

Estimation Approach. Because of the condi-tional nature of the dependent variable, we esti-mate Equation (1) using a simplified “marginalProbit” estimation. In our estimation, we reportmarginal coefficients of our independent vari-able. Because of this, the coefficients can beinterpreted as the probability that the genderinequality of educational attainment increased ordecreased.

IV. RESULTS

As shown in Table 2, without regard to dis-aggregation, the percentage of papers findingsignificant gender inequality against girls from1980 to the 2000s and across all schools andgrade levels is 66%. This means that in 66%of the studies, girls were found to be at a dis-advantage in educational attainment comparedwith boys. Three percent of the studies foundthat boys have significantly lower educationalattainment than girls. If one only considers thismost aggregated of statistics, gender inequality


TABLE 2Gender Inequality in Educational Attainment

(in the Aggregate) in China, 1980s, 1990s, and2000s

Numberof Studies

Percentage (Column1, Rows 1 or 2

Divided by Row 3).

Girls do not sufferfrom genderinequalitya

56 34

Girls suffer fromgender inequality

111 66

Total 167 100

aIn category we combine the counts of studies that findno gender differences and gender inequality against boys.It should be noted that only 3% studies found evidence ingender inequality against boys.

in educational attainment in China in the ReformEra (around 1980 to the present) appears toremain an issue.

A more nuanced analysis, however, is pre-sented in Table 3, decomposing the results byTime, Space, Grade level, and Ethnicity. Thelow p value on the time indicator (.02) sug-gests that the level of gender inequality thata study finds against girls differs significantlyacross time. Specifically, the instances of findinggender inequality against girls in the literaturereduces dramatically from 81% in the 1980s to67% in the 1990s and finally to only 54% in the2000s.

Findings of gender inequality against girlsseem to differ significantly across space aswell. This can be seen most clearly in the databy examining the low p value (.00) for thearea indicator. There are nearly twice as manypapers finding gender inequality in rural areascompared to those finding it in urban areas,suggesting a wide urban–rural gap in genderinequality. In 7 out of 10 studies conducted inrural areas, the analysis shows that girls haveinferior access to education relative to their malecounterparts; in the urban areas, this is only truefor 36% of the studies.

Findings of gender inequality against girlsacross different grade levels seem to follow adifferent trend from that of time and space.The relatively high p value (.82) indicates nosignificant difference across grade levels (whentaken as a group). The consistent findings ofgender inequality against girls—between 62%and 71% across all levels of schooling—suggestthat girls suffer high levels of gender inequalityin educational attainment throughout the entireeducational system.

Finally, looking at the case of gender inequal-ity among minority students, we find that ahigher percentage of studies find gender inequal-ity among minority children (75%) than amongHan children (66%); however, the differencebetween these numbers is not statistically signif-icant. Because the percentages are fairly closeand because the number of studies lookingspecifically at gender inequality in educational

TABLE 3Gender Inequality (against Girls) in Educational Attainment by Time, Area, Grade Level, and

Ethnicity in China in the 1980s, 1990s, and 2000s

Percentage ofStudies Finding

Gender InequalityStandardDeviation

Numberof Studies p Valuea

Time 1980s 81 0.40 42 .021990s 67 0.47 732000s 54 0.50 52

Area Rural 68 0.47 53 .00Urban 36 0.49 33

Grade level Elementary 67 0.48 33 .82Lower secondary 62 0.49 61Upper secondary 71 0.46 45Tertiary 68 0.48 28

Ethnicity Han 66 0.48 151 .45Minority 75 0.45 16Total 66 0.47 167

aThe p values in this column can be used to test for the differences among the subcategories in each group (Time, Area,Grade level, and Ethnicity).


attainment among minority populations is small,it is unclear from the descriptive statistics ifthere is any real difference in the frequencyof reported gender inequality between Han andminorities. In other words, because the samplesize limited the power of our analysis, it is notpossible to be confident about the results fornon-Han girls.

A. Econometric Considerations

As mentioned, multiple studies are availablefor each paper. In order to eliminate the potentialbias from our use of multiple findings withina single paper, we employ one way to controlfor the undue influence a single paper mighthave. Table 4, column 1 presents the results ofthe simplified marginal probit regression withoutcontrolling for multiple use of a single paper. Incolumns 2–5, we adopt a weighting scheme: allstudies from a single paper are weighted withthe inverse of the number of studies containedin that paper. When doing so, the coefficient ofEthnicity becomes significant in the weightedmarginal probit regression shown in column2, indicating that the probability that a paperfinds a statistically significant gender disparityin educational attainment for minority femalestudents is higher than the probability of findingit for Han female students when we control forstudies coming from a single paper.

B. Gender Inequality across Time

Simultaneously looking at gender inequal-ity and our other variables using Equation (1),we find the results are mostly consistent withthe descriptive findings. The coefficients on thedifferent time periods are all negative and sig-nificantly distinguishable from zero (Table 4,column 1, rows 1 and 2), indicating that theprobability that a study finds statistically sig-nificant gender inequality in educational attain-ment against girls is declining significantly overtime. Compared with the 1980s, the probabilitythat a study finds gender inequality against girlsdecreased significantly by 27% in the 1990s.What is more, the coefficient on the 2000s is40.47, larger than that on the 1990s (26.72),indicating that the probability that a paperfinds statistically significant gender inequality islower in the 2000s relative to the 1990s.5 It is

5. In another robustness check, we also added a squaredyear term in addition to a linear year term (results not shownfor sake of brevity). If significant, it would suggest that the

important to remember, however, that althoughfindings of gender inequality have been trend-ing downwards over time, overall, a majorityof studies continue to find significant genderinequality in educational attainment in the 21stcentury (Table 3).6

C. Gender Inequality across Regions and Timeand Regions

Statistically significant gender inequality hasbeen found in both urban and rural areas. How-ever, urban and rural China are so differentthat each requires its own careful analysis. Thesignificant negative coefficient on urban areas(Table 4, column 1) shows that the probabilitythat a study finds a statistically significant gen-der disparity in educational attainment is 41%lower in urban areas than that in rural areas.This is consistent with our descriptive analysis(Table 3).

