WORKING PAPER SERIES NO. 2007-3 40251 The World Bank Human Development Sector Unit 1818 H Street, NW East Asia and the Pacific Region Washington, DC 20433 The World Bank USA Telephone: 202-473-1000 Facsimile: 202-477-6391 East Asia and Pacific Region Human Development Sector Unit http://www.worldbank.org CHINA Educational Resources and Impediments in Rural Gansu, China Elizabeth King Lynnette de la Cruz Perez Mario Taguiwalo with Yolanda Quitero May 2007 Working Paper Series on CHINA Educational Resources and Impediments in Rural Gansu, China Paper No. 2007-3 Emily Hannum and Peggy Kong The World Bank Washington, DC May 2007 Data collection for the Gansu Survey of Children and Families was supported by grants from The Spencer Foundation Small and Major Grants Programs (wave 1), by NIH Grants 1R01TW005930-01 and 5R01TW005930-02 (wave 2), and by a grant from the World Bank (wave 2). We thank Tanja Sargent for helpful comments. Currency Equivalents (Exchange Rate Effective May 2007) Currency Unit = China Yuan Renminbi (CNY) 1 CNY = US$ 0.13044 US$1 = 7.66610 CNY Fiscal Year January 1 to December 31 Regional Vice President: James W. Adams, EAPVP Country Director: David Dollar, EACCF Sector Director: Emmanuel Jimenez, EASHD Sector Manager: Christopher Thomas, EASHD Task Team Leader: Kin Bing Wu, EASHD TABLE OF CONTENTS EXECUTIVE SUMMARY ................................................................................................ I INTRODUCTION .............................................................................................................1 BACKGROUND................................................................................................................3 THE NATURE OF RURAL SCHOOLS........................................................................12 Basic Facilities..................................................................................................................12 Primary Schools......................................................................................................12 Junior High Schools...............................................................................................15 Financial Arrangements.................................................................................................16 Household Expenditures on Education..............................................................16 Tuition and Fees Reported by Schools................................................................18 Public, Private, and Hybrid Schools....................................................................24 Personnel Costs Reported by Schools.................................................................26 Teachers............................................................................................................................29 Primary Schools......................................................................................................29 Qualifications and Status......................................................................................32 Workloads and Work Activities ..........................................................................36 Professional Development....................................................................................38 Satisfaction with Various Dimensions of Work.................................................39 Attitudes about School Management and School Culture...............................42 Junior High Schools...............................................................................................47 Qualifications and Status......................................................................................48 Workloads and Work Activities ..........................................................................48 Professional Development....................................................................................49 Satisfaction with Various Dimensions of Work.................................................50 Attitudes about School Management and School Culture...............................51 Classroom Learning Environments..............................................................................52 Children's Perspectives.........................................................................................53 Teachers' Perspectives...........................................................................................61 SCHOOL PERSISTENCE: BARRIERS AND SUPPORTS..........................................64 Informant Reports on Reasons for School-Leaving ...................................................65 Analysis of Outcomes in 2004 .......................................................................................71 Strategy....................................................................................................................72 Change-In-Attainment Results ............................................................................75 Enrollment Results.................................................................................................78 Nine-Year Achievement Results..........................................................................79 Engagement and Performance as Predictors of Persistence............................80 Summary.................................................................................................................81 Determinants of Engagement and Performance ........................................................82 Strategy....................................................................................................................83 Engagement Results ..............................................................................................84 Performance Results..............................................................................................86 Summary.................................................................................................................87 CONCLUSIONS..............................................................................................................88 Finance Problems............................................................................................................88 Infrastructure issues .......................................................................................................91 Teacher Morale................................................................................................................94 Teaching and Learning...................................................................................................95 Student Engagement with Education and Performance...........................................97 Implications .....................................................................................................................99 REFERENCES CITED...................................................................................................101 APPENDIX.....................................................................................................................104 Educational Resources and Impediments in Rural Gansu, China, 2000-2004 EXECUTIVE SUMMARY This report seeks to provide a portrait of schools serving rural communities in northwest China, and to shed light on factors that encourage and discourage school persistence among children in this region. To achieve these goals, we analyze a survey of rural children and their families, schools, and teachers in Gansu province. The project interviewed children in the year 2000, when children were 9 to 12 years old, and again four years later. In part one of the paper, we provide a descriptive overview of the material, human, and cultural resources available in sampled primary and middle schools. Where possible, we note changes between 2000 and 2004. We describe the following types of resources: (1) basic facilities; (2) financial arrangements; (3) teachers, including their background, qualifications, working lives, professional development activities, satisfaction with work, and attitudes about school management and culture; and (4) classroom environments, as reported by teachers and by students. In this descriptive section of the paper, we highlight basic infrastructure issues, the complexity of financial arrangements at the time of the surveys, problems of teacher wage arrears and teacher morale, and the i pedagogies and learning environments in classrooms, as reported by teachers and students. In part two of the paper, we investigate reasons for school leaving reported by village leaders, families, and children themselves, and analyze contributors to subsequent enrollment, change in attainment, and attainment of nine years of compulsory education. Our models of family, teacher, and school effects on outcomes show that higher socio-economic status children are more likely to show grade attainment, continued enrollment, and attainment of nine years of basic education. In contrast, the gender story is mixed: girls are less likely to be enrolled, but have not gained less grades, nor are they less likely to achieve nine years of education. This finding suggests that boys may start later or repeat more. It is possible that boys are more likely to be encouraged to repeat a grade to complete it successfully or to increase high school exam scores. One significant finding is that the introduction of school and teacher effects, by and large, does not explain away the advantages of children in better- off families. School and teacher effects do not consistently matter across the three outcomes. Some interesting findings include that teacher absenteeism in 2000 is associated with less attainment between 2000 and 2004; children with better-paid home room teachers are more likely to attain nine years of school; and children in schools with minban teachers are less likely to attain nine years. ii However, there is not a consistent story of school characteristics that help or hinder childrens persistence. Reports by village-leaders, fathers, mothers, and children themselves indicate that, along with socioeconomic status, children's performance and engagement are significant factors in school continuation decisions in Gansu's rural villages. Multivariate analyses indicate that childrens early aspirations and performance matter for later outcomes. We close by discussing the most significant strengths and weaknesses identified among the school resources discussed in part one, and the most significant supports and hindrances to favorable educational outcomes considered in part two. iii Educational Resources and Impediments in Rural Gansu, China INTRODUCTION This report analyzes a longitudinal survey of rural children and schools in Gansu province, with two main goals. First, we seek to provide an overview of the material, human, and cultural resources available in primary and middle schools serving Gansu's rural communities. Second, we seek to offer insights on factors that support and hinder favorable educational outcomes of academic achievement, engagement with education, continued enrollment, and high parental aspirations. To achieve these goals, we analyze a survey of rural children and their families, schools, and teachers in Gansu province. The survey was implemented in the year 2000, when children were 9 to 12 years old,1 and again four years later, in 2004. We begin with a brief background section that explains why a better understanding of schools and school experiences in China's western regions is needed, from a policy perspective, and describes the data used in this report. 1For the sample of 2000 children selected based on registered birthdates, 10 households reported birthdates that made their exact ages less than 9, and 12 reported birthdates that made their exact ages greater than 13. Analyses that follow use ages rounded at mid-year to the nearest full year, rather than exact ages. 1 We then turn to the main empirical results. In part one of the main results, we provide an overview of the material, human, and cultural resources available in sampled primary and middle schools serving Gansu's rural communities. Where possible, we note changes between 2000 and 2004. We describe the following types of resources: (1) basic facilities; (2) financial arrangements; (3) teachers, including their background, qualifications, working lives, professional development activities, satisfaction with work, and attitudes about school management and culture; and (4) classroom environments, as reported by teachers and by students. In part two, we investigate reasons for school leaving reported by village leaders, families, and children themselves. Guided by these reports, we analyze performance and early attitudes about schooling. We then consider these factors, along with economic factors and school-level factors, as contributors to subsequent enrollment, change in attainment, and attainment of nine years of compulsory education. We close by discussing the most significant strengths and weaknesses identified among the school resources discussed in part one, and the most significant supports and hindrances to favorable educational outcomes considered in part two. 2 BACKGROUND The educational attainment of China's population has increased dramatically in the years since the establishment of the People's Republic in 1949. For example, estimates from a sample of the 2000 census indicate that just two percent of women in their 80s had a junior high school or higher level of education; among 25 to 29 year-olds, the figure was 68 percent (Hannum, Behrman, Wang, and Liu, forthcoming: Figure 2). The task of improving access has grown easier as school-aged cohorts have stopped expanding and begun to decline (see the population pyramids in Figure 1). The data underlying these pyramids shows, for example, that the number of 5 to 9 year-olds in 1995 was about 21 percent higher than in 1990. The 2000 number was about 18 percent lower than the 1995 number, and the 2005 number was about 9 percent lower than the 2000 number (calculated from United States Census Bureau, 2005). 3 Figure 1: China Population Pyramid by Year Yet, important gaps in access persist in some areas. In particular, educational attainments in China's northwest and southwest continue to lag behind. For example, among 25 to 34 year-olds in the 2000 census, about 8 percent of young adults in the northwest and about 7 percent in the southwest have not attended formal schooling at all; these figures compare to 2 percent or less for all other macro-regions in China (Hannum et al., forthcoming: Table 6). Rural children in the interior regions are at particularly high risk of early school leaving. China's government has long sought to remedy the persistent rural poverty in the west. Improving chances for social mobility via educational expansion in this region has been a key element of long-term efforts toward 4 poverty alleviation. Yet, while the educational disadvantage of children in the west is well-recognized, little systematic data has been collected about the educational conditions these children face. Moreover, beyond the important problem of school costs that burden poor rural families, factors that improve or reduce chances of school continuation are poorly understood. To address this dearth of information, we offer a description of the schools attended by rural children in Northwest China, and a preliminary analysis of factors--economic, performance-related, and attitude-related--that might matter for school continuation. We focus on rural areas in one province in the Northwest region: Gansu. Gansu stretches across parts of the Gobi desert, mountainous and hilly areas, and vast grasslands. Gansus per capita GDP in 2005 was 7,341 RMB (about $891)-- just about 16 percent of that of Beijings per capita GDP in the same year (44,969 RMB, about $5,457) (China Internet Information Center 2005a, b). Most of Gansu's geography is mountainous or highland plateau, with an elevation of more than 1,000 meters (UNESCAP, 2005). In the year 2000, Gansu province had a population of 25.62 million, 76 percent of whom resided in rural areas (UNESCAP, 2005)2. Importantly, rural residents in Gansu are among the poorest 2By 2005, the population was estimated to be 25.944 million (China Internet Information Center 2005b). 5 in China, with only rural residents in Guizhou having lower per capita income in recent years (National Bureau of Statistics (China) 2005a). In 2004, average rural per capita income was 60 percent of the national average, about 30 percent of that of Beijing, about one-fourth of that of Shanghai, and about the same as that of Tibet (National Bureau of Statistics (China) 2005a). According to statistics compiled by UNESCAP (2005), Gansu has the second-highest illiteracy rate in China. Gansu is experiencing a contraction of child cohorts. Figure 2 shows that while the overall population of Gansu increased by about 2.8 percent between 1999 and 2003, the population of children ages 0 to 14 declined by about 13.5 percent in the same period. 6 Figure 2. Gansu Total Population and Cohorts Ages 0 to 14 by Year 30000000 25000000 +2.8% 20000000 15000000 1999 2003 10000000 5000000 -13.5% 0 0-14 Population Estimate Total Population Estimate 1999 6936475.41 25378073.77 2003 5996945.01 26097759.67 Our analyses draw on a unique data set, The Gansu Survey of Children and Families (GSCF), a multi-level survey designed to increase understanding of rural children's schooling and welfare in the context of poverty. The GSCF includes two waves of data collected from the same sample of children in 2000 and 2004. 7 The GSCF-1 (2000) surveyed 2000 children aged 9 to 12 and their families in rural areas of 20 counties in Gansu Province. GSCF-2 re-visited the same children in 2004. Map 1 shows the distribution of the sample across counties in Gansu. The sample was drawn using a multi-stage, clustered design with random selection procedures employed at each stage (county, township, village, child). Several minority autonomous counties were excluded from the sampling frame 8 due to travel restrictions to these areas, language barriers, limited transportation, and sparse and dispersed populations in these counties.3 At the final stage, children were sampled from birth records in 100 selected villages. In China, the urban-rural designation is official, clearly defined, and consequential for access to services, and so drawing a sample of rural villages was a clear-cut task. In terms of income, our sample is broadly representative of rural Gansu: the per capita incomes of 46 percent of the households in our sample (or 920 households) were above the provincial average. The remaining 54 percent of households (or 1,080 households) had per capita incomes that were below the provincial average (Gansu Statistics Bureau, 2000). In addition to interviewing children and their parents, we collected data from all village leaders, all principals in primary schools, and all teachers in primary schools serving sample villages. In 2004, we added data from all principals and teachers in the main junior high schools serving sample villages. 3Unfortunately, the sample does not contain sufficient numbers of minority children for meaningful analysis. With this caveat, the GSCF is representative of children in rural areas of Gansu, and includes wealthier and poorer rural counties. 9 Table 1. Child Samples and Percent Enrolled in 2000 and 2004 Year 2000 2004 At ages 9-12* At ages 13-16 Total enumerated children (based on household reports) n=2000 n=1918 Among which, percent enrolled 98.85 86.81 Descriptive tables on child attitudes: sample of children with valid observations on the child questionnaire in both 2000 and 2004 n=1776 n=1776 Among which, percent enrolled 99.72 89.58 Multivariate analyses: Sample with valid observations on 2000 child, family, and school predictors. (Variability in sample size across multivariate tables stems from different numbers of valid observations on the dependent variables). n=1693 n=1693 Among which, percent enrolled 99.76 87.06 Source: GSCF-1 and GSCF-2. Note: Ns for enrollment status in the 2000 household data were actually 1775 and 1692. 