Column 4 in Table 4 shows estimates forinteractions between time and region. The sig-nificant positive coefficients on the interactionvariables suggest that the probability that a studyfinds gender inequality against girls in ruralareas decreased more over time than did theprobability of finding gender inequality againstgirls in urban areas. Nevertheless, despite theprogress that has been made since the economicreform of the late 1970s, the probability ofa study finding gender inequality against ruralgirls in the 1990s is still significantly higher thanthe probability of a study finding gender inequal-ity against urban girls (p value is .00). By the2000s, however, the difference in these proba-bilities has shrunk even further (p value is .10),suggesting that the difference in gender inequal-ity between rural and urban areas is narrowingover time.

D. Gender Inequality across Grade Levels andEthnicity

The results of the multivariate analysis ofdifferences in gender inequality across grade

rate of change of gender inequality in education was notlinear. We tried this and found that in all specifications thecoefficients on the squared year variable were insignificant.This implies one of two things: (a) the rate of changewas more or less linear or (b) we do not have enoughobservations to pick up differential rates of change.

6. In short, the meta-regression analysis demonstrates aclear declining trend of gender inequality against girls ineducation. And we do not know exactly why (is it due torising demand for the education of girls or rising supply ofschools).


TABLE 4Marginal Probit Regression Analysis of Gender Inequality in Educational Attainment

(Using Decades as Variables)

(1) (2) (3) (4) (5)

Dependent Variable: Gender Inequality against Girls (1 = Yes, 0 = Neutral or against Boys)(1) Time (ref. = 1980s)

1990s −26.72∗∗∗(3.17)

−10.99(0.81)

−171.32∗∗∗(8.54)

−13.31(0.93)

−37.22∗∗(2.24)

2000s −40.47∗∗∗(4.42)

−32.13∗∗∗(2.56)

−183.71∗∗∗(11.36)

−32.81∗∗(2.50)

−210.79∗∗∗(10.86)

(2) Area (ref. = rural areas)Urban −41.09∗∗∗

(4.53)−35.13∗∗∗

(2.77)−192.61∗∗∗

(11.33)−36.93∗∗∗

(3.02)−30.07∗∗

(2.58)Nationwide 7.04

(0.96)2.8

(0.28)−149.23∗∗∗

(6.46)5.1

(0.49)6.55

(0.66)(3) Grade level (ref. = elementary)

Lower secondary 12.1(1.32)

8.18(0.68)

11.35(0.88)

8.53(0.69)

−23.88(1.15)

Upper secondary 27.68∗∗∗(2.70)

24.47∗(1.87)

30.49∗∗(2.27)

24.40∗(1.83)

−11.79(0.52)

Tertiary 32.42∗∗∗(2.91)

42.60∗∗∗(3.34)

46.58∗∗∗(3.68)

45.04∗∗∗(3.50)

157.87∗∗∗(9.12)

(4) Ethnicity (ref. = Han) 6.68(0.56)

27.29∗∗(2.19)

23.62∗(1.67)

143.96∗∗∗(7.39)

23.69∗(1.67)

(5) Interaction Time × Area1990s × urban 158.66∗∗∗

(5.53)1990s × nationwide 165.47∗∗∗

(5.92)2000s × urban 165.81∗∗∗

(6.44)2000s × nationwide 147.69∗∗∗

(5.68)(6) Interaction Time × Ethnicity

1990s × non-Han −124.44∗∗∗(4.34)

2000s × non-Han −122.88∗∗∗(4.54)

(7) Interaction Time × Grade1990s × lower secondary 42.92∗

(1.68)1990s × upper secondary 35.32

(1.29)1990s × tertiary −119.63∗∗∗

(4.92)2000s × lower secondary 191.43∗∗∗

(6.42)2000s × upper secondary 215.35∗∗∗

(7.03)2000s × Tertiary 38.58∗

(1.74)Observations 167 167 167 167 167

Notes: Robust t-statistics are in parentheses. In column 2, we use the inverse of number of the studies from a singlepaper as the weight. In columns 3–5, we do the interaction for the variables of Time and Area, Ethnicity, and Grade usingthe weighted marginal probit model. ref., reference.

∗Significant at 10%; ∗∗significant at 5%; ∗∗∗significant at 1%.

levels differ from the results of the descrip-tive analysis. Once Time, Area, and Ethnicityare taken into account, Grade level is shown to

be significantly correlated with findings of gen-der inequality against girls (Table 4, column 1,rows 5–7). More specifically, while there is no


significant difference between studies of ele-mentary school and studies of lower secondaryschool in the probability of finding genderinequality (Table 4, column 1, row 5), however,there is a significantly higher probability thatstudies of high school find gender inequality.China appears to have made noticeable progressachieving gender equality in elementary andlower secondary education, particularly in termsof enlarging and equalizing access (Wang 2010).The higher enrollment rate and fewer find-ings of gender inequality in elementary andlower secondary education may reflect the lowor non-existent fees for compulsory school-ing and the lower opportunity cost of keepingyoung children out of the workforce, as farminghas become less important over the course ofChina’s development during the 1980s, 1990s,and 2000s (Song and Appleton 2006).

Once students ascend beyond compulsoryeducation, however, the probability that a studyfinds gender inequality against girls is positiveand statistically significant (Table 4, columns 1,rows 6–7). Papers looking at gender inequal-ity at the high school level found that girlsare 27.68% more likely to suffer from genderinequality in access to school, relative to girlsof primary school age. The situation becomeseven worse when girls enter tertiary schools,where the probability that a study finds a statisti-cally significant gender disparity in educationalattainment increases by 32.42% when girls entertertiary school. Beyond the compulsory educa-tion system, therefore, our evidence suggeststhat gender inequality is still a significant prob-lem.

The regression results in Table 4 (column1) show no detectable difference across ethnicgroups in the probability that a study findsgender inequality. However, this may be dueto the limited number of studies that considerthe gender gap among minorities. Indeed, whenwe use the weighted marginal probit regression(column 2), the coefficient of Ethnicity becomespositive and significant, indicating that studiesof minority girls are 27.29% more likely to findevidence of gender disparities than are studiesof Han girls.

Although it makes some of the interpre-tation of the results less intuitive, using theactual year instead of a decade dummy retainsmore of the information in our meta-analysis.This alternative way of coding the data is pre-sented in Table A1. The only difference betweenTables A1 and 4 is that we use a variable called

“year.” This variable is coded as the year inwhich the study’s data are collected. Whenwe replace the decade dummies with “year,”the nature of our findings does not change.The coefficient of the “year” variable is neg-ative and statistically significant in the mostbasic version of the equation, implying thatover time, gender inequality in China has beenfalling.7

V. CONCLUSION

In this article, we review the existing liter-ature on the gender inequality in education inChina. We investigated 55 articles covering 167studies from the 1980s to present day. Meta-regression analysis allows us to review and com-pare these studies in a concise and systematicway and offers more convincing evidence forthe change in gender inequality against girls.

What have we learned about gender inequal-ity in educational attainment from the meta-regression analysis? Inequality against girls stillexists in China today. However, our analysissuggests the existence of a downward trendover time. Girls’ access to education improvednoticeably with China’s economic developmentduring the 1980s, 1990s, and 2000s, whichwas concomitant with a series of govern-ment policies which addressed issues that likelyaffected education inequality. Gender inequal-ity in educational attainment varies betweenurban and rural areas. In urban areas, genderinequality reduced dramatically and has nowall but disappeared; indeed, urban girls seemto have advantages in educational opportuni-ties. By contrast, the educational penalty forliving in a rural area is substantially greaterfor girls than boys, and somewhat greater forminorities than for Han. There is nearly nogender inequality against girls within the com-pulsory education system, even in poor areas.Beyond the compulsory level, however, genderis still linked to educational attainment. Girlsare still significantly less likely to matriculate tosenior high school than are boys, and they areless represented in higher education. In short,females from rural areas—especially minoritiesand all rural girls attending high school andabove—face the greatest obstacles to enrollmentin schools.

7. When we replace the decade dummies with year, andinteract year with the key variables of interest, the nature ofour findings does not change fundamentally.


APPENDIX

TABLE A1Marginal Probit Regression Analysis of Gender Inequality in Educational Attainment

(Using Years as Variables)

(1) (2) (3) (4) (5)

Dependent Variable: Gender Inequality against Girls (1 = Yes, 0 = Neutral or against Boys)

(1) Time (years) −2.33∗∗∗(5.67)

−2.56∗∗∗(4.35)

−2.82∗∗∗(3.19)

−2.57∗∗∗(4.09)

−5.65∗∗∗(3.63)

(2) Area (ref. = rural areas)

Urban −42.49∗∗∗(4.88)

−40.52∗∗∗(3.50)

−49.02∗∗(1.96)

−40.56∗∗∗(3.45)

−33.95∗∗∗(3.00)

Nationwide 4.84(0.67)

3.81(0.38)

−0.60(0.02)

3.76(0.37)

6.62(0.70)

(3) Grade level (ref. = elementary)

Lower secondary 10.39(1.19)

8.47(0.77)

9.06(0.82)

8.4(0.76)

−38.68(1.26)

Upper secondary 24.73∗∗(2.49)

23.97∗∗(1.98)

24.55∗∗(1.97)

23.89∗∗(1.97)

−39.99(1.40)

Tertiary 37.75∗∗∗(3.37)

51.85∗∗∗(4.13)

51.60∗∗∗(4.21)

51.81∗∗∗(4.12)

27.1(0.96)

(4) Ethnicity (ref. = Han) 5.68(0.47)

22.80∗(1.84)

22.73∗(1.81)

21.61(0.71)

23.88∗(1.82)

(5) Interaction Time × Area

Time × urban 0.61(0.41)

Time × nationwide 0.27(0.18)

(6) Interaction Time × Ethnicity

Time × non-Han 0.07(0.05)

(7) Interaction Time × Grade

Time × lower secondary 3.71∗(1.75)

Time × upper secondary 4.83∗∗(2.45)

Time × tertiary 2.34(1.26)

Observations 167 167 167 167 167

Notes: Robust t-statistics are in parentheses. In column 2, we use the inverse of number of the studies from a singlepaper as the weight. In columns 3–5, we do the interaction for the variables of Time and Area, Ethnicity, and Grade usingthe weighted marginal probit model. ref., reference.

∗Significant at 10%; ∗∗significant at 5%; ∗∗∗significant at 1%.


TA

BL

EA

2L

ist

ofPa

pers

Use

din

the

Met

a-A

naly

sis

No.