2000 enrollment status is not used in subsequent analyses, which are based on the Ns listed in the table. *For the sample of 2000 children selected based on registered birthdates, 10 households reported birthdates that made their exact ages less than 9, and 12 reported birthdates that made their exact ages greater than 13. Analyses that follow use ages rounded at mid-year to the nearest full year, rather than exact ages. Table 1 shows our sample of children. In 2000, our sample included 2000 children, among whom virtually all (98.85 percent) were enrolled in school.4 In 2004, we re-visited households of 1,918 of these children, among which group 1,665 or about 87 percent were still enrolled in school, according to household 4This number is similar to the enrollment rate for school-aged children in Gansu reported by UNESCAP for the year 2000, at 98.83 percent. However, the UNESCAP number is almost certainly an enrollment ratio (enrollments from one source, base population from another), rather than a rate based on a survey of children. Whether it is a gross or net enrollment ratio, and the age range to which it refers, are not specified (UNESCAP, 2005). 10 reports. A number of these children were no longer living at home, and while we collected information from other family members about these children, we were unable to administer questionnaires to these children. For tables based on child questionnaires, we employ a sample with valid observations on all child questionnaires items in both years, totaling 1,776. For multivariate analyses, we employ a sample of children with valid observations on all predictors measured in 2000, plus valid observations on the outcome variables of interest, for a maximum sample size of 1,693 (estimation samples vary in size depending on the valid observations on outcome variables of interest). The enrollment rate among this group was also 87 percent. Table 2. Number of Schools, Teachers, and Villages in the Sample in 2000 and 2004 Year 2000 2004 Primary Schools 131 140 Junior High Schools 77 Primary School Teachers 1009 1018 Junior High School Teachers 1332 Villages 100 100 Source: GSCF-1 and GSCF-2 Principal, Teacher, and Village Questionnaires. Table 2 shows our sample of primary schools, junior high schools, primary and junior high school teachers, and villages. Our 2000 sample included 131 primary school principals and 1,009 primary school teachers. Our 2004 sample included 140 primary school principals, 1,018 primary school teachers, 77 11 junior high school principals, and 1,332 junior high school teachers. In both 2000 and 2004, our sample included 100 village leaders. THE NATURE OF RURAL SCHOOLS In this section, we describe attributes of schools, including their basic facilities, the background, working conditions, and work-related attitudes of their teachers, finance, and classroom learning environments. Basic Facilities Primary Schools Table 3 shows the percent of schools with various types of infrastructure, by year and by level. Table 4 shows means for various types of school resources by year and by level. In terms of general infrastructure, primary schools had an average of 10 classrooms in 2004 (Table 4). Just over half (58 percent in 2000 and 54 percent in 2004) had common teacher workrooms (Table 3). In 2004, just 2 percent of schools lacked electricity (Table 3). About 30 percent of classrooms failed to meet safety standards in 2004 (Table 4). 12 Table 3. School Facilities by School Level in 2000 and 2004 Primary School (t-values) Junior High School 2000 2004 2004 Are there enough desks and chairs for all students? n=131 n=140 1.51 n=77 Percent Yes 84.73 90.71 90.91 Does this school have electricity? n=131 n=140 0.47 n=77 Percent Yes 96.95 97.86 100.00 Does this school have a library? n=131 n=140 1.02 n=77 Percent Yes 70.99 76.43 97.40 Does this school have a science laboratory? n=131 n=140 1.32 n=77 Percent Yes 25.19 18.57 84.42 Does this school have a teacher workroom? n=131 Percent Yes 58.02 Does the school have a teacher workroom or jiaoyan office+ n=140 n=77 Percent Yes 54.29 72.73 Does this school have computers? n=140 N/A n=77 Percent Yes 37.14 96.10 Does this school provide teacher housing? n-140 n=77 Percent Yes 90.00 N/A 93.51 Does this school have boarding students? n=131 n=140 1.20 n=77 Percent Yes 3.82 7.14 46.75 Source: GSCF-1 and GSCF-2 Principal Questionnaires. + Response category changed in 2004 to include Jiaoyan Office H0: µPRIMARY2000=µPRIMARY2004 *=p<.05 13 Table 4. School Resources by School Level in 2000 and 2004 (Means) Primary School (t-values) Junior High School 2000 2004 2004 Number of teachers who live at the school n=129 n=140 2.16* n=77 5.60 7.66 36.03 Number of classrooms in the school n=140 n=77 10.04 26.55 Percent of unsafe classrooms in the school n=140 n=77 28.97 16.26 Percent of classrooms that can be used in the rain n=140 n=77 78.55 82.37 Percent of classrooms that have a glass chalkboard n=140 n=77 10.57 37.45 Percent of classrooms that have a cement chalkboard n=140 n=77 85.64 50.17 Percent of classrooms that have a magnetic chalkboard n=140 n=77 6.10 11.84 Source: GSCF-1 and GSCF-2 Principal Questionnaires. H0: µPRIMARY2000=µPRIMARY2004 *=p<.05 In 2004, about 90 percent of primary schools provided teacher housing, with an average of about 8 teachers living at the school. Consistent with what might be expected given recent efforts to consolidate primary schools in sparsely populated areas, under 4 percent of primary schools reported boarding students in 2000, but 7 percent did so in 2004 (Table 3). We next considered the kinds of infrastructure that might more directly support teaching and learning. In 2000, about 15 percent of principals reported insufficient desks and chairs for students, and by 2004, just under 10 percent did so (Table 3). However, this change might be related to declining numbers of 14 school-aged children, rather than to upgrades in schools. As described earlier, China's primary-aged cohort is decreasing, and the child population in Gansu is also decreasing (Figures 1 and 2). About 85 percent of classrooms in 2004 had cement chalkboards (Table 4). In 2004, roughly 11 percent of classrooms had higher-quality glass chalkboards and roughly 6 percent of classrooms had higher-quality magnetic chalkboards. Over one-third of schools reported having computers (2004 data only). Substantial numbers of schools lacked libraries, but this percentage is going down, in our sample areas: 29 percent in 2000 and 24 percent in 2004. Few schools (25 percent in 2000 and 19 percent in 2004) had science laboratories (Table 3). Junior High Schools For the 77 junior high schools serving sample villages in 2004, about 9 percent reported insufficient desks and chairs for students (Table 3). All junior high schools had electricity. Virtually all had a library (97 percent) and had computers (96 percent). Most schools had a science lab (84 percent) and a common teacher workroom (73 percent). Among junior high schools, 94 percent boarded teachers, and 47 percent boarded students. 15 Financial Arrangements Household Expenditures on Education Table 5 shows average semester educational expenditures on the target child in 2000 and in 2004. All costs enumerated in 2000 totaled 157 RMB; all costs enumerated in 2004 totaled 374 RMB. Thus, as target children grew older and started junior high school, parents paid more. This interpretation of the rise in costs is consistent with the costs disaggregated by school level (in the right three columns of Table 5), which highlight the substantially higher level of junior high school fees. 16 Table 5. Average Semester Educational Expenditures for the Target Child in 2000 and 2004 All Children3 Enrolled Primary Students Enrolled Junior High Students 2000 2004 2000 2004 2004 Tuition and Textbooks n=2000 n=1918 n=1905 n=451 n=1127 94.08 208.86 93.93 115.38 229.55 School Supplies+ n=2000 n=1918 n=1905 n=451 n=1127 20.37 46.26 20.47 29.02 55.84 Food, Board, and Transport++ n=2000 n=1905 10.76 9.67 Board and Transport+++ n=1918 n=451 n=1127 21.33 2.46 19.26 Food+++ n=1918 n=451 n=1127 66.37 9.71 68.54 Tutoring n=2000 n=1918 n=1905 n=451 n=1127 0.34 n=9.49 0.28 0.88 12.70 Uniforms n=2000 n=1918 n=1905 n=451 n=1127 18.67 6.55 18.23 4.13 8.95 Other Costs n=2000 n=1918 n=1905 n=451 n=1127 13.24 15.17 12.99 6.51 18.76 All Educational Costs1 n=2000 n=1918 n=1905 n=451 n=1127 157.46 374.03 155.57 168.08 413.60 Comparable Educational Costs2 n=2000 n=1918 n=1905 n=451 n=1127 126.33 240.07 125.43 126.89 269.96 Source: GSCF-1 and GSCF-2 Household Questionnaires. +The wording in 2004 included a dictionary and fudao materials, which were not included in 2000. ++Answer option wording in 2000. +++Answer option wording in 2004 1Total educational costs are computed as a sum of all enumerated cost items in each year (tuition and textbooks, school supplies, food, room, and board, tutoring, uniforms, and other costs). 2Comparable educational costs is calculated as a sum of cost items enumerated in both years (tuition and textbooks, tutoring, uniforms, and other costs). 3All children reported in household regardless of enrollment status. Table 6. Total Educational Expenditures for Households in 2000 and 2004 2000 2004 Total household educational expenditure1 2 n=2000 n=1918 356.77 910.17 Source: GSCF-1 and GSCF-2 Household Questionnaires. 1 Total educational costs are computed as a sum of all enumerated cost items in each year (tuition and textbooks, school supplies, food, room, and board, tutoring, uniforms, and other 2 All children reported in household regardless of enrollment status. 17 Table 6 shows more strikingly the rising burden on families as children age, given that most families have more than one child. On average, total enumerated semester costs in 2000 for sample child households were 356.77 RMB. In 2004, the total enumerated semester costs were 910 RMB. Do these per semester school costs represent a heavy burden on families? It seems likely that the answer is a resounding yes. According to available official estimates for the full year of 2004, total rural per capita net income in Gansu averaged 1,852 RMB (National Bureau of Statistics (China) 2005). In the GSCF sample in 2004 (N=1916), the corresponding figure was 1,926 RMB (calculations not shown). Tuition and Fees Reported by Schools We next turn to the question of the fees that school principals report that they are charging. In general, we present averages calculated in two ways: across all schools, and across schools that reported each fee type. Readers may be interested in viewing the all-school average, to see what the overall contribution of a given fee is in the fee structure of all schools, or the valid- observation average, to see what the fee average is just among those schools that charge the fee. Fees reported are per semester. We begin by discussing mandatory fees. Between 2000 and 2004, many schools began to adopt a new "one fee system" (yifeizhi, ), in which schools charged just one total fee. However, among schools that reported having 18 adopted the one fee system, reporting on fees varied. Among principals that declared that they were using the one fee system, some reported only a breakdown of fees, some reported only the one fee system amount, some reported both (with the breakdown summing to the one fee amount), and still others reported a one fee system, plus some additional fees by type. For this reason, we do not attempt to aggregate the fees, but rather present averages for each fee type by grade, as reported by principals. Table 7 shows the average across all schools and Table 8 shows the same set of numbers averaged across non-zero, non-missing cases. Table 7. Average Semester Required School Fees by Grade in 2000 and 2004 (Average Across All Schools) Required Fees Primary 1 Primary 2 Primary 3 Primary 4 Primary 5 Primary 6 Junior 1 Junior 2 Junior 3 2000 2004 2000 2004 2000 2004 2000 2004 2000 2004 2000 2004 2004 2004 2004 Tuition n=131 n=140 n=131 n=140 n=131 n=140 n=131 n=140 n=131 n=140 n=131 n=140 n=77 n=77 n=77 13.94 24.79 13.70 26.41 14.62 27.00 12.59 27.11 12.42 26.49 3.11 22.11 32.39 32.55 32.58 Book fee n=131 n=140 n=131 n=140 n=131 n=140 n=131 n=140 n=131 n=140 n=131 n=140 n=77 n=77 n=77 53.44 19.68 55.28 20.03 70.93 23.29 64.94 24.36 77.29 23.42 15.06 17.91 66.31 66.43 54.57 Heating fee n=131 n=140 n=131 n=140 n=131 n=140 n=131 n=140 n=131 n=140 n=131 n=140 n=77 n=77 n=77 7.64 4.29 7.65 4.29 7.63 4.08 7.54 4.29 6.87 3.86 2.08 3.50 8.44 8.44 8.44 Class fee n=131 n=140 n=131 n=140 n=131 n=140 n=131 n=140 n=131 n=140 n=131 n=140 n=77 n=77 n=77 1.53 0.61 1.48 0.61 0.89 0.61 1.39 0.61 1.80 0.66 0.34 0.64 2.43 2.43 2.43 Other fee n=131 n=140 n=131 n=140 n=131 n=140 n=131 n=140 n=131 n=140 n=131 n=140 n=77 n=77 n=77 2.09 0.76 2.24 0.77 2.63 0.71 2.67 1.15 2.24 1.28 1.19 1.17 10.21 10.31 10.47 One fee system+ n=140 n=140 n=140 n=140 n=140 n=140 n=77 n=77 n=77 38.95 38.84 38.76 38.36 37.81 13.54 61.65 61.17 62.77 Source: GSCF-1 and GSCF-2 Principal Questionnaires. + Response category in 2004 Table 8. Average Semester Required School Fees by Grade in 2000 and 2004 (Average Across Schools with Non-Missing and Non-Zero Values)+ Required Fees Primary 1 Primary 2 Primary 3 Primary 4 Primary 5 Primary 6 Junior 1 Junior 2 Junior 3 2000 2004 2000 2004 2000 2004 2000 2004 2000 2004 2000 2004 2004 2004 2004 Tuition n=127 n=71 n=127 n=70 n=127 n=70 n=122 n=68 n=108 n=63 n=41 n=45 n=46 n=46 n=46 14.38 48.87 14.13 52.81 15.08 54.01 13.52 55.82 15.06 58.87 9.93 68.80 54.22 54.48 54.54 Book fee n=126 n=70 n=126 n=70 n=126 n=68 n=122 n=67 n=108 n=62 n=41 n=45 n=46 n=46 n=46 55.56 39.36 57.47 40.06 73.75 47.94 69.73 50.91 93.75 52.89 48.10 55.73 111.00 111.20 91.35 Heating fee n=80 n=27 n=80 n=27 n=79 n=26 n=78 n=27 n=71 n=24 n=19 n=20 n=20 n=20 n=20 12.51 22.26 12.53 22.26 12.65 21.96 12.66 22.26 12.68 22.54 14.34 24.50 32.50 32.50 32.50 Class fee n=67 n=23 n=67 n=23 n=67 n=23 n=66 n=22 n=60 n=22 n=26 n=21 n=31 n=31 n=31 2.99 3.70 2.89 3.70 1.74 3.70 2.77 3.86 3.93 4.18 1.71 4.24 6.03 6.03 6.03 Other fee n=22 n=11 n=23 n=11 n=23 n=11 n=23 n=11 n=20 n=12 n=10 n=10 n=18 n=18 n=19 12.44 9.73 12.74 9.82 14.97 9.00 15.21 14.64 14.65 14.92 15.64 16.40 43.67 44.11 42.42 One fee system++ n=73 n=73 n=72 n=70 n=68 n=23 n=35 n=35 n=35 74.70 74.49 75.36 76.73 77.85 82.39 135.63 134.57 138.09 Source: GSCF-1 and GSCF-2 Principal Questionnaires. + Some schools reported having a one fee system and reported a single fee, other schools reported having a one fee system and reported one fee and component fees, still other schools reported having a one fee system and did not report a one fee, but did report component fees. All of these configurations are included in this table. ++Response category in 2004 19 The structure of fees and fee changes between 2000 and 2004 are presented so that readers may peruse the individual fee items based on interest. Given the differences in adoption of the one fee system, reporting by principals, and patterns in the data, it is difficult to characterize the changing fee structure in a straightforward way. Among primary schools in the sample villages that reported charging tuition, the fee was substantially higher in 2004 than in 2000-- ranging from about 49 to 69 RMB for different grades in the latter year, compared to 10 to 15 RMB in the earlier year. Book fees were another costly item among required fees, especially at the junior high school grades (see Table 8). Next, we consider the one fee system in 2004. If we consider Table 8, which calculates averages only across non-zero, non-missing observations, it appears that at the primary level, the one fee system schools were charging between 74 and 82 RMB per semester in 2004. The semester fee jumped to 135 to 138 RMB for junior high schools. Again, it is likely that, at least in some schools, these fees do not represent the full fees paid by students, in light of the reporting of principals. "Voluntary" fees can also be charged by schools, for items ranging from insurance to textbooks to meal and boarding fees. From reporting by principals, the degree to which these voluntary fees are included in the one fee system is unclear. Table 9 shows voluntary school fees by type and year, averaged across all schools, and Table 10 shows the same fees averaged across 20 schools with non-zero, non-missing values. We focus here on Table 10, given the generally small numbers of schools reporting each voluntary fee type. Table 9. Average Semester Voluntary School Fees by Grade in 2000 and 2004 ( Average Across All Schools) Voluntary Fees Primary 1 Primary 2 Primary 3 Primary 4 Primary 5 Primary 6 Junior 1 Junior 2 Junior 3 2000 2004 2000 2004 2000 2004 2000 2004 2000 2004 2000 2004 Safety insurance fee n=131 n=131 n=131 n=131 n=131 n=131 12.87 12.47 13.80 11.74 12.20 2.97 Tutoring fee n=131 n=131 n=131 n=131 n=131 n=131 0.00 0.00 0.00 0.08 0.08 0.00 Boarding fee n=131 n=131 n=131 n=131 n=131 n=131 0.46 0.46 0.61 0.69 0.69 0.53 Meal fee n=131 n=131 n=131 n=131 n=131 n=131 2.21 2.21 2.21 2.21 2.21 2.21 Library fee n=131 n=131 n=131 n=131 n=131 n=131 0.00 0.00 0.00 0.00 0.00 0.00 Minban fee n=131 n=131 n=131 n=131 n=131 n=131 0.29 0.29 0.29 0.29 0.29 0.23 Over enrollment fee n=131 n=131 n=131 n=131 n=131 n=131 0.00 0.00 0.00 0.00 0.00 0.00 Textbook fee n=131 n=131 n=131 n=131 n=131 n=131 20.81 22.63 25.25 27.01 24.07 10.07 Exercise books fee n=131 n=131 n=131 n=131 n=131 n=131 3.64 3.81 4.94 5.10 4.38 2.43 Public security fee n=131 n=131 n=131 n=131 n=131 n=131 0.05 0.05 0.05 0.05 0.05 0.00 Fund for teacher retirement n=131 n=131 n=131 n=131 n=131 n=131 0.00 0.00 0.00 0.00 0.00 0.00 Winter vacation homework fee n=131 n=131 n=131 n=131 n=131 n=131 2.15 2.31 2.38 2.33 2.11 0.59 Graduation certification fee n=131 n=131 n=131 n=131 n=131 n=131 0.10 0.11 0.11 0.11 1.33 0.35 Other fee n=131 n=131 n=131 n=131 n=131 n=131 0.14 0.14 0.14 0.14 0.17 0.10 Health insurance+ n=140 n=140 n=140 n=140 n=140 n=140 n=77 n=77 n=77 2.69 2.69 2.68 2.71 2.36 2.01 5.77 5.25 6.42 Accident insurance+ n=140 n=140 n=140 n=140 n=140 n=140 n=77 n=77 n=77 7.75 7.34 7.31 7.40 6.60 5.47 7.30 7.10 6.99 Other insurance+ n=140 n=140 n=140 n=140 n=140 n=140 n=77 n=77 n=77 0.57 0.36 0.36 0.43 0.46 0.66 1.84 2.08 0.00 Tutoring fee+ n=140 n=140 n=140 n=140 n=140 n=140 n=77 n=77 n=77 0.00 0.00 0.00 0.14 0.14 0.14 2.08 2.08 7.10 Boarding fee+ n=140 n=140 n=140 n=140 n=140 n=140 n=77 n=77 n=77 2.18 2.18 2.18 2.18 2.46 2.54 10.52 10.58 11.62 Assistance fee+ n=140 n=140 n=140 n=140 n=140 n=140 n=77 n=77 n=77 0.25 0.25 0.25 0.25 0.25 0.25 0.00 0.00 0.00 Other+ n=140 n=140 n=140 n=140 n=140 n=140 n=77 n=77 n=77 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Source: GSCF-1 and GSCF-2 Principal Questionnaires. + Response category wording in 2004 21 Table 10. Average Semester Voluntary School Fees by Grade in 2000 and 2004 (Average Across Schools with Non-Missing and Non-Zero Values) Voluntary Fees Primary 1 Primary 2 Primary 3 Primary 4 Primary 5 Primary 6 Junior 1 Junior 2 Junior 3 2000 2004 2000 2004 2000 2004 2000 2004 2000 2004 2000 2004 2004 2004 2004 Safety insurance fee n=64 n=64 n=64 n=64 n=59 n=25 26.34 25.52 28.24 24.02 27.09 15.59 Tutoring fee n=0 n=0 n=0 n=1 n=1 n=0 0.00 0.00 0.00 10.00 10.00 0.00 Boarding fee n=2 n=2 n=3 n=4 n=4 n=3 30.00 30.00 26.67 22.50 22.50 23.33 Meal fee n=2 n=2 n=2 n=2 n=2 n=2 145.00 145.00 145.00 145.00 145.00 145.00 Library fee n=0 n=0 n=0 n=0 n=0 n=0 0.00 0.00 0.00 0.00 0.00 0.00 Minban fee n=3 n=3 n=3 n=3 n=3 n=1 12.67 12.67 12.67 12.67 12.67 30.00 Over enrollment fee n=0 n=0 n=0 n=0 n=0 n=0 0.00 0.00 0.00 0.00 0.00 0.00 Textbook fee n=78 n=78 n=78 n=76 n=66 n=29 34.96 38.01 42.41 46.56 47.77 45.51 Exercise books fee n=58 n=58 n=59 n=59 n=50 n=23 8.22 8.61 10.96 11.32 11.46 13.86 Public security fee n=1 n=1 n=1 n=1 n=1 n=0 6.00 6.00 6.00 6.00 6.00 0.00 Fund for teacher retirement n=0 n=0 n=0 n=0 n=0 n=0 0.00 0.00 0.00 0.00 0.00 0.00 Winter vacation homework fee n=75 n=76 n=76 n=73 n=63 n=23 3.75 3.98 4.10 4.17 4.39 3.34 Graduation certification fee n=5 n=5 n=5 n=5 n=34 n=24 2.56 2.76 2.76 2.76 5.12 1.93 Other fee n=4 n=4 n=4 n=4 n=5 n=2 4.63 4.63 4.63 4.63 4.50 6.75 Health insurance+ n=19 n=19 n=19 n=19 n=17 n=12 n=8 n=8 n=8 19.79 19.84 19.74 19.95 19.41 23.50 55.50 50.50 61.75 Accident insurance+ n=54 n=53 n=53 n=53 n=48 n=33 n=28 n=28 n=28 20.09 19.40 19.32 19.55 19.25 23.21 20.07 19.54 19.21 Other insurance+ n=2 n=2 n=2 n=2 n=2 n=3 n=1 n=1 n=0 40.00 25.00 25.00 30.00 32.00 30.67 142.00 160.00 0.00 Tutoring fee+ n=0 n=0 n=0 n=1 n=1 n=1 n=4 n=4 n=7 0.00 0.00 0.00 20.00 20.00 20.00 40.00 40.00 78.14 Boarding fee+ n=3 n=3 n=3 n=3 n=4 n=5 n=17 n=17 n=18 101.67 101.67 101.67 101.67 86.25 71.00 47.65 47.94 49.72 Assistance fee+ n=1 n=1 n=1 n=1 n=1 n=1 n=0 n=0 n=0 35.00 35.00 35.00 35.00 35.00 35.00 0.00 0.00 0.00 Other+ n=0 n=0 n=0 n=0 n=0 n=0 n=0 n=0 n=0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Source: GSCF-1 and GSCF-2 Principal Questionnaires. + Response category wording in 2004 Common types of voluntary school fees in both primary and junior high schools include insurance and, in 2000, textbook fees. At the primary school level in 2004, accident insurance ranged from 19 RMB to 23 RMB (Table 10). Accident insurance in junior high school was roughly 20 RMB. In primary schools, the most expensive voluntary school fee observed was the boarding fee, but primary school boarding fees were exceedingly rare--3 to 5 schools in 2004 and 2 to 4 schools in 2000, depending on the grade considered. Overall, the low percentage of schools citing these fees is likely related to these costs being subsumed, for many schools, under the one fee system. 