Pap

erN

ame

Aut

hor

Sour

ce

1A

Stud

yof

Edu

catio

nal

Dis

pari

tyin

Urb

anC

hina

(194

9–

2003

)(i

nC

hine

se)

Dah

aiH

aoSo

cial

Scie

nce

inC

hina

,6,

2007

2A

cces

sto

Hig

her

Edu

catio

n:Ta

rget

edR

ecru

itmen

tR

efor

mun

der

Eco

nom

icD

evel

opm

ent

Plan

sin

the

Peop

le’s

Rep

ublic

ofC

hina

Vilm

aSe

eber

gH

ighe

rE

duca

tion

,25

(2),

1993

,16

9–

88

3A

dole

scen

tT

rans

ition

sto

Adu

lthoo

din

Ref

orm

-era

Chi

naE

mily

Han

num

and

Jiho

ngL

iuN

atio

nal

Res

earc

hC

ounc

il.T

heC

hang

ing

Tran

siti

ons

toA

dult

hood

inD

evel

opin

gC

ount

ries

:Se

lect

edSt

udie

s,ed

ited

byC

ynth

iaL

loyd

,Je

reB

ehrm

an,

Nel

lySt

rom

quis

t,an

dB

arne

yC

ohen

.W

ashi

ngto

nD

C:

The

Nat

iona

lA

cade

mie

sPr

ess,

2005

4B

asic

Edu

catio

nin

Chi

na’s

Rur

alA

reas

:A

Leg

alO

blig

atio

nor

anIn

divi

dual

Cho

ice?

Feng

shu

Liu

Inte

rnat

iona

lJo

urna

lof

Edu

cati

onal

Dev

elop

men

t,24

,20

09,

5–

21

5C

hang

esin

Edu

catio

nal

Ineq

ualit

yin

Chi

na,

1990

–20

05:

Evi

denc

efr

omth

ePo

pula

tion

Cen

sus

Dat

aX

iaog

ang

Wu

and

Zhu

oni

Zha

ngG

loba

liza

tion

,Cha

ngin

gD

emog

raph

ics,

and

Edu

cati

onal

Cha

llen

ges

inE

ast

Asi

aR

esea

rch

inSo

ciol

ogy

ofE

duca

tion

,17

,20

10,

123

–52

6C

hild

ren’

sSo

cial

Wel

fare

inC

hina

,19

89–

1997

:A

cces

sto

Hea

lthIn

sura

nce

and

Edu

catio

nJe

nnif

erA

dam

san

dE

mily

Han

num

The

Chi

naQ

uart

erly

,18

1,20

05,

100

–21

7C

ultu

ral

Cap

ital,

the

Stat

e,an

dE

duca

tiona

lIn

equa

lity

inC

hina

,19

49–

1996

Yux

iao

Wu

Soci

olog

ical

Per

spec

tive

s,51

(1),

2008

,20

1–

27

8D

eter

min

ants

ofSc

hool

Enr

ollm

ent

and

Com

plet

ion

of10

to18

Yea

rO

lds

inC

hina

Rac

hel

Con

nelly

and

Zhe

nzhe

nZ

heng

Eco

nom

ics

ofE

duca

tion

Rev

iew

,22

,20

03,

379

–88

9E

cono

mic

Tra

nsiti

on,

Scho

olE

xpan

sion

and

Edu

catio

nal

Ineq

ualit

yin

Chi

na,

1990

–20

00X

iaog

ang

Wu

Res

earc

hin

Soci

alSt

rati

ficat

ion

and

Mob

ilit

y,28

,20

10,

91–

108

10E

duca

tion

and

Ear

ning

sin

the

Peop

le’s

Rep

ublic

ofC

hina

Dea

nJa

mis

onan

dJa

cque

sV

anD

erG

aag

Eco

nom

ics

ofE

duca

tion

Rev

iew

,6(

2),

1987

,16

1–

66

11E

duca

tion

and

Pove

rty

inR

ural

Chi

naPh

ilip

Bro

wn

and

Alb

ert

Park

Eco

nom

ics

ofE

duca

tion

Rev

iew

,21

,20

02,

523

–41

12E

duca

tion

and

the

Pove

rty

Tra

pin

Rur

alC

hina

John

Kni

ght,

Li

Shi,

and

Den

gQ

uhen

gR

esea

rch

Con

sort

ium

onE

duca

tiona

lO

utco

mes

and

Pove

rty,

RE

CO

UP

Wor

king

Pape

r,N

o.11

,20

0713

Edu

catio

nin

the

Ref

orm

Era

Em

ilyH

annu

m,

Jere

Beh

rman

,M

eiya

nW

ang,

and

Jiho

ngL

iu

InC

hina

’sG

reat

Eco

nom

icTr

ansf

orm

atio

n,ed

ited

byL

oren

Bra

ndt

and

Tho

mas

Raw

ski,

Cam

brid

geU

nive

rsity

Pres

s,20

07

14E

duca

tion,

Gen

der

and

Isla

min

Chi

na:

The

Plac

eof

Rel

igio

usE

duca

tion

inC

halle

ngin

gan

dSu

stai

ning

“Und

ispu

ted

Tra

ditio

ns”

amon

gC

hine

seM

uslim

Wom

en

Mar

iaJa

scho

k,H

auM

ing,

and

Vic

kyC

han

Inte

rnat

iona

lJo

urna

lof

Edu

cati

onal

Dev

elop

men

t,29

,20

09,

487

–94


TA

BL

EA

2C

ontin

ued

No.

Pap

erN

ame

Aut

hor

Sour

ce

15E

duca

tiona

lA

cces

sfo

rC

hina

’sPo

st-C

ultu

ral

Rev

olut

ion

Gen

erat

ion

Enr

ollm

ent

Patte

rns

in19

90R

ache

lC

onne

llyan

dZ

henz

hen

Zhe

ngIn

Edu

cati

onan

dR

efor

min

Chi

na,

edite

dby

Em

ilyH

annu

man

dA

lber

tPa

rk.

Lon

don:

Rou

tledg

e,20

0716

Edu

catio

nal

Atta

inm

ent

and

Rur

al-U

rban

Div

ide

inC

hina

John

Kni

ght

and

Li

Shi

Oxf

ord

Bul

leti

nof

Eco

nom

ics

and

Stat

isti

cs,

58(1

),19

9617

Edu

catio

nal

Stra

tifica

tion

byE

thni

city

inC

hina

:E

nrol

lmen

tan

dA

ttain

men

tin

the

Ear

lyR

efor

mY

ears

Em

ilyH

annu

mD

emog

raph

y,39

(1),

2002

,95

–11

7

18E

duca

tiona

lSt

ratifi

catio

nin

Urb

anC

hina

:19

49–

1994

Xue

guan

gZ

hou,

Phyl

lisM

oen,

and

Nan

cyT

uma

Soci

olog

yof

Edu

cati

on,

71(3

),19

98,

199

–22

2

19E

nrol

lmen

tan

dG

radu

atio

nPa

ttern

sas

Chi

na’s

Ref

orm

sD

eepe

n,19

90–

2000

Xue

guan

gZ

hou,

Phyl

lisM

oen,

and

Nan

cyB

rand

onT

uma

InE

duca

tion

and

Ref

orm

inC

hina

,ed

ited

byE

mily

Han

num

and

Alb

ert

Park

.L

ondo

n:R

outle

dge,

2007

20E

thni

cG

roup

san

dE

duca

tiona

lIn

equa

litie

s:A

nE

mpi

rica

lSt

udy

ofth

eE

duca

tiona

lA

ttain

men

tof

the

Eth

nic

Min

oriti

esin

Wes

tern

Chi

na(i

nC

hine

se)

Rac

hel

Con

nelly

and

Zhe

nzhe

nZ

heng

Soci

ety,

2,20

10

21E

thni

cSt

ratifi

catio

nin

Nor

thw

est

Chi

na:

Occ

upat

iona

lD

iffe

renc

esbe

twee

nH

anC

hine

sean

dN

atio

nal

Min

oriti

esin

Xin

jiang

,19

82–

1990

Em

ilyH

annu

man

dY

uX

ieD

emog

raph

y,35

(3),

1998

,32

3–

33

22E

xam

inat

ions

and

Edu

catio

nal

Opp

ortu

nity

inC

hina

:M

obili

tyan

dB

ottle

neck

sfo

rth

eR

ural

Poor

Em

ilyH

annu

m,

Xue

hui

An,

Hua

Yu,

and

Seba

stia

nC

hern

gO

xfor

dR

evie

wof

Edu

cati

on,

37(2

),20

11,

267

–30

5

23E

xpan

sion

ofH

ighe

rE

duca

tion

and

Ineq

ualit

yin

Opp

ortu

nity

ofE

duca

tion:

ASt

udy

onE

ffec

tof

“Kuo

zhao

”Po

licy

onE

qual

izat

ion

ofE

duca

tiona

lA

ttain

men

t(i

nC

hine

se)

Chu

nlin

Li

Soci

olog

ical

Stud

ies,

3,20

10

24E

xpan

sion

ofH

ighe

rE

duca

tion

inC

hina

and

Ineq

ualit

yin

Ent

ranc

eO

ppor

tuni

ties:

1978

-200

3(i

nC

hine

se)

Jing

min

gL

iuSo

ciet

y,3,

2006

,15

8–

79

25Fa

mily

Bac

kgro

und,

Fina

ncia

lC

onst

rain

tsan

dH

ighe

rE

duca

tion

Atte

ndan

cein

Chi

naW

enli

Li

Eco

nom

ics

ofE

duca

tion

Rev

iew

,26

(6),

2007

,72

4–

34

26Fa

mily

Bac

kgro

und,

Gen

der

and

Edu

catio

nal

Atta

inm

ent

inU

rban

Chi

naM

ontg

omer

yB

road

edan

dC

hong

shun

Liu

The

Chi

naQ

uart

erly

,14

5,19

96,

53–

86

27Fa

mily

Sour

ces

ofE

duca

tiona

lG

ende

rIn

equa

lity

inR

ural

Chi

na:

AC

ritic

alA

sses

smen

tE

mily

Han

num

a,Pe

ggy

Kon

g,an

dY

upin

gZ

hang

Inte

rnat

iona

lJo

urna

lof

Edu

cati

onal

Dev

elop

men

t,29

,20

09,

474

–86

28G

ende

rIn

equa

lity

and

Hig

her

Edu

catio

nJe

rry

Jaco

bsA

nnua

lR

evie

wof

Soci

olog

y,22

,19

96,

153

–85


TA

BL

EA

2C

ontin

ued

No.

Pap

erN

ame

Aut

hor

Sour

ce

29G

ende

rIn

equa

lity

inU

rban

Chi

na:

Edu

catio

nan

dE

mpl

oym

ent

John

Bau

er,

Wan

gFe

ng,

Nan

cyR

iley,

and

Xia

ohua

Zha

oM

oder

nC

hina

,18

,19

92,

333

30G

ende

red

Path

way

sto

Rur

alSc

hool

ing:

The

Inte

rpla

yof

Wea

lthan

dL

ocal

Inst

itutio

nsD

ebor

ahD

avis

,Pi

erre

Lan

dry,

Yus

heng

Peng

,an

dJi

nX

iao

The

Chi

naQ

uart

erly

,89

,20

07,

60–

82

31G

irls

’E

duca

tion

inG

ansu

,C

hina

Huh

uaC

aoan

dFe

nglia

nL

eiSe

ptem

ber

6th,

post

edat

http

://w

ww

.chi

naea

m.u

otta

wa.

ca/ic

as5/

book

/22.

pdf

32G

irls

inG

ansu

,C

hina

:E

xpec

tatio

nsan

dA

spir

atio

nsfo

rSe

cond

ary

Scho

olin

gE

mily

Han

num

and

Jenn

ifer

Ada

ms

Gan

suSu

rvey

ofC

hild

ren

and

Fam

ilies

Pape

rs,

Uni

vers

ityof

Penn

sylv

ania

,20

0833

Gir

ls’

Acc

ess

toE

duca

tion

inC

hina

:A

ctor

s,C

ultu

res

and

the

Win

dmill

ofD

evel

opm

ent

Man

agem

ent

Xia

ojun

Wan

gC

onso

rtiu

mfo

rR

esea

rch

onE

duca

tiona

lA

cces

s,T

rans

ition

san

dE

quity

,C

reat

ePa

thw

ays

toA

cces

sR

esea

rch

Mon

ogra

ph,

No.