22 An additional school cost is uniforms. A little over one-third of primary schools reported using uniforms in 2000, with the figure rising to about 46 percent in 2004 (see Table 11). About 58 percent of junior high schools in 2004 reported using uniforms. Among schools reporting use of uniforms, the average cost was about 45 RMB at the primary level in both years, and about 51 RMB at the junior high school level in 2004. Table 11. Percent of Schools with Partial Fees and Uniforms by School Grade and Year Primary School Junior High School 2000 2004 2004 Percent of schools with partial fee n=131 n=140 n=77 52.67 78.57 89.61 Average percent of students who pay a partial fee in schools with partial fees n=68 n=108 n=69 8.16 6.60 3.41 Percent of schools with uniforms n=131 n=140 n=77 35.88 45.71 58.44 Average cost of uniform n=46 n=63 n=45 45.30 45.33 51.04 Source: GSCF-1 and GSCF-2 Principal Questionnaires. A final complicating factor in the school fees story is that many schools allow some percent of students to pay just a partial fee (see Table 11). In 2000, just over half of primary schools reported having students paying only partial fees, while in 2004, the number rose to 79 percent. Among primary schools reporting partial fees, about 8 percent of students paid partial fees in 2000, and about 7 percent did so in 2004. Among junior high schools reporting partial fees, the average percent of students paying partial fees was 3 percent. 23 In 2004, the one semester costs reported by parents were much more than the semester one fee reported by school principals, for primary and junior high schools (see Tables 5 and 8). This disparity is due in part to the fact that some of the costs incurred by parents are not paid to the schools. Public, Private, and Hybrid Schools Tsang (2000) notes that there is a wide variety of arrangements of management and finance, such that many schools are not easy to categorize as fully private or fully public. This makes research on the private sources of support for education complex. Table 12 shows the distribution of primary and junior high schools by financial arrangement and, for primary schools, by year. In 2000, the questionnaire asked whether a school was "public" or "other", and among the 131 primary schools in the sample, about 91 percent reported being public schools. Thus, while the vast majority of primary schools were public, almost one in ten schools had some other arrangement of finances. 24 Table 12. Distribution of Schools by Financial Arrangement by School Level in 2000 and 2004 (Percent) Primary School Junior High School Schools in 2000 n=131 Public 90.84 Private 0.00 Other 9.16 Schools in 2004 n=140 n=77 Public 75.00 98.70 Private 2.86 1.30 Gongban Minzhu 20.71 0.00 Other 1.43 0.00 Source: GSCF-1 and GSCF-2 Principal Questionnaires. In 2004, the categories for school type were laid out in a more detailed way, and thus we have a little more information about financial arrangements. Categories in the 2004 questionnaire included public, private, a hybrid category called gongban minzhu (, literally "publicly run, people supported") and other. Gongban minzhu schools are considered public schools, with faculty and staff provided by the government, but with private funding for school buildings. With this set of options available to principals, just three-fourths of school principals reported their school as a public school, while 21 percent reported gongban minzhu, 2.86 percent reported private, and 1.43 percent reported other types of arrangements. Thus, while public schools remain the majority of primary schools in rural Gansu, almost one in ten schools had some arrangement other than being a public school by 2000, and a hybrid arrangement of public- private funds characterized about one in five schools by 2004. 25 In our sample of 77 junior high schools serving sample villages, almost all (98.70 percent) remained public as of 2004. It is important to note that these public schools, and also primary schools officially classified as public, also receive some form of private funds, via fees collected from students. Personnel Costs Reported by Schools Official teachers are called gongban teachers (, literally "publicly-run"), which means that they are credentialed and are employed by the government. The average salary for an official teacher reported by principals of primary schools was 577 RMB per month in 2000 and 935 RMB per month in 2004. Average bonuses reported in 2000 were 83 RMB. In 2004, average bonuses for those who received them were 232 RMB. The smaller number of schools that hired substitute daike teachers () paid them much less, on average: just 154 RMB in 2000 and 188 RMB in 2004 (Table 13). Daike teachers are not credentialed teachers, may or may not have the appropriate education level, and are employed by the local community. In junior high schools, data from principals for 2004 show that the average monthly salary paid to gongban teachers was 953 RMB and the average bonus for those receiving bonuses was 292 RMB. For the smaller number of junior high schools reporting salaries for daike teachers, the average was 337 RMB. 26 Table 13. Principal Reports of Teacher Salary by School Level in 2000 and 2004 Primary School (t-values) Junior High School 2000 2004 2004 Average monthly salary of gongban teacher n=129 n=135 16.96* n=77 577.19 935.49 952.58 Average monthly salary of daike teacher n=72 n=84 1.53 n=32 153.75 188.30 337.03 Average annual teacher bonus n=128 82.89 Average annual bonus for those who received them+ n=75 n=54 231.64++ 292.30++ Source: GSCF-1 and GSCF-2 Principal Questionnaires. + The phrasing of the question changed in 2004 ++ This average is for principals who reported non-zero values. Including principals who reported zero values, the average 2004 primary school bonus was 124.09 (n=140) and the average 2004 Junior high school bonus was 204.99 (n=77). H0: µPRIMARY2000=µPRIMARY2004 *=p<.05 Salary averages can also be calculated from reports of teachers about their actual salaries, rather than from principal reports of school averages. Table 14 shows teacher reports of remuneration and arrears in 2000 and 2004. Primary school teachers reported an average total salary of 527 RMB per month in 2000. In 2004, the average total salary was significantly higher, at 875 RMB. Also in 2004, junior high school teachers reported an average total salary of 956 RMB. 27 Table 14. Teacher Reports of Remuneration and Arrears by School Level in 2000 and 2004 Primary School (t-values) Junior High School 2000 2004 2004 Average monthly total teacher salary n=1002 n=1017 22.74*1 n=1327 526.91 875.29 955.57 School pays salary on time+ n=1005 n=1018 N/A n=1332 Always 7.86 39.49 32.21 Usually 4.68 17.98 23.57 Sometimes 55.82 35.95 34.16 Never 31.64 6.58 10.06 Does the school owes you any wages right now?++ n=923 n=1018 n=1332 Percent Yes 90.57 42.04 46.70 Source: GSCF-1 and GSCF-2 Teacher Questionnaires. +The question wording for 2000 and 2004 were different. In 2000, Does your school pay your salary on time? In 2004, From September 2003-June 2004 did you receive your salary on time? ++The phrasing of the question changed slightly, but the English translation remains the same. 1H0: µPRIMARY2000=µPRIMARY2004 *=p<.05 One important problem in recent years has been delayed payment of teacher salaries. In 2000, just 13 percent of primary school teachers reported that their salary was usually or always paid on time. About 91 percent of primary school teachers reported that they were owed wages at the time of the survey. In 2004, about 57 percent of primary school teachers reported that their salary was usually or always paid on time. About 42 percent of primary teachers were owed wages at the time of the survey. These findings suggest that meeting personnel costs remains an important problem at rural schools, but seems to be less of a problem than in the past. This improvement might be related to efforts to recentralize finance to the county level (Tsang and Ding, 2005: 29 [note xxix]). 28 Teachers Primary Schools - Background Table 15 shows average teacher age, years of experience, and years a teacher has taught in the same school in 2000 and 2004. The average age of primary school teachers in 2000 was 36 years old. Teachers had an average of 15 years of teaching experience, and had taught in the current place of employment for an average of 7 years. The situation in 2004 was similar. The average age of a primary school teacher was 37 years old, and he or she had an average of 16 years of experience, with 8 years of experience in the current school. Table 15. Average Teacher Age and Experience by School Level in 2000 and 2004 Primary School (t-values) Junior High School 2000 2004 2004 Average teacher age (in years) n=1006 n=1018 2.15* n=1332 35.5 36.57 32.13 Average number of years teaching+ n=1009 n=1018 n=1332 14.77 16.03 10.72 Average number of years in current school n=1009 n=1018 2.19* n=1332 7.24 8.00 6.59 Source: GSCF-1 and GSCF-2 Teacher Questionnaires. + The phrasing of the question changed slightly, but the English translation remains the same. H0: µPRIMARY2000=µPRIMARY2004 *=p<.05 29 Table 16 shows additional teacher background characteristics for 2000 and 2004, including the following: local versus non-local status, gender, marital status, credentials, and participation in outside work. This table shows that in our sample, more than 90 percent of teachers in both years are from the same county. 30 Table 16. Teacher Characteristics by School Level in 2000 and 2004 Primary School chi2/( t-values) Junior High School 2000 2004 2004 Teacher birthplace n=1009 n=1018 22.56 (4)* n=1332 Same village 35.08 28.78 20.35 Same township, different village 33.20 29.86 25.38 Same county, different township 26.56 35.76 45.42 Same province, different county 4.16 4.81 7.06 Other province 0.99 0.79 1.80 Gender of teacher n=1004 n=1018 21.46 (1)* n=1332 Male 62.45 52.26 71.85 Female 37.55 47.74 28.15 Ethnicity n=1009 n=1018 N/A n=1332 Han 97.22 97.94 98.72 Mongol 0.10 1.08 0.53 Hui 1.98 0.39 0.30 Tibetan 0.30 0.20 0.08 Tujia 0.30 0.39 0.23 Dong Xiang+ 0.10 Manchu++ 0.00 0.08 Tu++ 0.00 0.08 Marital status n=1009 n=1018 4.13 (3) n=1332 Never Married 17.24 16.70 23.20 Currently Married 81.96 81.63 76.43 Divorced 0.40 0.49 0.30 Widowed 0.40 1.18 0.08 Teacher is a credentialed teacher n=1009 N/A Percent Yes 80.28 Teacher is: n=1018 N/A n=1332 A Gongban Teacher 83.20 92.64 A Minban Teacher 3.93 0.75 A Daike Teacher 11.59 5.63 Other 1.28 0.98 Teacher is a farmer n=1007 n=1018 4.13*1 n=1332 Percent Yes 48.86 39.78 35.51 Teacher has a part-time job during winter and summer holiday n=1009 n=1018 7.90*1 n=1332 Percent Yes 19.92 7.96 7.28 Source: GSCF-1 and GSCF-2 Teacher Questionnaires. + Answer Option only in 2000 ++ Answer Options only in 2004 1 H0: µPRIMARY2000=µPRIMARY2004 *=p<.05 31 Qualifications and Status There are three main teacher categories: official (gongban, ), unofficial (minban, literally "people run"), and substitute (daike, ). Minban teachers are not credentialed, may or may not have the sufficient educational attainment for the job, and are employed at the local community level. Daike teachers are considered temporary substitute teachers. Like minban teachers, daike teachers are not credentialed and may not have an appropriate educational level for the position they hold, and they are employed at the local community level. In 2000, teachers responded to a single question about whether they were a credentialed teacher: "Are you a credentialed teacher?" Credentialed teachers are gongban teachers employed by the government. Most primary teachers in 2000 were gongban teachers: 80 percent of primary teachers reported being a credentialed teacher. The 2004 questionnaire included more detailed teacher response categories including gongban, minban, daike, and other. In 2004, 83 percent of primary teachers reported being a gongban teacher. Roughly 4 percent of primary school teachers reported being a minban teacher in 2004. About 12 percent of 32 primary school teachers were daike teachers in 2004. About 1 percent of primary school teachers in 2004 reported being in the "other" category. Table 17 shows teacher educational levels for primary and junior high schools in 2004 within their teaching category (gongban, minban, daike, and other). The wording of questions about teacher educational levels was more detailed in 2004 (primary school, junior high school, senior high school, vocational middle school, vocational college, four-year college, graduate school, and other) than in 2000. In 2000, the focus was on the teacher graduation year and type of school, limited to middle school, senior high school, and vocational and four-year colleges. Due to the different question wording in the two years, we present teacher educational levels for year 2004 only. 33 Table 17. Teacher Education Level by Teacher Category and School Type in 2004 (Percentages) Primary School Junior High School Teacher Category Gongban Teacher n=847 n=1234 Primary School 0.00 0.00 Junior High School 2.48 0.16 Senior High School 15.47 1.30 Vocational Middle School 39.91 7.37 Vocational College 40.14 73.74 Four-Year Degree 2.01 17.34 Graduate School 0.00 0.08 Other 0.00 0.00 Minban Teacher n=40 n=10 Primary School 0.00 0.00 Junior High School 10.00 0.00 Senior High School 35.00 30.00 Vocational Middle School 35.00 20.00 Vocational College 20.00 50.00 Four-Year Degree 0.00 0.00 Graduate School 0.00 0.00 Other 0.00 0.00 Daike Teacher n=118 n=75 Primary School 0.00 0.00 Junior High School 8.47 0.00 Senior High School 29.66 4.00 Vocational Middle School 41.53 12.00 Vocational College 19.49 78.67 Four-Year Degree 0.85 5.33 Graduate School 0.00 0.00 Other 0.00 0.00 Other Teacher n=13 n=13 Primary School 0.00 0.00 Junior High School 7.69 0.00 Senior High School 15.38 0.00 Vocational Middle School 46.15 15.38 Vocational College 30.77 84.62 Four-Year Degree 0.00 0.00 Graduate School 0.00 0.00 Other 0.00 0.00 Source: GSCF-2 Teacher Questionnaires. 34 In 2004, as many as 40 percent of gongban primary teachers had attended a vocational college, 40 percent had attended a vocational middle school, and two percent had a four-year tertiary degree. Just 2.5 percent had only a junior high school degree. There were few minban teachers: just 40. Of these few, ten percent had a junior high school degree, 70 percent had a secondary level degree (half of these vocational), and 20 percent had a vocational college degree. Similarly, 8.47 percent of daike primary teachers had only a junior high school degree, while about 30 percent had a senior high school degree and 41.5 percent had a vocational middle school degree. About 20 percent had any kind of tertiary degree. The sample size of "other" teachers was too small to make an interpretation meaningful. Table 18. Teacher Rank by School Level in 2000 and 2004 (Percentages) Primary School Junior High School 2000 2004 2004 n=895 n=999 n=1327 Intern 10.39 13.81 10.40 Primary, Level Two 20.67 17.62 3.01 Primary, Level One 48.60 40.44 7.99 Primary, Level High 16.76 24.02 2.56 Middle School, Level Three+ 0.60 15.90 Middle School, Level Two 2.68 2.70 41.82 Middle School, Level One 0.89 0.80 17.56 Middle School, Level High 0.00 0.00 0.75 Source: GSCF-1 and GSCF-2 Teacher Questionnaires. +Answer Option Available in 2004 only 35 Teachers in China are ranked each year by the principal and school district. In primary school, the levels include 3, 2, 1, and high, where level 3 is the lowest ranking, corresponding to an internship status, and high is the highest ranking. Table 18 shows teacher ranks by year. More teachers held a primary, level high rank in 2004 than in 2000. In 2000, most primary school teachers reported that they held the rank of primary, level one (49 percent) with fewer holding a primary, level high (17 percent). Roughly 40 percent of teachers held a rank of primary, level one in 2004 and about 24 percent held the rank of primary, level high. Workloads and Work Activities Substantial numbers of teachers work outside of teaching, though the percentage doing so may be dropping. In 2000, 49 percent of primary school teachers reported also working as farmers, and in 2004, 40 percent of primary school teachers reported working as farmers (Table 16). In 2004, fewer primary school teachers held a part-time job during the winter or summer holiday (8 percent) than in 2000 (20 percent). Table 19 describes teacher workloads and work activities in 2000 and 2004. Consistent with findings about outside work in Table 16, Table 19 shows that teachers in 2004 spent less time at part-time outside work (2 hours per week) than teachers in 2000 (6 hours per week). 36 Table 19. Teacher Workload by School Level in 2000 and 2004 Primary School (t-values) Junior High School 2000 2004 2004 Average number of classes taught each week n=1018 n=1332 20.06 13.56 Average number of hours spent preparing for class each week n=1009 n=1018 n=1332 12.28 9.72 11.24* 9.92 Average number of hours spent grading papers each week n=1009 n=1018 n=1332 11.79 10.32 6.07* 10.70 Average number of hours spent lecturing each week n=1009 14.95 Average number of hours spent on career development each week n=1000 5.19 Average number of hours spent on political and professional study each week n=1018 n=1332 3.13 3.12 Average number of hours spent on part-time work each week, including farming, business, and other money making activities n=1009 n=1018 n=1332 5.53 2.06 11.26* 1.66 Source: GSCF-1 and GSCF-2 Teacher Questionnaires. H0: µPRIMARY2000=µPRIMARY2004 *=p<.05 Primary school teachers reported spending an average of 15 hours lecturing each week in 2000. In 2004, primary school teachers reported teaching 20 classes each week. In 2000, primary school teachers reported spending 12 hours a week preparing for classes and 12 hours grading papers; these numbers dropped to 10 hours for preparing for classes and 10 hours a week grading papers in 2004. In 2000, primary school teachers reported spending 5 hours a week on career development. In 2004, primary school teachers spent an average of 3 hours on political and professional development each week. 37 Professional Development Table 20 shows professional development activities conducted by teachers in 2004. The 2004 questionnaire included questions about professional development within the school, at another school, at district-level training sessions, at teacher training institutes, and at training sessions led by educational project experts. Each school has a jiaoyanzu (), which is a teaching and research section that conducts meetings and provides support for teachers. There are separate subject-specific jiaoyanzu, such as for math, language, and science. Each jiaoyanzu organizes peer observations, shares new materials, and disseminates information from the upper levels of administration to teachers. Primary school teachers in 2004 participated more often in professional development activities in their own school--such as through the jiaoyanzu--than in activities conducted at another school or conducted by someone outside their own school. For example, at the school level, 48 percent of primary school teachers participated in activities organized by their own school once a week, 38 percent participated in peer observation and evaluation of teaching in their school once a week, and 48 percent were involved in teaching and research activities in their school once a week. Roughly 32 percent of primary school teachers participated in teaching and research at another school one to two times a semester; 8 percent did so once a month, and 3 percent did so on a weekly basis. 38 Less than half of primary school teachers had ever received training conducted by an expert of an educational project. Table 20. Teacher Professional Development Activities by School Level in 2004 One to Two Once a Week Once a Month Times a Semester Once Never Peer observation and evaluation of teaching Primary School (n=1018) 38.11 27.21 26.42 3.63 4.62 Junior High School (n=1332) 35.81 29.20 25.00 6.53 3.45 Model lesson at this school or another school Primary School (n=1018) 10.90 20.63 48.04 10.41 10.02 Junior High School (n=1332) 11.64 23.95 38.81 13.59 12.01 Short term training activities at a teacher-training institute Primary School (n=1018) 1.96 2.46 27.31 25.74 42.53 Junior High School (n=1332) 2.78 3.83 29.43 26.65 37.31 Short term training conducted by an expert of an educational project Primary School (n=1018) 1.67 3.34 18.76 23.58 52.65 Junior High School (n=1332) 1.43 2.10 17.34 28.90 50.23 District level teaching and research activities (not including model lessons) Primary School (n=1018) 3.93 19.06 41.45 19.25 16.31 Junior High School (n=1332) 4.95 10.89 29.05 19.89 35.21 Teaching and research activities at another school Primary School (n=1018) 3.44 8.25 32.22 21.71 34.38 Junior High School (n=1332) 3.08 5.33 22.67 20.12 48.80 Teaching and research activities at your own school (including the teaching and research section) Primary School (n=1018) 47.54 25.25 18.07 4.13 5.01 Junior High School (n=1332) 53.83 24.77 13.21 5.18 3.00 Activities organized by your own school Primary School (n=1018) 47.64 22.89 16.01 4.22 9.23 Junior High School (n=1332) 34.46 31.08 20.65 4.88 8.93 Source: GSCF-2 Teacher Questionnaire. Satisfaction with Various Dimensions of Work Table 21 describes teacher satisfaction with various dimensions of work. In 2000, almost all primary school teachers liked teaching (96 percent) and were content with teaching (96 percent). Yet, almost one in five primary school teachers (18 percent) wanted to change their profession. 39 Table 21. Teacher Reports of Satisfaction by School Level in 2000 and 2004 (Percentages) Primary School Junior High School 2000 2004 2004 Do you like teaching? n=1009 Percent Yes 95.74 Are you content teaching? n=1009 Percent Yes 95.94 I am satisfied with teaching as job. n=1018 n=1332 Strongly Disagree 2.16 3.00 Disagree 10.31 16.97 Agree 58.64 61.71 Strongly Agree 28.88 18.32 Do you want to change your profession? n=1006 Percent Yes 17.59 I want to change my career. n=1018 n=1332 Strongly Disagree 15.32 12.09 Disagree 61.00 51.88 Agree 19.06 29.43 Strongly Agree 4.62 6.61 I am satisfied with my salary. n=1018 n=1332 Strongly Disagree 17.68 17.57 Disagree 34.18 42.94 Agree 41.45 36.71 Strongly Agree 6.68 2.78 Apart from my salary, my job benefits are good. n=1018 n=1332 Strongly Disagree 44.50 46.10 Disagree 44.30 41.82 Agree 9.63 10.06 Strongly Agree 1.57 2.03 Are you satisfied with the assessment of teacher titles? n=991 Percent Yes 74.57 The evaluation of teachers is fair and equitable n=1018 n=1332 Strongly Disagree 10.02 11.34 Disagree 24.95 30.33 Agree 60.61 55.03 Strongly Agree 4.42 3.30 Are you satisfied with the selection of outstanding teachers? n=989 Percent Yes 80.79 The evaluation of outstanding teachers is fair and equitable n=1018 n=1332 Strongly Disagree 5.50 9.98 Disagree 23.77 30.41 Agree 64.64 56.46 Strongly Agree 6.09 3.15 Source: GSCF-1 and GSCF-2 Teacher Questionnaires. 40 To seek out a broader scale of attitudes about teaching, we worded the questions differently in 2004, asking teachers to rate degree of agreement with various statements about teaching. This wording continued to reveal a high degree of reported satisfaction, but again a considerable degree of desire to change careers. Roughly 88 percent of primary school teachers agreed or strongly agreed with the statement, "I am satisfied with teaching as a job". Almost one in four primary school teachers (25 percent) agreed or strongly agreed with the statement, "I want to change my career". To understand why teachers might want to change careers, our 2004 questionnaire examined teacher satisfaction with salary, job benefits and evaluation of teachers. In 2004, less than half of teachers were satisfied with their salary (48 percent agreed or strongly agreed with the statement, "I am satisfied with my salary"). Interestingly, very few teachers were satisfied with job benefits. Few primary school teachers in 2004 agreed (10 percent) or strongly agreed (2 percent) with the statement, "Apart from my salary, my job benefits are good". Another dimension of job satisfaction is that associated with the evaluation of teachers. Here again, we used different wording in 2000 and 2004, and so draw insights from both years but cannot assess change by comparing the years. Most teachers, but not a vast majority, report satisfaction with this 41 dimension of work. In 2000, 75 percent of primary school teachers responded "yes" to a question about whether they were satisfied with teacher assessment, and 81 percent responded affirmatively to a question about whether they were satisfied with the selection of outstanding teachers. In 2004, the questions suggested more strongly that a substantial minority of teachers was not satisfied with this dimension of their jobs. For example, 65 percent of primary school teachers agreed or strongly agreed with the statement, "The evaluation of teachers is fair and equitable". Roughly 71 percent of primary school teachers in 2004 agreed or strongly agreed with the statement, "The evaluation of outstanding teachers is fair and equitable". Attitudes about School Management and School Culture Table 22 shows distributions of teacher responses--strongly disagree to strongly agree--to statements about principals' leadership. Table 23 shows responses to statements about principal school management strategies, and Table 24 shows responses to statements about relationships with principals. On the 2004 teacher questionnaire, teachers were asked their views on the school culture and the principal's leadership skills (Table 22). Most teachers in 2004 felt that the principal promoted cooperation among teachers and worked to improve the school environment. Roughly 89 percent of primary school teachers agreed or strongly agreed with the statement, "The principal works hard to 42 improve the school environment and the construction of school culture". About 89 percent of primary school teachers agreed or strongly agreed with the statement, "The principal emphasizes teacher cooperation", and 84 percent agreed or strongly agreed with the statement, "The principal does well in organizing teachers to work together". Most principals were reported by teachers to hold regular staff meetings. About 88 percent of primary teachers agreed or strongly agreed with the statement, "The principal regularly holds staff meetings" (Table 22). Table 22. Teacher Responses to Statements about Principals Leadership by School Level in 2004 (Percentages) Strongly Disagree Disagree Undecided Agree Strongly Agree The principal does not give new teachers guidance. Primary School Teachers (n=1018) 9.53 56.68 18.37 13.65 1.77 Junior High School Teachers (n=1332) 9.83 55.03 17.87 14.41 2.85 The principal is a very good source of information for teaching. Primary School Teachers (n=1018) 2.16 9.82 21.91 55.70 10.41 Junior High School Teachers (n=1332) 1.95 10.06 25.00 52.25 10.74 The principal does well in organizing teachers to work together. Primary School Teachers (n=1018) 1.18 4.72 10.12 60.31 23.67 Junior High School Teachers (n=1332) 1.28 7.36 11.26 58.93 21.17 The principal encourages teachers to use different teaching approaches. Primary School Teachers (n=1018) 0.20 1.47 5.60 65.32 27.41 Junior High School Teachers (n=1332) 0.83 3.68 7.06 67.19 21.25 The principal regularly holds staff meetings. Primary School Teachers (n=1018) 1.57 3.24 7.37 68.57 19.25 Junior High School Teachers (n=1332) 0.83 6.46 4.95 64.26 23.50 The principal emphasizes cooperation among teachers. Primary School Teachers (n=1018) 0.69 2.75 7.47 71.02 18.07 Junior High School Teachers (n=1332) 0.53 4.05 7.73 72.67 15.02 The principal works hard to improve the school environment and the construction of school culture. Primary School Teachers (n=1018) 1.38 3.24 6.19 63.06 26.13 Junior High School Teachers (n=1332) 0.68 5.11 5.18 63.36 25.68 The principal communicates with all the school staff and makes them aware of their importance to the school. Primary School Teachers (n=1018) 1.47 7.17 13.56 59.43 18.37 Junior High School Teachers (n=1332) 2.10 9.98 17.04 55.93 14.94 Source: GSCF-2 Teacher Questionnaire. 43 A majority of teachers felt that their principal upheld high standards for students: 83 percent of primary school teachers agreed or strongly agreed with the statement, "The principal has strict requirements for the students" (Table 23). In addition, Table 23 shows that most teachers, though not the vast majority, reported that the principal communicated with and received support from parents: 70 percent of primary school teachers disagreed or strongly disagreed with the statement, "The principal does not communicate much with parents". Similarly, 71 percent of primary school teachers agreed or strongly agreed with the statement, "The principal receives support from parents". Most teachers reported that their principal was open to new ideas. About 72 percent of teachers disagreed or strongly disagreed with the statement, "The principal is resistant to new ideas" (Table 23). 44 Table 23. Teacher Responses to Statements about Selected Principal School Management Strategies by School Level in 2004 (Percentages) Strongly Disagree Disagree Undecided Agree Strongly Agree The principal carries out fundraising activities for the school. Primary School Teachers (n=1018) 6.88 19.55 18.86 47.84 6.88 Junior High School Teachers (n=1332) 8.03 19.74 21.77 46.10 4.35 The principal has strict requirements of students. Primary School Teachers (n=1018) 0.88 6.39 9.63 69.55 13.56 Junior High School Teachers (n=1332) 3.30 10.96 8.78 64.86 12.09 The principal has poor disciplinary management. Primary School Teachers (n=1018) 10.02 57.47 11.69 19.25 1.57 Junior High School Teachers (n=1332) 9.38 54.28 12.54 22.22 1.58 The principal is resistant to new ideas. Primary School Teachers (n=1018) 11.98 59.72 14.44 11.79 2.06 Junior High School Teachers (n=1332) 13.96 59.83 14.71 9.08 2.40 The principal does not communicate much with parents. Primary School Teachers (n=1018) 11.20 58.64 15.52 13.06 1.57 Junior High School Teachers (n=1332) 9.38 55.26 16.67 17.42 1.28 The principal receives support from parents. Primary School Teachers (n=1018) 1.96 6.88 20.24 58.74 12.18 Junior High School Teachers (n=1332) 0.68 9.31 26.05 55.41 8.56 The principal emphasizes a cooperative relationship between the school and the community. Primary School Teachers (n=1018) 1.18 4.72 12.57 63.46 18.07 Junior High School Teachers (n=1332) 0.53 6.83 15.47 59.23 17.94 The principal uses resources appropriately. Primary School Teachers (n=1018) 1.77 7.07 17.19 63.56 10.41 Junior High School Teachers (n=1332) 3.30 9.38 16.52 61.49 9.31 Source: GSCF-2 Teacher Questionnaire. In addition, most teachers reported that their principals supported teaching innovation: roughly 93 percent of teachers agreed or strongly agreed with the statement, "The principal encourages teachers to use different teaching approaches" (Table 22). Moreover, our data suggest a good deal of supervision of teacher behavior in the classroom. Three-fourths of primary school teachers disagreed or strongly disagreed with the statement, "The principal has never observed me teaching in the classroom" (Table 24). 45 Interestingly, although principals met regularly with teachers, encouraged cooperation, and encouraged new teaching methods, less than half of primary school teachers felt that the principal included them in making school management decisions: 48 percent agreed or strongly agreed with the statement, "The principal lets me participate in school management decisions"(Table 24). On the other hand, referring back to Table 23, there is not a widespread impression of miss-allocation of resources on the part of principals. 74 percent of teachers agreed or strongly agreed with the statement, "The principal uses resources appropriately" (Table 23). Table 24. Teacher Reponses to Statements about Relationships with Principals by School Level in 2004 (Percentages) Strongly Disagree DisagreeUndecided Agree Strongly Agree The principal gives me opportunities for personal development. Primary School Teachers (n=1018) 1.77 7.86 16.99 61.98 11.39 Junior High School Teachers (n=1332) 1.28 12.69 16.82 60.06 9.16 The principal has never observed me in class. Primary School Teachers (n=1018) 15.91 59.53 8.25 13.75 2.55 Junior High School Teachers (n=1332) 13.81 61.41 6.91 15.54 2.33 The principal draws conclusions about me without having observed my teaching. Primary School Teachers (n=1018) 12.97 58.55 16.99 10.02 1.47 Junior High School Teachers (n=1332) 10.44 57.51 18.09 12.61 1.35 The principal uses rewards and punishment as a means to influence my teaching. Primary School Teachers (n=1018) 6.68 39.00 15.52 35.46 3.34 Junior High School Teachers (n=1332) 5.63 33.11 17.87 37.16 6.23 The principal has high expectations for me. Primary School Teachers (n=1018) 1.77 7.96 24.46 56.29 9.53 Junior High School Teachers (n=1332) 1.20 9.91 27.40 53.98 7.51 The principal lets me participate in school management decisions. Primary School Teachers (n=1018) 5.70 24.07 22.00 43.71 4.52 Junior High School Teachers (n=1332) 11.34 34.38 21.92 29.43 2.93 The principal is very respectful towards me. Primary School Teachers (n=1018) 1.67 8.94 19.74 58.74 10.90 Junior High School Teachers (n=1332) 1.50 7.96 20.65 63.36 6.53 Source: GSCF-2 Teacher Questionnaire. 46 On a personal level, Table 24 shows that most teachers feel respected and supported by their principals. About 70 percent of primary school teachers agreed or strongly agreed with the statement, "The principal is very respectful towards me". About 73 percent of teachers agreed or strongly agreed with the statement, "The principal gives me opportunities for personal development". Junior High Schools - Background We refer back to Table 15 and see that in 2004, a junior high school teacher was 32 years old, on average, with 11 years of teaching experience, and had taught in the same school for 7 years. Table 16 provides basic information about teachers and their hometown, gender, ethnicity, and teaching credentials. Similar to primary school teachers, over 90 percent of junior high school teachers are teaching in the same county. Junior high school teachers were predominantly male (roughly 72 percent) and overwhelmingly Han Chinese (roughly 99 percent) in 2004. Over three-fourths of junior high school teachers were married: 76 percent in 2004. In 2004, roughly 93 percent of junior high school teachers reported being a gongban teacher. Less than 1 percent of junior high school teachers reported being a minban teacher or other teacher. About 6 percent of junior high school teachers were daike teachers. 47 Qualifications and Status We refer back to Table 17 to discuss educational qualifications and status for junior high school teachers. As discussed above, response categories for teacher educational levels in 2004 included primary school, junior high school, senior high school, vocational middle school, vocational college, four-year college, graduate school, and other. Table 17 shows that junior high school gongban teachers had primarily attended vocational colleges and four-year colleges: 74 percent had attended a vocational college and 17 percent had attended a four-year college. Junior high schools hired very few minban, daike, or other kinds of teachers: there were just 10 minban teachers in the sample, 75 daike teachers, and 13 other teachers. Among those daike teachers present, most had attended vocational college (79 percent). Table 18 describes teacher ranks in 2004. There are four junior high school teacher rankings, including levels 3, 2, 1, and high, where 3 is the lowest and high is the highest ranking. In 2004, most junior high school teachers reported to have a middle school, level two 2 or level 1 qualification: 42 percent and 18 percent, respectively. Workloads and Work Activities We refer back to Table 16 and Table 19 to investigate the workload of junior high school teachers. In 2004, junior high school teachers reported 48 spending an average of 10 hours preparing classes each week, 11 hours grading papers, and 3 hours a week on political and professional study. About 36 percent of junior high school teachers reported being farmers and 7 percent reported having a part-time job in 2004 (Table 16). Teachers spent about 2 hours per week at other work, including farming (Table 19). Professional Development As described above, Table 20 shows professional development activities for teachers. Similarly to primary school teachers, junior high school teachers participated more frequently in professional development activities in their own school than at other schools, at district-level sessions, or at teacher training institutions. At their own school, 36 percent of junior high school teachers reported participating in peer observation and evaluation of teaching once a week, 54 percent participated on a weekly basis in teaching and research activities led by their teaching and research section (jiaoyanzu) and 34 percent participated weekly in activities organized by their school in 2004. Roughly 49 percent of junior high school teachers had never participated in teaching and research activities in another school, and 35 percent had never participated in district-level teaching and research activities in 2004. 