39,

2010

34H

ouse

hold

Dec

isio

nsan

dG

ende

rIn

equa

lity

inE

duca

tion

inR

ural

Chi

naD

anke

Li

and

Mun

Tsan

gC

hina

:A

nIn

tern

atio

nal

Jour

nal,

1(2)

,20

03,

224

–48

35In

equa

lity

inC

hine

seE

duca

tion

Xue

Lan

Ron

gan

dT

ianj

ian

Shi

Jour

nal

ofC

onte

mpo

rary

Chi

na,

10(2

6)20

01,

107

–24

36In

stitu

tion

and

Ineq

ualit

y:T

heH

ukou

Syst

emin

Chi

naZ

hiqi

ang

Liu

Jour

nal

ofC

ompa

rati

veE

cono

mic

s,33

,20

05,

133

–57

37In

tern

atio

nal

Mig

ratio

nan

dE

duca

tion

ofL

eft-

Beh

ind

Chi

ldre

nin

Fujia

n,C

hina

Hid

eki

Mor

ooka

and

Zai

Lia

ngA

sian

and

Pac

ific

Mig

rati

onJo

urna

l,18

(3),

2009

,34

5–

70

38M

arke

tT

rans

ition

,E

duca

tiona

lD

ispa

ritie

s,an

dFa

mily

Stra

tegi

esin

Rur

alC

hina

:N

ewE

vide

nce

onG

ende

rSt

ratifi

catio

nan

dD

evel

opm

ent

Em

ilyH

annu

mD

emog

raph

y,42

(2),

2005

,27

5–

99

39M

igra

ntO

ppor

tuni

tyan

dth

eE

duca

tiona

lA

ttain

men

tof

You

thin

Rur

alC

hina

Ala

nde

Bra

uwan

dJo

hnG

iles

IZA

Dis

cuss

ion

Pape

r,N

o.23

26,

2006

,po

sted

atht

tps:

//ww

w.m

su.e

du/∼

gile

sj/a

djg1

final

June

5_20

08.p

df40

Pare

ntal

Inve

stm

ent

inC

hild

ren’

sH

uman

Cap

ital

inU

rban

Chi

naL

inda

Yue

hA

ppli

edE

cono

mic

s,38

,20

06,

2089

–11

141

Phys

ical

Wel

l-B

eing

and

Scho

olE

nrol

lmen

t:A

Com

pari

son

ofA

dopt

edan

dB

iolo

gica

lC

hild

ren

inO

ne-C

hild

Fam

ilies

inC

hina

Jiho

ngL

iua,

Gra

ceW

ysha

k,an

dU

llaL

arse

nSo

cial

Scie

nce

and

Med

icin

e,59

,20

04,

609

–23

42Po

litic

alC

hang

ean

dth

eU

rban

-Rur

alG

apin

Bas

icE

duca

tion

inC

hina

,19

49–

1990

Em

ilyH

annu

mC

ompa

rati

veE

duca

tion

Rev

iew

,43

(2),

1999

,19

3–

211


TA

BL

EA

2C

ontin

ued

No.

Pap

erN

ame

Aut

hor

Sour

ce

43Po

vert

yan

dB

asic

Edu

catio

nin

Rur

alC

hina

:V

illag

es,

Hou

seho

lds,

and

Gir

ls’

and

Boy

s’E

nrol

lmen

tE

mily

Han

num

Com

para

tive

Edu

cati

onR

evie

w,

47(2

),20

03

44R

esea

rch

into

Gir

ls’

Edu

catio

nin

Four

Wes

tern

Prov

ince

sof

Chi

naW

eiZ

hou,

Tie

dao

Zha

ng,

Wen

puL

iu,

Yuq

inM

a,an

dJi

ayun

Peng

Chi

nese

Edu

cati

onan

dSo

ciet

y,33

(5),

2001

,4

–28

45So

nPr

efer

ence

and

Edu

catio

nal

Opp

ortu

nitie

sof

Chi

ldre

nin

Chi

na—

“IW

ish

You

Wer

ea

Boy

!”W

endy

Wan

gG

ende

rIs

sues

,22

(2),

2005

,3

–30

46T

heE

duca

tiona

lC

onse

quen

ces

ofM

igra

tion

for

Chi

ldre

nin

Chi

naZ

aiL

iang

and

Yiu

Por

Che

nSo

cial

Scie

nce

Res

earc

h,36

,20

07,

28–

4747

The

Eff

ect

ofPa

rent

alL

abor

Mig

ratio

non

Chi

ldre

n’s

Edu

catio

nal

Prog

ress

inR

ural

Chi

naC

had

Mey

erho

efer

and

C.