49 Satisfaction with Various Dimensions of Work We refer back to Table 21 to examine junior high school teacher satisfaction with various dimensions of work. Roughly 80 percent of junior high school teachers agreed or strongly agreed with the statement, "I am satisfied with teaching as a job" in 2004--one in five teachers was not satisfied. Interestingly, a full 36 percent of junior high school teachers agreed or strongly agreed with the statement, "I want to change my career". We next examined teacher satisfaction with salary and benefits in order to understand how 80 percent of teachers were satisfied with teaching as a job, yet 36 percent want to change their careers. We found that salary was an area of dissatisfaction: just 39 percent of junior high school teachers agreed or strongly agreed with the statement, "I am satisfied with the salary". Still fewer teachers believed that their job benefits were good: only 12 percent of junior high school teachers agreed or strongly agreed with the statement, "Apart from my salary, my job benefits are good". Teacher interest in changing careers might also be linked to a low level of satisfaction with teacher evaluations. Small majorities of junior high school teachers reported fair and equitable evaluation of teachers (58 percent agreed or strongly agreed with the statement, "The evaluation and ranking of teachers in 50 the school is fair and equitable"), and reported satisfaction with the evaluation of outstanding teachers (60 percent). Attitudes about School Management and School Culture We refer back to Tables 22, 23, and 24 to understand junior high school teacher's views on school management and school culture. In 2004, junior high school teachers' views on school management were similar to those of primary school teachers. Tables 22 and 23 show that teachers felt that their principal supported cooperation among teachers, among the school and community, and that the principal worked hard to improve the school environment and culture. For example, about 89 percent of teachers agreed or strongly agreed with the statement, "The principal works hard to improve the school environment and the construction of school culture". Teacher cooperation may be an important piece of the school environment. In our sample, roughly 88 percent of teachers agreed or strongly agreed with the statement, "The principal emphasizes teacher cooperation" (Table 22). Having a strong relationship with the community may also be important for improving the school environment, and 77 percent of teachers agreed or strongly agreed with the statement, "The principal emphasizes a cooperative relationship between the school and the community" (Table 23). Teachers reported that principals supported their use of different teaching methods: 88 percent of teachers agreed 51 or strongly agreed with the statement, "The principal encourages teachers to use different teaching approaches" (Table 22). Most teachers view principals as open-minded: 74 percent of teachers disagreed or strongly disagreed with the statement, "The principal is resistant to new ideas") (Table 23). Three-fourths of teachers reported that their principals observed their teaching (75 percent disagreed or strongly disagreed with the statement, "My principal has not observed me in class") (Table 24). Less than 35 percent of teachers reported that their principals included them in management decisions, (32 percent of teachers agreed or strongly agreed with the statement, "The principal lets me participate in management decisions") (Table 24). Roughly 71 percent of teachers felt that the principal used resources appropriately (agreed or strongly agreed with the statement, "The principal uses resources appropriately") (Table 23). Principals treated teachers respectfully, in teachers' views. About 70 percent of teachers agreed or strongly agreed with the statement "The principal is very respectful towards me" (Table 24). Classroom Learning Environments Some of the goals of the curriculum and teaching reforms being implemented in China at present are intended to make the classroom a more caring, engaging, and interactive place for children. For example, Ministry of 52 Education (2002) documents cited in Sargent (2005) characterize the new quality education as follows: Classroom teaching should lay emphasis on the students own thinking process...This requires the teacher to be good at creating an open classroom environment, fostering a positive and comfortable atmosphere and encouraging students expressions of new, different and unconventional ideas...Teachers must work hard to protect students curiosity, desire for learning, and imagination...Learning arises out of questioning...Teachers should respect students individual personalities, and pay attention to the differences between the students in order to satisfy the learning needs of individuals, should create an educational environment that ... enables every student to develop their potentialities fully (Ministry of Education 2002: 9). Our survey asked both children in the target sample and teachers to report on several related aspects of classroom environments, as well as on broader aspects of school quality. Below, we describe some of the findings from these reports. Children's Perspectives Table 25 shows student perceptions about teachers' attitudes toward students, tabulated by survey year and gender. Across these categories, a substantial majority of students agreed with statements that the teachers will 53 praise them for industriousness, that the teachers listen to them, that teachers care about students, and that teachers treat students fairly. Table 25. Childrens Views on Teachers by Gender in 2000 and 2004 (Percentages) 2000 2004 Girls Boys chi2 Girls Boys chi2 If I study hard the teacher will praise me. n=831 n=945 4.33 (3) n=831 n=945 9.73 (3)* Strongly Disagree 8.90 6.67 9.39 6.24 Disagree Somewhat 10.47 11.53 25.75 25.50 Agree Somewhat 44.16 46.88 47.53 53.12 Strongly Agree 36.46 34.92 17.33 15.13 Most teachers listen to me. n=831 n=945 2.24 (3) n=831 n=945 2.80 (3) Strongly Disagree 10.83 12.70 13.96 12.17 Disagree Somewhat 23.95 24.55 34.18 37.46 Agree Somewhat 41.88 41.48 41.16 39.37 Strongly Agree 23.35 21.27 10.71 11.01 The teachers at my school care about the students. n=831 n=945 3.93 (3) n=831 n=945 4.18 (3) Strongly Disagree 1.81 2.33 2.41 2.65 Disagree Somewhat 4.33 5.61 8.42 10.48 Agree Somewhat 36.58 38.84 52.83 54.39 Strongly Agree 57.28 53.23 36.34 32.49 Teachers at our school treat students fairly. n=831 n=945 2.87 (3) n=831 n=945 .90 (3) Strongly Disagree 3.97 3.49 5.42 4.76 Disagree Somewhat 9.87 8.15 16.73 15.77 Agree Somewhat 47.05 50.37 47.89 48.15 Strongly Agree 39.11 37.99 29.96 31.32 Source: GSCF-1 and GSCF-2 Child and Household Questionnaires. *=p<.05 The vast majority of students--about 90 percent of girls and boys in the 2000 and 2004 waves--felt that teachers cared about students. A small, but still substantial, majority felt that students were treated fairly by teachers. 86 percent of girls and 88 percent of boys in 2000 agreed that there was fair treatment; comparable figures were 78 and 79 percent in 2004. In 2004, roughly 82 percent 54 of children in school and 84 percent of children out of school felt that they were fairly treated. 55 Table 26. Childrens Views on Schooling by Gender in 2000 and 2004 (Percentages) 2000 2004 Girls Boys chi2 Girls Boys chi2 The teaching quality at my school is good. n=831 n=945 1.76 (3) n=831 n=945 1.93 (3) Strongly Disagree 3.49 3.60 5.54 4.66 Disagree Somewhat 12.15 10.26 19.98 19.68 Agree Somewhat 43.68 45.50 48.13 51.01 Strongly Agree 40.67 40.63 26.35 24.66 Going to school is important for my future. n=831 n=945 3.46 (3) n=831 n=945 9.62 (3)* Strongly Disagree 1.44 1.59 0.96 1.27 Disagree Somewhat 5.17 3.92 1.68 4.13 Agree Somewhat 32.37 29.84 33.45 31.96 Strongly Agree 61.01 64.66 63.90 62.65 We believe we can learn our lessons well. n=831 n=945 10.15 (3) * n=831 n=945 2.57 (3) Strongly Disagree 3.25 1.27 1.32 2.01 Disagree Somewhat 4.21 5.19 6.62 5.29 Agree Somewhat 50.78 48.68 45.61 46.03 Strongly Agree 41.76 44.87 46.45 46.67 We are assigned a heavy homework load. n=831 n=945 0.10 (3) n=831 n=945 3.60 (3) Strongly Disagree 10.47 10.26 11.67 9.95 Disagree Somewhat 26.96 27.20 44.16 48.15 Agree Somewhat 33.81 34.29 33.57 32.59 Strongly Agree 28.76 28.25 10.59 9.31 Schoolwork is easy for me. n=831 n=945 3.00 (3) n=831 n=945 .80 (3) Strongly Disagree 12.15 11.43 12.15 10.90 Disagree Somewhat 32.61 35.24 51.50 51.75 Agree Somewhat 34.78 35.66 28.52 28.99 Strongly Agree 20.46 17.67 7.82 8.36 If I study hard, I can do well in my studies. n=831 n=945 11.44 (3) * n=831 n=945 3.05 (3) Strongly Disagree 2.17 2.12 1.81 2.33 Disagree Somewhat 5.42 2.43 8.06 6.88 Agree Somewhat 39.83 39.15 42.00 45.19 Strongly Agree 52.59 56.30 48.13 45.61 I like to participate in group activities at school. n=831 n=945 2.28 (3) n=831 n=945 2.52 (3) Strongly Disagree 3.49 2.33 4.57 3.28 Disagree Somewhat 7.34 6.98 9.87 11.01 Agree Somewhat 50.78 51.85 55.84 56.51 Strongly Agree 38.39 38.84 29.72 29.21 Source: GSCF-1 and GSCF-2 Child and Household Questionnaires. *=p<.05 56 Table 26 shows a selection of children's answers to questions about schooling, more broadly. Most children believe that the teaching quality at their school is good. In 2000, 84 percent of girls and 86 percent of boys agreed or strongly agreed that teaching quality was good; in 2004, corresponding figures were 74 and 76 percent. An overwhelming majority of children, over 90 percent across the board, feel that education is important for their futures. On day-to-day aspects of the school experience, we also asked a number of questions. More than 92 percent of students, across the board, agreed with the sentiment, "We believe that we can learn our lessons well." In 2000, 63 percent of girls and boys agreed that, "we are assigned a heavy homework load". In 2004, corresponding numbers were roughly 44 percent for girls and 42 for boys. Just over half of children in 2000 and a little over a third in 2004 agreed or strongly agreed that schoolwork was easy. There is a high degree of agreement with the notion that studying pays off: under one in ten girls or boys in either year disagreed with the notion that "If I study hard, I can do well in my studies." The final item in Table 26 addresses whether students enjoy participating in group activities at school. Most students do. In 2000, 11 percent of girls and 9 percent of boys disagreed, somewhat or strongly, with this statement. In 2004, 14 percent of both groups did so. 57 Table 27. Childrens Feelings about School by Gender in 2000 and 2004 (Percentages) 2000 2004 Girls Boys chi2 Girls Boys chi2 I feel happy in school. n=831 n=945 3.47(3) n=831 n=945 1.46 (3) Strongly Disagree 5.66 5.71 6.14 4.97 Disagree Somewhat 10.71 11.64 15.64 16.83 Agree Somewhat 48.74 51.85 55.23 55.24 Strongly Agree 34.90 30.79 22.98 22.96 I often do not want to go to school. n=831 n=945 1.63 (3) n=831 n=945 3.61 (3) Strongly Disagree 49.70 48.04 47.53 45.29 Disagree Somewhat 28.28 30.69 35.38 34.60 Agree Somewhat 10.47 10.79 13.12 14.50 Strongly Agree 11.55 10.48 3.97 5.61 I often feel bored in school . n=831 n=945 7.58 (3) n=831 n=945 5.83 (3) Strongly Disagree 31.17 25.40 32.97 31.64 Disagree Somewhat 43.44 47.20 54.63 52.06 Agree Somewhat 15.76 17.67 9.99 12.59 Strongly Agree 9.63 9.74 2.41 3.70 I often feel lonely in school. n=831 n=945 2.50 (3) n=831 n=945 5.98 (3) Strongly Disagree 34.06 31.43 31.77 33.23 Disagree Somewhat 39.35 40.85 50.54 48.15 Agree Somewhat 16.49 15.87 13.84 12.49 Strongly Agree 10.11 11.85 3.85 6.14 I like to learn new things in school. n=831 n=945 1.49 (3) n=831 n=945 2.91 (3) Strongly Disagree 5.42 4.76 4.21 2.75 Disagree Somewhat 11.07 9.63 9.27 9.74 Agree Somewhat 43.20 44.23 49.94 50.79 Strongly Agree 40.31 41.38 36.58 36.72 Source: GSCF-1 and GSCF-2 Child and Household Questionnaires. *=p<.05 In Table 27, we present responses to questions about children's feelings at school. In 2000, about 84 percent of girls and 83 percent of boys reported feeling happy at school; the number was 78 for both groups in 2004. About one in five enrolled students, and nearly one in four out-of-school children, reported disagreement with this statement. 58 Boredom remains an issue for primary school children. In 2000, about one-fourth of girls and boys agreed with the statement, "I often feel bored at school." Corresponding numbers in the year 2004 were 12 percent and 16 percent. Many children also reported feeling lonely at school. In 2000, about one quarter of girls and boys agreed with the statement "I often feel lonely at school." Less than one in five girls and boys agreed in 2004. A final item in Table 27 regards whether children like learning new things at school. Fewer than one in five children in the full cohort disagreed with this statement in either year. 59 Table 28. Childrens Views on Teaching Styles by Gender in 2000 and 2004 (Percentages) 2000 2004 Girls Boys chi2 Girls Boys chi2 The teacher often pays attention to me in class. n=831 n=945 .11 (3) n=831 n=945 10.90 (3)* Strongly Disagree 10.59 10.37 10.59 6.77 Disagree Somewhat 24.55 24.23 40.79 38.94 Agree Somewhat 38.63 39.37 37.91 41.59 Strongly Agree 26.23 26.03 10.71 12.70 We usually discuss problems together in class animatedly. n=831 n=945 3.82 (3) n=831 n=945 1.46 (3) Strongly Disagree 9.39 6.98 4.69 4.02 Disagree Somewhat 18.17 18.62 21.42 19.79 Agree Somewhat 43.80 46.24 51.74 52.70 Strongly Agree 28.64 28.15 22.14 23.49 In class the teacher lectures while we listen. n=831 n=945 3.23 (3) n=831 n=945 3.76 (3) Strongly Disagree 6.38 4.87 8.30 6.98 Disagree Somewhat 15.16 13.65 33.21 35.45 Agree Somewhat 38.27 40.74 42.12 43.70 Strongly Agree 40.19 40.74 16.37 13.86 The teacher encourages me to ask questions. n=831 n=945 3.44 (3) n=831 n=945 5.66 (3) Strongly Disagree 3.97 2.43 1.56 2.86 Disagree Somewhat 7.58 7.83 6.50 7.41 Agree Somewhat 44.16 44.55 50.78 52.38 Strongly Agree 44.28 45.19 41.16 37.35 Source: GSCF-1 and GSCF-2 Child and Household Questionnaires. *=p<.05 Our last table on children's perspectives on schooling addresses children's perceptions of teaching styles in the classroom. The first question asks students to rate agreement with the statement "The teacher often pays attention to me in class." More than a third of children disagreed with this statement in 2000 and just about half did so in 2004. We asked a number of questions related to the progressive teaching styles encouraged in the recent curriculum reforms. About three-fourths of girls and boys in both years agreed with the sentiment, "We usually discuss problems 60 together in class animatedly". Additionally, we see a substantial drop in the perception of lecturing as a teaching style by children in the sample between 2000 and 2004. In 2000, about 78 percent of girls and 81 percent of boys agreed with the statement, "In class, the teacher lectures, while we listen." In 2004, comparable percentages were 58 percent for both girls and boys. We do not test this difference statistically here, but speculate that such a difference could be linked plausibly to changing practices over time, or changing practices by school level. Teachers' Perspectives In 2004, we also asked teachers in our sample a number of questions about their practices in the classroom. Their responses are probably best interpreted as a combination of what they believe that they are doing in the classroom, and what they believe that they should be doing in the classroom. Most of these items pertained to the use of traditional versus "new curriculum" teaching practices. Responses to a selection of these items are reported in Table 29, and attest to the reported use of both kinds of practices. Teachers report use of a number of traditional teaching techniques. Among primary school teachers, about 42 percent said that they frequently drill, and the number was similar--40 percent--for junior high school teachers. About 53 percent of primary school teachers and 59 percent of junior high school teachers reported frequent lecturing. 61 About 42 percent of primary school teachers and 37 percent of junior high school teachers reported frequently asking students to find answers in their textbooks. One interesting exception is memorization. Just four percent of primary or junior high school teachers reported frequent use of memorization; about 37 percent of primary school teachers and 30 percent of junior high school teachers actually say they never use this technique. 62 Table 29. Teacher Self-Reports of Teaching Techniques by School Level in 2004 (Percentages) Never Sometimes Frequently Drill Primary School Teachers (n=1018) 3.93 53.83 42.24 Junior High School Teachers (n=1332) 2.70 57.21 40.09 Ask Students Open-Ended Questions Primary School Teachers (n=1018) 1.87 46.56 51.57 Junior High School Teachers (n=1332) 1.28 53.90 44.82 Include Classroom Discussions Primary School Teachers (n=1018) 0.39 23.48 76.13 Junior High School Teachers (n=1332) 0.30 32.36 67.34 Facilitate Small Group Activities Primary School Teachers (n=1018) 1.77 38.61 59.63 Junior High School Teachers (n=1332) 1.73 47.45 50.83 Student Choral Response Primary School Teachers (n=1018) 3.14 63.65 33.20 Junior High School Teachers (n=1332) 2.85 60.21 36.94 Encourages Individual Student Response Primary School Teachers (n=1018) 0.59 25.44 73.97 Junior High School Teachers (n=1332) 0.38 29.05 70.57 Lectures Primary School Teachers (n=1018) 1.77 45.68 52.55 Junior High School Teachers (n=1332) 1.43 39.26 59.31 Memorization Primary School Teachers (n=1018) 36.84 59.04 4.13 Junior High School Teachers (n=1332) 29.73 65.84 4.43 Inquiry Learning Primary School Teachers (n=1018) 4.03 58.55 37.43 Junior High School Teachers (n=1332) 3.75 63.06 33.18 Ask Students to Find Answers in the Textbook Primary School Teachers (n=1018) 8.25 49.71 42.04 Junior High School Teachers (n=1332) 5.71 57.36 36.94 Ask Students to Participate in Activities that Require the Connection of Theory and Practice Primary School Teachers (n=1018) 2.06 59.43 38.51 Junior High School Teachers (n=1332) 3.00 60.74 36.26 Source: GSCF-2 Teacher Questionnaire. 63 Considering "new curriculum" techniques, about half of primary and junior high school teachers said that they frequently ask open-ended questions. Three-quarters of primary school teachers and two-thirds of junior high school teachers said that they include classroom discussion in their practice. Well over one-half of primary teachers, and about half of junior high school teachers, said that they frequently facilitate small-group activities. Just about a third of both groups reported frequent use of student choral response (students responding to questions in unison). About three-quarters of primary and junior high school teachers reported encouraging individual student responses. About one-third of primary and junior high school teachers reported use of inquiry learning. SCHOOL PERSISTENCE: BARRIERS AND SUPPORTS We turn now to part two of the report, which consists of an analysis of barriers and supports to school persistence. We begin by showing that the desired levels of education, as expressed in childrens stated preferences, are extremely high. Table 30 shows children's educational aspirations for children in the years 2000 and 2004. The results are tabulated by sex,. In 2000, about half of children (46 percent of girls and 51 percent of boys) aspired to a college education. In 2004, children aspiring to tertiary level schooling had risen to 62 percent for girls and 64 percent for boys. What are the barriers that keep these 64 high aspirations from being realized? Educational bottlenecks at the upper secondary level are a serious problem (Wang 2006). However, here, we consider the barriers that face children of an age to be in primary and lower secondary education--stages of education that the government seeks to universalize as compulsory. Table 30. Childrens Educational Aspirations by Gender in 2000 and 2004 (Percentages) 2000 2004 Girls Boys chi2 Girls Boys chi2 What is the highest level of schooling you want to complete? n=831 n=945 13.42 (5)* n=831 n=945 15.73 (5)* Primary School 5.42 3.39 2.41 0.63 Junior High School 9.99 8.89 6.98 7.94 Senior High School 19.13 15.13 15.16 14.92 Junior Trade School 9.99 10.16 4.57 2.65 Senior Trade School 9.03 11.53 8.78 10.16 University or Higher 46.45 50.90 62.09 63.70 Source: GSCF-1 and GSCF-2 Child and Household Questionnaires. *=p<.05 Informant Reports on Reasons for School-Leaving We begin by looking at reports of reasons for non-enrollment in Gansu villages, as reflected in village leaders' reports of the contributors to school- leaving in the local village (Table 31). This question was asked in both 2000 and 2004, separately for primary and junior high school levels of education. For each item, we asked the village leader to state whether or not that item was a contributor to school leaving. The percentage refers to the percent of leaders who reported that each item contributed to school leaving. 65 Table 31. Village Leaders Assessment of Factors Contributing to School Leaving in Sampled Gansu Villages (% responding that the item is a factor in school-leaving). Primary Junior High School 2000 2004 2000 2004 Poor school quality 7 2 6 8 School too distant 8 5 17 12 School is over-subscribed 3 0 3 5 Tuition/costs too high 41 11 53 49 Parents unwilling to send children to school 15 4 10 16 Family needs labor at home 10 6 23 25 Child does not want to go to school 11 19 27 56 Childs grades are poor 14 18 41 65 Total Number of Villages=100 Source: GSCF-1 and GSCF-2 Village Leader Questionnaires. Table 31 shows that in the year 2000, 41 of the 100 village leaders stated that primary tuition being too high was a contributor to school leaving; 15 felt that parental unwillingness to send children to school was a problem; 14 believed that low grades was a problem; 11 percent felt that child's own unwillingness to attend school; and 10 felt that parental labor needs were a problem. By 2004, tuition was cited as a problem by only 11 village leaders, parental unwillingness was cited by only 4 leaders, and parental labor needs, by only 6. However, child- related factors were mentioned more in the latter year: 19 leaders cited children not wanting to attend school and 18 percent cited poor grades. At the junior high school level, school supply issues were still perceived as concerning by a minority of village leaders. Distance was cited by 17 village leaders in 2000 and 12 leaders in 2004, and quality was cited by 6 leaders in 2000 66 and 8 leaders in 2004. Over-enrollment was reported by only 3 villages in 2000 and 5 in 2004. The significance of household economic constraints is clear at the junior high school level: 53 percent of leaders cited tuition in 2000 and 49 percent cited it in 2004 as a reason for non-enrollment. After tuition, parental need for labor was reported by about one-fourth of leaders, and parental unwillingness to send children to school by 10 leaders in 2000 and 16 in 2004. However, village leader reports are also consistent with an important, and possibly even rising, role of children's own preferences and performance. In the year 2000, more than a fourth of leaders reported children not wanting to attend school as a problem; by the year 2004, this was perceived as an issue by more than half of village leaders. In 2000, 41 percent of leaders reported children's poor grades as a reason for leaving school; by 2004, 65 percent of leaders reported this issue as a contributing factor to non-enrollment. These findings suggest that even as economic considerations remain crucial, children's own contributions to the schooling decisions are perceived as being important, and perhaps increasingly so. Significantly, reports by household heads, mothers and children themselves are consistent with notions about the dual contributions of family economy and children themselves to decisions about schooling. Figure 3 shows the distribution of household head reports of the main reason that the child is not in school for households of sample 67 children who are not in school. The modal category was child's unwillingness to attend school (40 percent of heads), followed by financial difficulty (32 percent of heads) and poor school performance (11 percent of heads). Figure 3. Percent Distribution of Household Head Reports of Primary Reason for School Leaving by Target Child, 2004 (N=255 ) 70 60 50 40 40 32 30 20 11 10 4 6 5 0 0 1 0 0 ytil of r the ol alic ts' n ytl ying ess old e th he qual g ttin onissi hocs iculty ren jectio fl ud hg edt elvel Ot diff ssengn Pa ob Se St ifficud ou ing hoocS m naniF Sickn Ge ad het illiwnu ginebt en ecp onita No ex Finish duce 68 Figure 4. Out-of-School Childrens Mothers Reports of Factors Contributing to School Leaving, 2004 (% responding that the item is a factor, N=218) 70 60 50 50 41 40 32 30 20 10 8 12 8 10 5 3 0 from ntien ve wayar hetfoytil to h to s dr l nd ug ot sel oo go tayl em or ru oorp fo pes sd'lhic conin aft eno obl ol sch ot e dlhic ot e pr ho s uaq no to fa elih th ev th ling ha d ntawt poyr l ve htal sc oni at o tosilo verysiit, ngih ootsi uld co child oo not ted ede no imselfh/reh ddi hoocs he eact oolhcs e ol th milyaf d hwtrowt noy p neo hel d m sch ne h d did dlhic ha viola gulerro no an choS meoh ehT ddiyli uc e sen m dlihc chos child Th asw maf ehT dlhic e It atht nds heT ha heT heT Th Figure 5. Out-of-School Childrens Reports of Factors Contributing to School Leaving, 2004 (% responding that the item is a factor, N=130) 70 60 48 50 42 40 32 30 20 17 10 8 5 4 6 6 0 ol r fa hocs o ngih d to s e ytial orf n milyaf aft atrxe th s og es alth heT tosaw ol he ed eact tuitio qu e dn' e tonsaw ghih y het ition milaf ed nd tons to ho adrg meblo lesurlo ht tu ha sc or My pr owL Th oulc th It gn orw The eedn waI illiw Po atloivI scho 69 We asked mothers of non-enrolled children a different question, about whether or not each of a series of factors contributes to non-enrollment (see Figures 4 and 5). Among mothers, half reported child unwillingness and one- third reported children's poor achievement as factors in school non-enrollment. Family inability to afford school costs was cited by 41 percent of mothers; needing child labor at home, by 12 percent of mothers, and school not being worth the tuition, by 8 percent. School quality and school distance were cited by 5 and 10 percent of mothers, respectively. Among non-enrolled children who were available to complete the child questionnaire, nearly half, 48 percent, report unwillingness to attend school, while 42 percent report poor grades as a contributor to non-enrollment. The household economy explanations are also important here: 32 percent of children reported that their families could not afford tuition, and 17 percent reported that the family needed extra hands at home. Few children reported school quality or distance as barriers. Overall, these descriptive reports show that village leaders, families, and children themselves perceive that both family economic considerations and children's feelings about schooling and performance are important contributors to schooling decisions. Using these reports as a guide, we first consider factors 70 that predict performance and engagement with schooling. We then consider whether performance and engagement, along with household socio-economic status and other factors measured in 2000, can predict subsequent school outcomes. Analysis of Outcomes in 2004 We now turn to an analysis of factors that predict school continuation, and those that predict school discontinuation. While there are many ways to approach these questions, we focus on three issues: First, to what extent are there socioeconomic and gender differences in school continuation? Second, can attributes of schools and teachers in 2000 be linked to school continuation in 2004? Finally, we consider whether school engagement and performance in 2000 can be linked to subsequent school continuation, as reasons for school-leaving presented in Table 31 and Figures 1 to 3 would suggest. 71 Strategy We model the following outcome variables:5 Attainment Change. Regardless of a childs enrollment status in 2000 or 2004, we calculated a difference measure of the change in highest attainment reported by the household head between 2000 and 2004. Enrollment Status. We report enrollment status as marked in the household questionnaire. This variable combines reported enrollment status for children at home and children not living at home. Attainment of Nine Years of Education. Regardless of a childs enrollment status, we also consider whether or not the child has achieved the nine-year target for compulsory education. We restrict analyses of this outcome to children who were 12 in 2000, and thus old enough to have completed nine years if attending and moving up the grades in a timely manner. We model these outcomes in the year 2004 with the following child socio- demographic characteristics, school characteristics, and teacher characteristics measured in the earlier 2000 survey: 5We estimate random intercepts models, with random effects at the level of school and teacher. For continuous outcomes (change in attainment, and for the 2000 analyses, aspirations, academic confidence, industriousness, and math and language grades), we use the STATA procedure XTMIXED. For binary outcomes (enrollment or nine-year achievement), we use the GLLAMM procedure. 72 Child Socio-Demographic Characteristics. Human capital of parents and the material resources in the household lend themselves fairly easily to measurement in terms of parents' education and family wealth. Here, we use mother's and father's years of education as indicators of human capital. Wealth (logged) is used as a measure of family material resources. Family wealth was constructed from detailed measures of household assets, including the value of housing, fixed capital, and household durable goods. Because gender has historically played an important role in conditioning educational opportunity in rural China, we include an indicator for child gender in all models. Because siblings are commonly perceived to "dilute" family resources--economic, social, or cultural--we also control for number of siblings. Because there may be developmental changes in children's attitude reporting and because older children are less likely to be in school, we include age as a control variable. School Characteristics. The school characteristics that we include are those reported by the principal. They include school type, whether public or other, whether the school has boarders,6 teacher absenteeism (whether the typical teacher 6Very few primary schools had boarding students. We include this variable due to its substantive interest, but we are cautious in interpretation due to small sample size. 73 was absent four or more times per semester), teacher average monthly salary (logged), the presence of daike teachers or minban teachers, the principals gender (female principal) and experience teaching and experience as a principal. Teacher Characteristics. Analyses include the log of teacher's monthly income (logged), including salary and bonuses. We include teacher gender in all analyses, as teacher gender is often cited as a salient factor in perceived quality of education in developing countries. We include teacher's status as a formal, credentialed teacher, years of experience, teachers' lower secondary, upper secondary or tertiary educational attainment,7 and whether or not the teacher is local, meaning from the village. In addition, we are interested in knowing how a students engagement and performance matter or do not matter for subsequent persistence. We use the following measures: 7High school level education includes those who graduate from regular high school or from a specialized teacher training school attended after middle school (zhongzhuan, ). College-level education includes those that graduate from regular universities (very few) or from a normal college (dazhuan, ) following the completion of high school. There are some teachers who take correspondence courses to receive accreditation for dazhuan without ever having completed high school. Because of the design of the survey instrument, we have difficulty distinguishing these cases from teachers who receive zhongzhuan degrees, and so they are categorized as having high school-level education. Thus "tertiary" is equal to one only when the teacher completes both high school and college-level schooling, and "upper secondary" captures all other educational outcomes except those teachers who hold only a middle school degree. 74 Engagement. We define four dimensions of engagement, each based on child's self-report: educational aspirations, academic confidence, industriousness, and alienation from school. Aspirations were converted from levels shown in Appendix 2 to years corresponding to the completion of those levels. Academic confidence, industriousness, and alienation were generated by standardizing the component variables shown in Appendix 2, summing them, and dividing by the number of components. Performance. Our measures of performance are last semester's average math and language grades, reported by teachers (reported as scores). Change-In-Attainment Results 75 Table 32. Mixed Models of Change in Attainment with Child, Teacher, and School Characteristics Change in Attainment, 2000-2004 (1) (2) (3) (4) Demo- (1) + (1)+Teache (1)+School graphics School r Char.s + Teacher Char.s Char.s Female -0.072 -0.071 -0.074 -0.074 Wealth 0.109*** 0.110*** 0.104*** 0.108*** Fathers education (years) 0.023** 0.021** 0.022** 0.021** Mothers education (years) 0.006 0.006 0.005 0.006 Siblings 0.031 0.031 0.038 0.035 Age -0.052* -0.046* -0.050* -0.043* Teacher monthly income 0.056 0.052 0.079 School type=other (ref.=public) 0.217 0.227 School has boarders -0.26 -0.208 Typical teacher absent 4+times per semester -0.300* -0.308* Teacher average monthly salary 0.01 -0.006 School has daike teachers 0.021 0.029 School has minban teachers -0.083 -0.07 Female principal -0.268 -0.281 Principal experience (years teaching) 0.003 0.003 Principal experience (years as principal) 0.003 0.002 Female teacher 0.076 0.071 Formal teacher 0.032 0.052 Teacher experience (years) 0.003 0.002 Teacher education=upper secondary 0.125 0.117 Teacher education=tertiary 0.056 0.05 Local teacher 0.02 0.006 Constant 3.216*** 2.717** 2.726*** 2.454* Random effects parameters SD (Constant--Teacher) -1.507*** -1.601*** -1.529*** -1.585*** SD (Constant--School) -1.229*** -1.234*** -1.210*** -1.221*** SD (Residual) -0.115*** -0.115*** -0.115*** -0.116*** Notes: N=1693. * p<0.05; ** p<0.01; *** p<0.001 Change-in-attainment results are presented in Table 32. We show four specifications: (1) a model with only child background characteristics; (2) a model with child background and school characteristics; (3) a model with child 76 background and teacher characteristics; and (4) a model with child background, school, and teacher characteristics. In these specifications, it is clear that socioeconomic status matters: regardless of whether school and teacher characteristics are accounted for, logged wealth and fathers education are significant predictors of childrens grade advancement between the surveys. A standard deviation increase in the logged wealth measure is associated with about a tenth of a year more attainment, and a standard deviation increase in fathers years of education, with about .08 years more attainment. The scope of the effect is similar in the models with or without school and teacher measures, indicating that these factors do not explain socioeconomic effects on change in attainment. Perhaps more surprising than the presence of socioeconomic differences is the absence of gender differences: gender is not statistically significant in these models. Among school and teacher measures, the only statistically significant finding is that teacher absenteeism in 2000 is associated with slower advancement between 2000 and 2004. 77 Enrollment Results Table 33. GLLAMM Models of Continued Enrollment and Nine-Year Achievement with Child, Teacher, and School Characteristics Enrollment, 2004 Nine-Year Achievement, 2004 (1) (2) (3) (4) (1) (2) (3) (4) Demo- (1) + School (1)+Teacher Char.s (1)+School + Demo-graphics (1) + School (1)+Teacher (1)+School + graphics Char.s Teacher Char.s Char.s Char.s Teacher Char.s Female -0.397* -0.419* -0.391* -0.414* - 0.39 - 0.416 - 0.379 - 0.402 Wealth 0.323** 0.344*** 0.328** 0.339*** 0.597** 0.481* 0.457* 0.416* Fathers education (years) 0.062* 0.065* 0.060* 0.062* 0.045 0.038 0.052 0.045 Mothers education (years) 0.049 0.056* 0.049 0.056* 0.100* 0.095* 0.098* 0.099* Siblings 0.075 0.069 0.074 0.08 0.022 0.101 0.165 0.18 Age -0.955*** -0.951*** -0.937*** -0.941*** 0.739* 0.678* 0.735* 0.678* Teacher monthly income - 0.175 - 0.037 2.065*** 1.452* School type=other (ref.=public) 0.055 0.083 - 0.209 0.131 School has boarders - 0.21 0.027 2.296* 1.605 Typical teacher absent 4+times per semester 0.17 0.11 - 0.142 - 0.388 Teacher average monthly salary - 0.016 - 0.045 1.171 1.121 School has daike teachers 0.087 0.183 - 0.33 - 0.186 School has minban teachers 0.353 0.435 -2.201*** -2.015*** Female principal 0.65 0.67 - 0.21 0.765 Principal experience (years teaching) - 0.005 - 0.005 0.02 0.027 Principal experience (years as principal) - 0.011 - 0.012 - 0.045 - 0.049 Female teacher 0.056 0.107 - 0.471 - 0.577 Formal teacher - 0.354 - 0.27 1.467 0.927 Teacher experience (years) 0.006 0.006 -0.085** -0.074* Teacher education=upper secondary 0.373 0.417 - 0.275 - 0.301 Teacher education=tertiary - 0.022 0.034 0.292 0.129 Local teacher - 0.075 - 0.094 - 0.262 - 0.1 Constant 9.604*** 9.469** 10.528*** 9.698** -15.268*** -20.125* -26.962*** -28.194** Random effects parameters SD (Constant--Teacher) 0.478 0.518 0.359 0.425 2.600*** 2.527*** 2.665*** 2.578*** SD (Constant--School) 0.601*** 0.542** 0.610*** 0.543** 0.43 0. 0. 0. SD (Residual) Notes: Enrollment Models: N=1693. Nine-Year Attainment Models: N=629. * p<0.05; ** p<0.01; *** p<0.001 The same set of specifications are presented in Table 33 for the enrollment outcome. Here again, in each model specification, a family's wealth and paternal education matter for continued enrollment (maternal education matters except in specification 3). One standard deviation increase in logged wealth translates to roughly a 35 percent increase in odds of enrollment, depending on specification. A year more of fathers or mothers education translates to about a 6 percent increase in odds of enrollment (though the mothers education effect is not always significant at conventional levels). In enrollment, gender differences are significant. Girls odds of enrollment are about a third lower than boys. These statements are true whether or not account is taken of earlier school and teacher 78 characteristics, and none of the school or teacher characteristics matter. School and teacher effects are not significant in the enrollment models. Nine-Year Achievement Results Our models of nine-year achievement are presented in Table 33 and use the same four specifications as for the earlier outcomes of change in attainment and enrollment, but again, for an age-restricted sample. Considering the outcome of whether or not children have achieved nine years of schooling, wealth and mothers education matter. Gender shows no significant effect. For this outcome, some teacher and school characteristics mattered: Children with better paid teachers were more likely to complete nine years, while children in schools with minban teachers were much less likely to have completed nine years. Anomalously, teacher experience is negatively associated with student nine-year achievement, and, before taking into account teacher factors, schools with boarders are positively associated. 