J.C

hen

Rev

iew

ofE

cono

mic

sof

the

Hou

seho

ld,

9(3)

,20

11,

379

–96

48T

heE

ffec

tof

Sibs

hip

Size

onE

duca

tiona

lA

ttain

men

tin

Chi

na:

Peri

odV

aria

tions

Yao

Lu

and

Don

ald

Tre

iman

Am

eric

anSo

ciol

ogic

alR

evie

w,

73(5

),20

08,

813

–34

49T

heE

ffec

tsof

the

Gen

der

ofC

hild

ren

onE

xpen

ditu

rePa

ttern

sin

Rur

alC

hina

:A

Sem

iPa

ram

etri

cA

naly

sis

Xia

odon

gG

ong,

Art

hur

Van

Soes

t,an

dPi

ngZ

hang

Jour

nal

ofA

ppli

edE

cono

met

ric,

20,

2005

,50

9–

27

50T

heIn

equa

lity

ofO

ppor

tuni

tyan

dIt

sC

hang

ing

inth

eB

asic

Edu

catio

nin

Chi

na(i

nC

hine

se)

Liu

Jing

min

gSo

cial

Scie

nce

inC

hina

,20

08,

5

51T

heO

ne-C

hild

Polic

yan

dSc

hool

Atte

ndan

cein

Chi

naJu

hua

Yan

gC

ompa

rati

veE

duca

tion

Rev

iew

,51

(4),

2007

,47

1–

9552

Urb

an-R

ural

Dis

pari

ties

inA

cces

sto

Prim

ary

and

Seco

ndar

yE

duca

tion

unde

rM

arke

tR

efor

ms

Em

ilyH

annu

m,

Mei

yan

Wan

g,an

dJe

nnif

erA

dam

sO

neC

ount

ry,T

wo

Soci

etie

s?R

ural

-Urb

anIn

equa

lity

inC

onte

mpo

rary

Chi

na,

edite

dby

Mar

tinK

ing

Why

te,

Har

vard

Uni

vers

ityPr

ess,

2008

53W

hat

Det

erm

ines

Bas

icSc

hool

Atta

inm

ent

inD

evel

opin

gC

ount

ries

?E

vide

nce

from

Rur

alC

hina

Men

gZ

hao

and

Paul

Gle

ww

eE

cono

mic

sof

Edu

cati

onR

evie

w,

29,

2010

,45

1–

60

54W

hoG

ets

Mor

eFi

nanc

ial

Aid

inC

hina

?A

Mul

tilev

elA

naly

sis

PoY

ang

Inte

rnat

iona

lJo

urna

lof

Edu

cati

onal

Dev

elop

men

t,30

,20

10,

560

–69

55W

hyD

oG

irls

inR

ural

Chi

naH

ave

Low

erSc

hool

Enr

ollm

ent?

Lin

aSo

ngan

dSi

mon

App

leto

nW

orld

Dev

elop

men

t,34

(9),

2006

,16

39–

53


REFERENCES

Brandt, L., and T. Rawski. “China’s Great Economic Trans-formation,” in China’s Great Economic Transforma-tion, edited by L. Brandt and T. Rawski. Cambridge:Cambridge University Press, 2008.

Brown, P., and A. Park. “Education and Poverty in RuralChina.” Economics of Education Review, 21, 2002,523–41.

Cai, F., A. Park, and Y. Zhao. “The Chinese Labor Marketin the Reform Era,” in China’s Great Economic Trans-formation, edited by L. Brandt and T. Rawski. Cam-bridge: Cambridge University Press, 2008, 167–214.

Cao, H., and F. Lei. “Girls Education in Gansu.” September6, 2008. Accessed September 10, 2011. http://www.chinaeam.uottawa.ca/icas5/book/22.pdf.

Crook, F. W. “Underreporting of China’s Cultivated LandArea: Implications for World Agricultural Trade,” inChina: Situation and Outlook Series: USDA Inter-national Agricultural Trade Reports, edited by U.S.Department of Agriculture. Washington, DC: U.S.Department of Agriculture, Economic Research Ser-vice, 1993.

Davis, D., P. Landry, Y. Peng, and J. Xiao. “GenderedPathways to Rural Schooling: The Interplay of Wealthand Local Institutions.” The China Quarterly, 189,2007, 60–82.

deBrauw, A., and S. Rozelle. “Reconciling the Returns toEducation in Off-Farm Wage Employment in RuralChina.” Review of Development Economics, 12(1),2008, 57–71.

Fleisher, B. M., H. Li, S. Li, and X. Wang. “Sorting, Selec-tion, and Transformation of the Return to CollegeEducation in China.” IZA Discussion Paper No. 1446,2004.

Glewwe, P., and M. Kremer. “Schools, Teachers and Educa-tion Outcomes in Developing Countries,” in Handbookof the Economics of Education, edited by E. Hanushek,S. Machin, and L. Woessmann. Amsterdam: NorthHolland, 2006.

Gustafsson, B., and S. Li. “The Ethnic Minority–MajorityIncome Gap in Rural China during Transition.” Eco-nomic Development and Cultural Change, 51(4), 2003,805–22.

Hannum, E., J. Behrman, M. Wang, and J. Liu. “Educationin the Reform Era,” in China’s Great Economic Trans-formation, edited by L. Brandt and T. Rawski. Cam-bridge: Cambridge University Press, 2008,215–49.

Hannum, E., M. Wang, and J. Adams. “Urban–Rural Dis-parities in Access to Primary and Secondary Educationunder Market Reforms. One Country, Two Societies?”in Rural–Urban Inequality in Contemporary China,edited by M.K. Whyte. Cambridge, MA: Harvard Uni-versity Press, 2008.

Hong, Y. “Ethnic Groups and Educational Inequalities: AnEmpirical Study of the Educational Attainment of theEthnic Minorities in Western China.” Social ChineseJournal of Sociology, 30(2), 2010, 45–73.

Knight, J., and S. Li. “Educational Attainment and theRural–Urban Divide in China.” Oxford Bulletin ofEconomics and Statistics, 58(1), 1996, 83–117.