79 Engagement and Performance as Predictors of Persistence Table 34. Models of Attainment Change, Enrollment and Nine-Year Completion with Achievement and Engagement 2004 Outcomes Change in Attainment Enrollment Nine Years (1) (2) (3) (1) (2) (3) (1) (2) (3) (2) + Child, (2) + Child, (2) + Child, (1) + (1) + (1) + 2000 Measures Engagement Teacher and Engagement Teacher and Engagement Teacher and Performance Performance Performance School Char.s School Char.s School Char.s Aspirations 0.025** 0.019* 0.018* 0.122*** 0.115*** 0.111*** 0.187*** 0.172** 0.140* Academic self confidence 0.193*** 0.120*** 0.112** 0.219 0.092 - 0.026 0.435 0.098 0.259 Industriousness 0.003 - 0.005 0.000 - 0.095 - 0.108 0.021 - 0.066 - 0.066 0.102 Alienation -0.088** - 0.059 - 0.054 0.018 0.075 0.037 -0.549* - 0.377 - 0.270 Math score 0.010*** 0.009** 0.020* 0.021* 0.019 0.009 Lang score 0.006 0.006 0.005 0.000 0.045 0.048* Female - 0.086 -0.391* - 0.420 Wealth 0.080** 0.288** 0.355 Fathers education (years) 0.013 0.059* 0.044 Mothers education (years) 0.000 0.045 0.098* Siblings 0.042 0.047 0.124 Female teacher 0.052 0.107 - 0.668 Formal teacher 0.034 - 0.327 0.664 Teacher experience (years) 0.003 0.013 - 0.050 Teacher education=upper secondary 0.080 0.382 - 0.365 Teacher education=tertiary 0.009 - 0.016 - 0.172 Local teacher - 0.047 - 0.278 - 0.234 Teacher monthly income 0.052 - 0.096 1.182* School type=other (ref.=public) 0.227 0.054 0.200 School has boarders - 0.285 - 0.356 2.185 Typical teacher absent 4+times per semester - 0.244 0.292 - 0.310 Teacher average monthly salary - 0.015 0.017 1.022 School has daike teachers 0.040 0.204 - 0.268 School has minban teachers - 0.023 0.585* -1.849*** Female principal - 0.282 0.623 0.331 Principal experience (years teaching) 0.003 - 0.003 0.043 Principal experience (years as principal) 0.001 - 0.014 - 0.049 Age - 0.034 -0.968*** 0.889** Constant 3.519*** 2.447*** 1.634 0.745* - 0.981 7.687* -2.484*** -6.915*** -34.048*** Random effects parameters SD (Constant--Teacher) -1.513*** -1.456*** -1.560*** 0.924*** 0.974*** 0.688** 2.598*** 2.713*** 2.299*** SD (Constant--School) -1.106*** -1.082*** -1.117*** 0.475* 0.484* 0.407 0.923* 0.624 0.000 SD (Residual) -0.145*** -0.171*** -0.172*** Notes: Enrollment Models: N=1693. Nine-Year Attainment Models: N=629. * p<0.05; ** p<0.01; *** p<0.001 Do indicators of engagement and performance matter? Table 34 shows engagement and performance effects on the three outcomes--change in attainment, continued enrollment, and nine-year attainment. For each outcome, we show engagement and performance coefficients for three specifications. The first specification includes engagement measures only; the second adds performance measures, and the third adds controls for all child, teacher and school variables shown in the full models (specifications marked (4)) in Tables 32 and 33. The point of these tables is to illuminate whether these measures of early engagement are predictive of subsequent outcomes, net of other factors thought 80 to matter. The table shows that early aspirations significantly predict favorable outcomes across all specifications and all outcomes. Math or language achievement is a significant predictor of each outcome in the full model specification. Summary Our models of family, teacher, and school effects on outcomes show that better-off children are more likely to show grade attainment, continued enrollment, and attainment of nine years of basic education. In contrast, the gender story is mixed: girls are less likely to be enrolled, but have not gained fewer grades, nor are they less likely to achieve nine years of education. This finding suggests that boys may start later or repeat more. It is possible that boys are more likely to be encouraged to repeat a grade to complete it successfully, or to increase high school exam scores. One significant finding is that the introduction of school and teacher effects, by and large, does not explain away the advantages of children in better- off families. School and teacher effects do not consistently matter across the three outcomes. Some interesting findings include that teacher absenteeism in 2000 is associated with less attainment between 2000 and 2004; children with better-paid home room teachers are more likely to attain nine years of school; and children in schools with minban teachers are less likely to attain nine years. 81 However, there is not a consistent story of school characteristics that help or hinder childrens persistence. Reports by village-leaders, fathers, mothers, and children themselves, suggest that children's performance and engagement, along with economic considerations, are important contributors to school-leaving in Gansu's rural villages. We do not yet know how to measure engagement well. However, models presented in this section suggest that overall and net of many school and teacher characteristics, aspirations and performance, along with socioeconomic status, are significant determinants of continuation in school.8 Determinants of Engagement and Performance Given that children's performance and preferences are highlighted as potentially important dimensions of enrollment decisions, along with economic factors, we next turn to an analysis of factors associated with favorable attitudes toward education and high grades. We consider aspirations and math and language scores (grades), as these are suggested by our analyses as being predictive of subsequent outcomes. However, we also consider the academic confidence, industriousness, and alienation measures. These measures may be 8The wealth effect on nine-year attainment dissipates with the inclusion of aspirations and performance in the model, suggesting the links among these characteristics. 82 considered indicative of a desirable classroom environment, regardless of whether they predict subsequent school continuation. Strategy Tables 35, 36, and 37 report estimates from models predicting aspirations and academic confidence (Table 35), industriousness and alienation (Table 36), and math and language grades (Table 37). For each outcome, we consider four different specifications. For each outcome, model (1) regresses measures on child characteristics only; model (2) adds school characteristics only; model (3) adds teacher characteristics only; and model (4) adds school and teacher characteristics. 83 Engagement Results Table 35. Mixed Models of Aspirations and Confidence with Child, Teacher, and School Characteristics Aspirations (Years) Academic Confidence (1) (2) (3) (4) (1) (2) (3) (4) Demo- (1) + (1)+Teacher (1)+School Demo- (1) + (1)+Teacher (1)+School graphics School Char.s + Teacher graphics School Char.s + Teacher Predictor Char.s Char.s Char.s Char.s Female -0.349* -0.354* -0.368* -0.367* 0.039 0.039 0.039 0.039 Wealth 0.151 0.163 0.154 0.157 -0.026 -0.027 -0.025 -0.026 Fathers education (years) 0.046 0.045 0.047* 0.046* 0.021*** 0.021*** 0.020*** 0.021*** Mothers education (years) 0.023 0.027 0.023 0.027 -0.005 -0.004 -0.004 -0.004 Siblings -0.233* -0.264* -0.257* -0.269* 0.027 0.026 0.025 0.023 Teacher monthly income -0.215 -0.235 -0.161 0.022 -0.006 0.025 School type=other (ref.=public) 0.507 0.629 -0.015 -0.021 School has boarders 2.008** 1.896** 0.102 0.126 Typical teacher absent 4+times per semester -0.566 -0.521 -0.271** -0.269** Teacher average monthly salary -0.297 -0.281 -0.14 -0.137 School has daike teachers 0.208 0.2 -0.078 -0.075 School has minban teachers 0.08 0.19 0.023 0.03 Female principal -0.153 -0.185 0.121 0.127 Principal experience (years teaching) -0.011 -0.01 -0.003 -0.002 Principal experience (years as principal) 0.005 0.002 0.007 0.007 Female teacher 0.434* 0.493* 0.077 0.076 Formal teacher -0.247 -0.233 -0.053 -0.01 Teacher experience (years) -0.011 -0.009 0.002 0.002 Teacher education=upper secondary -0.178 -0.124 0.024 0.02 Teacher education=tertiary 0.244 0.248 0.014 0.003 Local teacher 0.438* 0.393 0.101* 0.102* Constant (Dummies for grades not shown) 11.041*** 14.076*** 12.629*** 13.838*** 0.034 0.856 -0.003 0.685 Random effects parameters SD (Constant--Teacher) 0.003 0.011 -0.056 -0.027 -2.067*** -2.113*** -2.082*** -2.161*** SD (Constant--School) -0.31 -0.324 -0.308 -0.336 -1.887*** -1.921*** -1.890*** -1.928*** SD (Residual) 1.056*** 1.054*** 1.055*** 1.054*** -0.333*** -0.334*** -0.333*** -0.333*** Notes: N=1693. * p<0.05; ** p<0.01; *** p<0.001 Turning to the engagement measures, we begin with a discussion of child's aspirations. Here, we see that girls, children of less educated fathers (in specifications 3 and 4), and children with more siblings have lower aspirations, across the board. The few children in schools with boarders have higher aspirations, and children with female and local teachers have higher aspirations (results are significant or marginal, depending on specification). Turning to academic confidence, fathers education is the main child factor that matters. 84 Teacher absence is associated with less confidence, and local teachers, with greater confidence. Table 36. Mixed Models of Industriousness and Alienation with Child, Teacher, and School Characteristics Industriousness Scale Alienation Scale (1) (2) (3) (4) (1) (2) (3) (4) Demo- (1) + (1)+Teacher (1)+School Demo- (1) + (1)+Teacher (1)+School graphics School Char.s + Teacher graphics School Char.s + Teacher Predictor Char.s Char.s Char.s Char.s Female 0.139*** 0.139*** 0.137** 0.139** -0.081* -0.080* -0.077* -0.076* Wealth -0.016 -0.012 -0.012 -0.012 -0.025 -0.024 -0.021 -0.022 Fathers education (years) 0.01 0.01 0.01 0.01 -0.004 -0.003 -0.003 -0.003 Mothers education (years) -0.011 -0.011 -0.011 -0.01 -0.007 -0.006 -0.006 -0.006 Siblings 0.042 0.034 0.038 0.036 0.041 0.036 0.033 0.03 Teacher monthly income -0.031 -0.039 -0.015 -0.023 -0.053 -0.041 School type=other (ref.=public) -0.064 -0.085 0.082 0.063 School has boarders 0.162 0.187 0.295 0.251 Typical teacher absent 4+times per semester -0.347** -0.350** 0.016 0.024 Teacher average monthly salary -0.224 -0.228 -0.162 -0.146 School has daike teachers 0.029 0.029 -0.135* -0.149* School has minban teachers -0.012 -0.018 0.034 0.014 Female principal -0.095 -0.082 0.084 0.092 Principal experience (years teaching) -0.003 -0.003 0 0 Principal experience (years as principal) 0.004 0.004 -0.001 0 Female teacher 0.003 0.007 -0.078 -0.078 Formal teacher 0.043 0.07 -0.051 -0.016 Teacher experience (years) 0.003 0.003 0 0 Teacher education=upper secondary 0.051 0.055 -0.146** -0.146** Teacher education=tertiary 0.033 0.022 -0.054 -0.051 Local teacher 0.066 0.055 -0.008 -0.006 Constant (Dummies for grades not shown) -0.271 1.352 -0.254 1.057 0.513* 1.708* 0.976 1.860* Random effects parameters SD (Constant--Teacher) -1.529*** -1.555*** -1.567*** -1.584*** -1.516*** -1.586*** -1.533*** -1.598*** SD (Constant--School) -1.967*** -1.998*** -1.863*** -1.913*** -1.519*** -1.513*** -1.537*** -1.533*** SD (Residual) -0.199*** -0.199*** -0.201*** -0.200*** -0.424*** -0.422*** -0.424*** -0.422*** Notes: N=1693. * p<0.05; ** p<0.01; *** p<0.001 Looking at industriousness in Table 36, only two predictors matter: gender and teacher absence. Girls score higher on this scale, and children in schools with teacher absenteeism score lower. In addition, for alienation in the same table, gender matters: girls are less alienated, on average. The other significant predictors of alienation are having better-educated teachers and whether or not the school has daike teachers. Students in schools with these teachers report less alienation. 85 Performance Results Table 37. Mixed Models of Math and Language Grades with Child, Teacher, and School Characteristics Average Math Score (Grade) Average Language Score (Grade) (1) (2) (3) (4) (1) (2) (3) (4) Demo- (1) + (1) + Teacher (1) + School Demo- (1) + (1) + Teacher (1) + School + graphics School Char.s + Teacher graphics School Char.s Teacher Predictor Char.s Char.s Char.s Char.s Female 0.635 0.576 0.497 0.455 2.224*** 2.194*** 2.126*** 2.109*** Wealth 0.691 0.582 0.577 0.511 0.877* 0.812* 0.853* 0.803* Fathers education (years) 0.308** 0.307** 0.310** 0.309** 0.274** 0.277** 0.275** 0.279** Mothers education (years) 0.315** 0.300** 0.302** 0.293* 0.330** 0.316** 0.331** 0.316** Siblings -0.407 -0.324 -0.258 -0.228 -0.111 -0.061 -0.071 -0.055 School type=other (ref.=public) -0.402 0.004 -1.333 -1.439 School has boarders 0.392 0.019 -2.425 -2.536 Typical teacher absent 4+times per semester -2.669 -2.534 0.78 1.004 Teacher average monthly salary 1.133 0.058 -0.61 -0.902 School has daike teachers -1.106 -1.446 -0.563 -0.879 School has minban teachers -2.138 -1.955 -2.158 -2.286 Female principal -3.438 -2.643 -6.181 -5.815 Principal experience (years teaching) -0.121 -0.123 -0.159 -0.15 Principal experience (years as principal) 0.108 0.101 0.08 0.074 Female teacher 1.696 1.519 1.152 1.039 Formal teacher 6.365** 6.499** 2.608 2.497 Teacher experience (years) -0.053 -0.042 -0.039 -0.023 Teacher education=upper secondary 0.79 0.591 0.952 0.765 Teacher education=tertiary 0.73 0.638 0.514 0.555 Local teacher 1.912 1.615 2.019* 1.721 Teacher monthly income 3.841** 3.412* 1.451 0.819 Constant (Dummies for grades not shown) 63.990*** 61.272*** 32.953** 39.276* 57.679*** 66.359*** 44.577*** 58.897** Random effects parameters SD (Constant--Teacher) 1.336*** 1.327*** 1.314*** 1.328*** 1.463*** 1.456*** 1.426*** 1.439*** SD (Constant--School) 1.716*** 1.702*** 1.683*** 1.674*** 1.668*** 1.649*** 1.681*** 1.659*** SD (Residual) 2.555*** 2.557*** 2.556*** 2.556*** 2.406*** 2.407*** 2.405*** 2.406*** Notes: N=1693. * p<0.05; ** p<0.01; *** p<0.001 Performance results in Table 37, suggest that children's math and language grades are contingent on a number of school and family resources. For math grades, without controlling for school and teacher characteristics, children of better-educated mothers and fathers are doing better. With school and teacher characteristics controlled, fathers and mothers education continue to exert a significant positive effect. Children with better paid teachers and formal teachers receive higher grades in math. For language grades, girls, wealthier children, and children with better-educated parents are favored. Among school and 86 teacher characteristics, only having a local teacher shows a significant positive effect, and only in the teacher-only specification. Summary Aspirations and performance, the two measures of engagement that significantly condition subsequent enrollment, are themselves conditioned by socioeconomic status. For aspirations, being from a household with a better- educated father mattered in the full model. Female teachers seemed to help. Males and those with fewer siblings had higher aspirations. Being in primary schools with boarding students also showed a positive effect, though this category encompassed few students. For math grades, parent education level mattered. Having a better paid teacher and a formal teacher translated to better grades. For language grades, being female, wealthier, and from better-educated families helped. Overall, findings from part 2 of this report suggest that childrens attitudes and performance, along with family socioeconomic status, have a role to play in determining their educational trajectory. Findings also indicate that those factors at school that matter consistently for attitudes and performance, as for school continuation, are difficult to identify. Similarly, there is not a consistent story of school quality characteristics that directly help or hinder childrens persistence. 87 The pattern of gender effects is more surprising. While girls have somewhat lower aspirations and are somewhat less likely to still be enrolled in 2004, girls are not disadvantaged in the attainment change measure or in the achievement of nine years of education. Moreover, there is no gender difference in math performance or confidence, and girls fare better than boys in language grades, industriousness, and alienation. CONCLUSIONS This report has sought to provide, first, a broad-based description of conditions in rural Gansu schools, and second, an analysis of educational outcomes guided by informant reports of factors that might matter for children's enrollment and persistence in school. In this concluding section, we do not seek to characterize all of the findings, but instead highlight results of particular policy relevance. Finance Problems At the time of these surveys, the most pressing problem in rural education continued to be that of finance. As can be seen from the results presented in various sections of this report, schools are securing economic support from a variety of public and private sources, the fee system is complicated and not very unified across schools, family's costs for educating children are high, and a substantial proportion of children who leave school are reportedly doing so for 88 direct reasons of not being able to afford school costs. In our main analysis of enrollment, change in attainment, and nine-year achievement, wealth effects are significant across specifications that control for multiple other school, teacher, and family background factors. The government's capacity to reduce or eliminate costs as a barrier to educational access for the poorest families is crucial for the basic goal of extending an equal opportunity for education to children in these groups. This point is well-recognized by the government, which has renewed efforts to reduce or eliminate fees for basic education. For example, in 2005, the national and provincial governments in central and western China invested 7.2 billion RMB to expand a policy known as "two frees, one subsidy" (liang mian yi bu, ), which waives fees and provides boarding allowances for poor students during their nine years of compulsory education (Chang, 2006). According to a recent report in the China Development Brief (Chang, 2006), each beneficiary of liang mian yi bu receives 210-320 RMB a year, making school fees and textbooks virtually free, and some impoverished students in boarding schools also get living allowances. In 2006, Premier Wen Jiabao pledged that the government would eliminate all charges on rural students receiving a nine-year compulsory education before the end of 2007 (Peoples Daily, March 05, 2006). An 89 amendment to the compulsory education law, which came into effect September 1, 2006, commits to giving children in both cities and the countryside nine years of free compulsory education, though tuition charges will not be completely waived for a few years (Peoples Daily, June 30, 2006). The initiative to provide free, compulsory education was rolled out nationwide in 2007 (Wu, 2007). Central and local governments will have responsibility for expenditures, and local governments are to place expenditures for compulsory education in their budgets (Peoples Daily, June 30, 2006). According to a recent World Bank report, the cost sharing arrangement between the central government and the provincial and/or county governments is 80:20 in 12 western provinces and 60:40 in 10 central provinces (Wu, 2007). The transfer scheme began in spring 2006 in the western region and in spring 2007 in the central region. The coastal provinces will not be a part of the transfer scheme, although some poor areas are eligible to receive funds for reconstruction of dilapidated schools. The success of these efforts to reduce cost barriers to access is critical. Of course, finance problems have implications far beyond access. As we know from earlier work, there are dramatic differences across China in per pupil expenditures (Tsang and Ding, 2005), and these differences doubtless translate to considerable differences in the quality of education that children experience at 90 school. In particular, finance problems likely underlie both infrastructure problems and teacher morale issues--problems to which we turn below. Infrastructure Issues At the most basic level of infrastructure provision, two concerning findings emerged. First, in 2004, principals reported that about 29 percent of primary classrooms failed to meet official safety standards, as did about 16 percent of junior high school classrooms. Second, in 2000, about 15 percent of primary principals reported insufficient desks and chairs for all students, and in 2004, just under one in ten primary and junior high school principals reported this problem. Safety and a place to sit are very basic requirements. The trend for desks and chairs appears to be in a positive direction (probably aided by demographic trends), but addressing these basic deficiencies for the minority of schools still experiencing them will be an important step toward providing a basic educational environment for rural students. Another important feature of school infrastructure is whether or not schools have electricity. Here, our findings are encouraging: nearly all primary schools do, and all junior high schools do. Extending access to the last few primary schools that lack electricity is an important task. However, site visits to many rural Gansu schools suggest that the quality of lighting in many primary classrooms remains problematic, despite the fact that schools have electricity. 