Li, C. “Expansion of Higher Education and Inequality inOpportunity of Education: A Study on the Effectof “Kuozhao” Policy on Equalization of EducationalAttainment.” Sociological Studies, 3, 2010, 82–113.

Liu, C., L. Zhang, R. Luo, S. Rozelle, B. Sharbono, and Y.Shi. “Development Challenges, Tuition Barriers andHigh School Education in China.” Center for ChineseAgricultural Policy, Chinese Academy of Science,Working Paper 2009-01-23, 2009.

Liu, C., L. Zhang, R. Luo, S. Rozelle, and P. Loyalka. “TheEffect of Primary School Mergers on Academic Perfor-mance of Students in Rural China.” International Jour-nal of Educational Development, 30, 2010, 570–85.

Liu, C., L. Zhang, R. Luo, X. Wang, S. Rozelle, B. Shar-bono, J. Adams, Y. Shi, A. Yue, H. Li, and T. Glauben.“Early Commitment on Financial Aid and CollegeDecision Making of Poor Students: Evidence from aRandomized Evaluation in Rural China.” Economicsof Education Review, 2, 2011, 1–13.

Liu, F. “Basic Education in China’s Rural Areas: ALegal Obligation or an Individual Choice?” Interna-tional Journal of Educational Development, 24, 2004,5–21.

Ma, H., J. Huang, and S. Rozelle. “Reassessing China’sLivestock Statistics—An Analysis of Discrepanciesand the Creation of New Data Series.” EconomicDevelopment and Cultural Change, 52(2), 2004,445–73.

Mo, D., H. Yi, L. Zhang, R. Luo, S. Rozelle, and C. Brin-ton. “Transfer Paths and Academic Performance: ThePrimary School Merger Program in China.” LICOSDiscussion Paper Series. Discussion Paper 283, 2011.

National Bureau of Statistics of China (NBSC). China Sta-tistical Yearbook. Beijing: China Statistics Press, 1997.

. China Statistical Yearbook. Beijing: China StatisticsPress, 2001.



Naughton, B. Growing Out of the Plan: China’s EconomicReform, 1978–1993. Cambridge: Cambridge Univer-sity Press, 1994.

Phillips, J. “Farmer Education and Farmer Efficiency: aMeta-analysis.” Economic Development and CulturalChange, 43, 1994, 149–65.

Rawski, T. G. “What Is Happening to China’s GDP Statis-tics?” China Economic Review, 12(4), 2001, 347–54.

Song, L., and S. Appleton. “Why Do Girls in Rural ChinaHave Lower School Enrollment?” World Development,34(9), 2006, 1639–53.

Stanley, T. D. “Wheat from Chaff: Meta-Analysis as Quan-titative Literature Review.” Journal of Economic Per-spectives, 15(3), 2001, 131–50.

UNESCO. “Children Out of School: Measuring Exclu-sion from Primary Education.” UNESCO Institute forStatistics, 2005.

. “Education for All by 2015. Will We Makeit?” EFA Global Monitoring Report, United NationsEducational, Scientific and Cultural Organization,2008.

. “Comparing Education Statistics across the World:Focus on Secondary Education.” Global EducationDigest, UNESCO Institute for Statistics, 2011.

Wang, X. G. “Girls’ Access to Education in China: Actors,Cultures and the Windmill of Development Man-agement.” Consortium for Research on EducationalAccess, Transitions and Equity, Create Pathways toAccess Research Monograph, No. 39, 2010.

Wang, X., C. Liu, L. Zhang, R. Luo, T. Glauben, Y. Shi,S. Rozelle, and B. Sharbono. “What Is Keeping thePoor Out of College? Enrollment Rates, EducationalBarriers and College Matriculation in China.” ChinaAgricultural Economic Review, 3(2), 2011, 131–49.

Wong, C. P. “Central–Local Relations in an Era of FiscalDecline: The Paradox of Fiscal Decentralization inPost-Mao China.” The China Quarterly, 128, 1991,691–715.


Wong, C. P., and R. M. Bird. “China’s Fiscal System: AWork in Progress,” in China’s Great Economic Trans-formation, edited by L. Brandt and T. Rawski. Cam-bridge: Cambridge University Press, 2008, 429–66.

World Bank. “The World Bank Education Statistics Newslet-ter.” August, 2011, V(1).

. “World Development Report 2012: Gender andDevelopment.” Washington DC: World Bank, 2012.

Wu, X., and Z. Zhang. “Changes in Educational Inequal-ity in China, 1990–2005: Evidence from the Popula-tion Census Data,” in Globalization, Changing Demo-graphics, and Educational Challenges in East AsiaResearch in Sociology of Education, Vol. 17, edited byE. Hannum, H. Park, and Y. G. Butler. Bingley, UK:Emerald Group Publishing, 2010, 123–52.

Yang, J. “The One-Child Policy and School Attendance inChina.” Comparative Education Review, 51(4), 2007,471–95.

Zhang, J., and Y. Zhao. “Rising Schooling Returns in UrbanChina,” in Education and Reform in China, editedby E. Hannum and A. Park. London; New York:Routledge, 2007.

Zhang, L., J. Huang, and S. Rozelle. “Employment, Emerg-ing Labor Markets, and the Role of Education inRural China.” China Economic Review, 13, 2002,313–28.

Zhao, M., and P. Glewwe. “What Determines Basic SchoolAttainment in Developing Countries? Evidence fromRural China.” Economics of Education Review, 29,2010, 451–60.

GENDER INEQUALITY IN EDUCATION IN CHINA: A META...

Documents

Transcript of GENDER INEQUALITY IN EDUCATION IN CHINA: A META...