91 Schools may ask teachers to turn off lights to save on costs, or lighting may simply consist of a bulb that does not sufficiently light a full classroom. Dim classrooms may affect opportunities to learn, especially for children with vision problems. At a level up from these basic infrastructure issues, many primary schools still lack libraries: just under one-third in 2000 and just under one-fourth in 2004. Again, progress between the two years is encouraging, but many schools still lack access. Additionally, having a library may not directly translate into children accessing reading materials. Site visits to rural schools suggest that student access to libraries can be limited. Because books and reading materials are valuable, schools often keep the library, an extra office, locked up. Moreover, the keeper of the key to the library is often not present. Having accessible libraries may or may not be linkable to test scores, which is a high priority for schools serving communities where testing outcomes matter so much for children's subsequent economic circumstances. But access to reading materials beyond textbooks alone is crucial for children's broader development, which is a priority area under new educational policies in China. Some might argue that facilities like libraries are a luxury in settings as poor as are many communities in rural Gansu. Of course, a first priority should rightfully be placed on expanding basic access. But after this goal is achieved, it 92 will be important to begin thinking about how rural schools can address the reality that many families have few economic resources to provide enrichment materials or experiences to their children. Schools need to consider providing an environment where children who have limited means to purchase enrichment materials at home can access these materials at school, and libraries that are fully accessible to children could be an important element of such an environment. Finally, the strong presence of computers in rural schools in 2004--over a third of primary schools reported having computers, as did the vast majority of junior high schools--suggests that this resource is highly valued in rural schools. Like libraries, computers could be used to provide enrichment opportunities to children living with little. Certainly, the presence of computers does not necessarily imply that they are being used in a productive way, but lack of access is surely detrimental to children in a society where computer use is exploding. Moreover, the presence of computers raises significant policy questions: How are computer labs being financed? How are they being used, and how can we measure the quality of integration of computers into schools? What is the opportunity cost of computer labs, in terms of other needed facilities? These are important policy issues to be considered in further work on school infrastructure. 93 Teacher Morale Few would argue with the notion that teachers are the most critical resources in rural schools. Families and principals repeatedly raise the concern that rural areas struggle to retain good teachers. This concern is warranted, if attitudes expressed in our sample are any indication. Although the vast majority of teachers liked teaching and were satisfied with teaching as a job, many also said that they wanted to change jobs. Just under one in five primary school teachers wanted to change their profession in 2000; about one in four agreed with the sentiment of wishing to change careers in 2004. More than one-third of junior high school teachers wanted to change careers in 2004. It is likely that a major contributor to this problem is failure to pay teachers in a timely fashion. This finding is one of the most striking in our data. About 91 percent of primary school teachers were owed wages in 2000, as were 42 percent of primary school teachers and 47 percent of junior high school teachers in 2004. Just 13 percent of primary school teachers in 2000 said that their wages were usually or always paid on time, as did 57 percent of primary school teachers and 56 percent of junior high school teachers in 2004. These numbers, while improving quite dramatically between 2000 and 2004, suggest a persisting problem of major proportion in school finance. 94 Multivariate analyses with the 2000 primary teacher data linked wage delays to teacher dissatisfaction with their jobs (Sargent and Hannum, 2005). However, better paid teachers were not more satisfied in the 2000 survey. In 2004, a direct question about satisfaction with wages indicated that less than half of teachers were satisfied with this dimension of their work, even though wages rose substantially, at least in nominal terms, in the intervening four years. Dedicated, capable teachers are critical for rural educational improvement. Reducing wage delays will certainly not resolve the difficult problems of teacher morale and retention in remote rural schools. Even fewer teachers--just over ten percent of primary and junior high school teachers in 2004--were satisfied with benefits aside from salary. However, addressing wage arrears is a necessary, if not a sufficient, prerequisite for resolving teacher morale problems. Steps in this direction may be in progress, as the government is reportedly centralizing responsibility for teacher wages (Shanghai Daily, 2005). Teaching and Learning A third area of policy concern is classroom practices. The Chinese government has recently implemented major reforms that encourage progressive, student-centered pedagogy. Without placing a value judgment on whether this kind of pedagogy is preferable to more traditional practices, it remains of 95 significant policy interest whether teachers subscribe to the new practices, and whether they are implementing them in the classroom. Our results show that teachers report using a mix of traditional and new pedagogical approaches. Regarding traditional methods, the vast majority of teachers report sometimes or frequently using drilling, choral response, lecturing, and asking children to find answers in a textbook. Many teachers report sometimes using memorization, but few teachers report that they frequently use this method. Regarding new methods, the vast majority of teachers say that they sometimes or frequently ask open-ended questions, use classroom discussion, small-group activities, encourage individual student response, use inquiry learning, and ask students to participate in activities that connect theory to practice. Most teachers feel supported in the adoption of new methods: the vast majority agreed with a statement that the principal encourages them to use different teaching approaches, and a smaller majority agreed that the principal was a good source of information for teaching. Child reports of classroom occurrences also suggest a mix of styles. A substantial majority of students report being encouraged to ask questions, but also agree with a characterization of class as consisting of the teacher lecturing while students listen. A less substantial majority report that animated discussions usually occur in class. 96 Sargent's (2006) multivariate analysis of Gansu teacher and student reports on teaching styles suggest that both teacher and student reports of progressive methods in schools were greater in schools that had begun implementing the official curriculum reforms. Her classroom observation data also shows that progressive pedagogical practices are more likely to occur in schools that had implemented reforms. In summary, both traditional and progressive pedagogical approaches are occurring in rural Gansu schools, and progressive methods are linked directly to the new curriculum reforms. Are teaching styles associated with the new curriculum having an impact on outcomes such as achievement and subsequent enrollment? Are they having an impact on disparities in these outcomes? These will be key questions for the policy and research communities to address in the near future. Student Engagement with Education and Performance Our descriptive tables in part one lend a sense of the proportion of children who do not feel welcome at school. Our data suggests that roughly one in five children in both survey years often did not want to attend school. About fifteen to twenty percent did not feel happy at school. About one in four said that they were often lonely at school in 2000, as did just under one in five in 2004. Roughly one in three children in 2000, and over 45 percent in 2004, did not feel 97 that the teacher often paid attention to them in class. These numbers suggest that a substantial minority of students are disaffected. Of course, there is no real benchmark against which to judge these findings; this level of disaffection may be a reasonable level or even a very low level. As time passes, it will become increasingly important to know more about disaffection, and its implications for long-term educational outcomes. We have discussed the fact that school finance remained the critical barrier to access for China's poor rural children in 2004: schooling costs were a burden on parents, and many children and their families reported that costs are a consideration in school-leaving decisions. Yet, reports by village leaders, families, and children themselves indicate that children's disaffection and performance are also important for understanding educational decisions. These factors are likely to influence parental considerations as they face the mounting costs of higher levels of schooling. Children's engagement with schooling is influenced to some degree by conditions at school, as well as at home. While findings here do not suggest a consistent set of predictors of engagement and performance, other work with the same data (An, Sargent, and Hannum, 2007) has linked student perceptions of teacher behaviors in the classroom to student aspirations and alienation. 98 Given China's aggressive recent initiatives to address the cost barrier to education for the poorest families in China, the role of children's own willingness to stay in school will become increasingly important as an element of educational inequality. For this reason, understanding factors that are associated with performance and engagement will rise in policy significance in the future. Implications The Chinese government renewed commitments in recent years to support the development of poor rural areas, such as rural Gansu. The extension of meaningful educational opportunities to Gansu's young generation will be essential to the success of this broad endeavor. This report has suggested two kinds of barriers to educational improvement in Gansu. One kind of barrier is simply economic: many problems of infrastructure and economic deprivation in rural schools could, in theory, be quickly addressed, should funding be made available. This barrier is a major focus of China's current policy initiatives. Other barriers are not likely to be fully addressed with a simple infusion of economic resources. For example, low teacher morale, along with its implications for retaining good teachers in rural areas, is a difficult dilemma. If teacher labor markets develop further, and if teachers have increasing opportunities outside of education, this problem may intensify. 99 Another fundamental problem is that, to truly equalize the playing field with wealthier urban schools, rural schools may actually need to work harder to give children real opportunities for success. Our analyses in part two of this paper highlight the critical importance of family socio-economic status for many school outcomes, even at the stages of education agreed upon to be compulsory. The difficulties faced by many impoverished rural parents with little education as they seek to providing resources that their children need to flourish at school are immense. If a truly equal playing field in basic education is required, creative thinking will be needed--thinking about how schools might provide material resources, cultural materials, and enriching experiences that are tailored to the needs of the rural children and families that they serve. Finally, we have noted the problem of disaffection among a substantial minority of students, despite evidence that reforms designed to create more welcoming, nurturing environments in schools are, to some degree, being implemented. If the government is able to end poverty as a barrier to educational access, students' disaffection will become a more important source of educational decision-making. Work will be needed to understand how to sustain the engagement of impoverished rural children, once they are in the school system. 100 REFERENCES CITED Chang T. 2006. Rural education: Subsidies provide palliative, but not panacea. China Development Brief, Mon, 2006-10-09 11:11. (Electronic version posted to http://www.chinadevelopmentbrief.com/node/805 .) An, X., Hannum, E., & Sargent, T. 2007. Teaching Quality and Student Outcomes: Educational Engagement and Academic Achievement in Rural Northwest China. Forthcoming, China: an International Journal. China Internet Information Center. 2005a. An Introduction to Chinas Provinces, Municipalities, and Autonomous Regions: Beijing 2005 - the Year in Review. Beijing: State Council Information Office and the China International Publishing Group in Beijing, Retrieved 5/14, 2007 (http://www.china.org.cn/english/features/ProvinceView/163623.htm). ------.2005b. An Introduction to Chinas Provinces, Municipalities, and Autonomous Regions: Gansu 2005 - the Year in Review. Beijing: State Council Information Office and the China International Publishing Group in Beijing, Retrieved 5/14, 2007 (http://www.china.org.cn/english/features/ProvinceView/164352.htm). Gansu Bureau of Statistics. 2000. Gansu Survey of Children and Families Sampling Description. Lanzhou, China: Gansu Bureau of Statistics. Hannum, Emily, Jere Behrman, Meiyan Wang and Jihong Liu. Forthcoming. Education in the Reform Era. Chinas Economic Transition, edited by Loren Brandt and Thomas Rawski. 101 Ministry of Education. 2002. Suzhi Jiaoyu Guannian: Xuexi Tiyao (the Concept of Quality Education: Key Points for Study). Beijing, P. R. China: Ministry of Education. National Bureau of Statistics (China). 2005. China Statistics Yearbook 2005, Table 10-21: Per Capita Net Income of Rural Households by Region. Beijing: National Bureau of Statistics (China), Retrieved 5/28, 2006 (http://www.stats.gov.cn/tjsj/ndsj/2005/html/J1021e.xls). Peoples Daily. 2006a. China Adopts Amendment to Compulsory Education Law. (June 30, electronic version posted to http://english.people.com.cn/200606/30/eng20060630_278583.html.) Peoples Daily. 2006b. China Pledges Elimination of Rural Compulsory Education Charges in Two Years. (March 5, electronic version posted to http://english.peopledaily.com.cn/200603/05/print20060305_248042.html.) Sargent, Tanja. 2005. Institutionalizing Educational Ideologies: Organizational Control of Classroom Instruction in China. Philadelphia, PA: University of Pennsylvania Department of Sociology. Sargent, Tanja and Emily Hannum. 2005. Keeping Teachers Happy: Job Satisfaction among Primary School Teachers in Rural Northwest China. Comparative Education Review 49 (2): 173-204. Sargent, T. 2006. Institutionalizing Educational Ideologies: Organizational Control of Classroom Instruction in China. Unpublished manuscript. Shanghai Daily. (2005). Rural Teacher Salaries to be Centralized. (December 20, electronic version posted to http://china.org.cn/english/China/152548.htm.) Tsang, M. C. 2000. Education and National Development in China since 1949: Oscillating Policies and Enduring Dilemmas. China Review-An 102 Interdisciplinary Journal on Greater China, 579-618. (Electronic version posted to http://www.tc.columbia.edu/centers/coce/pdf_files/d1.pdf.) Tsang, M. C. and Y. Q. Ding. 2005. Resource Utilization and Disparities in Compulsory Education in China. China Review-an Interdisciplinary Journal on Greater China 5:1-31. United Nations Economic and Social Commission for Asia and the Pacific (UNESCAP). NA. Population and Family Planning in China by Province: Gansu Province. Bangkok, Thailand: United Nations Economic and Social Commission for Asia and the Pacific (UNESCAP), Retrieved 10/21, 2005 (http://www.unescap.org/esid/psis/population/database/chinadata/gansu. htm). United States Census Bureau. (2005). International Data Base. Retrieved 3/2, 2006 from http://www.census.gov/cgi-bin/ipc/idbagg . Wang, Jiayi. (2006). The Development of Basic Education in China: Achievements, Challenges and Policy Measures. Presentation at the China Institute for Public Affairs Conference, November 18, Baruch College, City University of New York. Wu, Kin Bing. 2007. Draft Report on Rural Education Finance and Service Delivery. Washington, D.C.: World Bank. Xinhuanet ( ). (2006). Wen Jiabao: In rural areas, widely implement compulsory education without miscellaneous fees ( ). Retrieved 3/5, 2006 from http://news.sina.com.cn/c/2006-03- 05/10018364484s.shtml. 103 Appendix 1. Descriptions of Selected Variables Employed in Multivariate Analysis Mean or Variables Comments Proportion SD N 2004 Change in Attainment Attained Grades 3.85 (0.98) 1693 Outcomes Enrollment Status (1=Yes) 0.87 (0.34) 1693 Nine Year Attainment (1=Yes) 0.53 (0.50) 629 2000 Math Grade (Average Score) 74.03 (14.74) 1677 Outcomes Chinese Grade (Average Score) 72.74 (13.29) 1671 Aspirations (Years) 13.08 (3.16) 1688 Academic Confidence Standardized Scale 0.00 (0.75) 1689 Industriousness Standardized Scale 0.00 (0.87) 1689 Alienation Standardized Scale -0.01 (0.75) 1680 2000 Gender (1=Female) 0.47 (0.50) 1693 Predictors Age (Years) 11.09 (1.15) 1693 Ln(Wealth) Ln(RMB) 9.21 (0.92) 1693 Fathers Education (Years) 6.99 (3.50) 1693 Mothers Education (Years) 4.15 (3.47) 1693 Siblings (Number) 1.31 (0.72) 1693 Teacher Gender (1=Female) 0.34 (0.48) 1693 Formal Teacher (1=Yes) 0.22 (0.41) 1693 Teacher Experience (Years) 14.90 (10.11) 1693 Teacher SHS Education (1=Yes) 0.62 (0.49) 1693 Teacher Tertiary Education (1=Yes) 0.14 (0.35) 1693 Teacher born in this Village (1=Yes) 0.40 (0.49) 1693 Ln(Teacher Monthly Income) Ln(RMB) 6.02 (0.75) 1693 School type=other (ref.=public) (1=Other) 0.09 (0.28) 1693 School has boarders (1=Has Boarders) 0.03 (0.17) 1693 Typical teacher absent 4+times per semester (1=Yes) 0.06 (0.24) 1693 Ln(Teacher av. monthly salary) Ln(RMB) 6.33 (0.26) 1693 School has daike teachers (1=Yes) 0.57 (0.50) 1693 School has minban teachers (1=Yes) 0.34 (0.47) 1693 Female principal (1=Yes) 0.02 (0.15) 1693 Principal experience (years teaching) (Years) 22.00 (8.34) 1693 Principal experience (years as principal) (Years) 9.17 (6.83) 1693 Highest Grade Attained (Grade) 3.35 (1.36) 1693 Note: Unit of analysis in this table and multivariate tables is the child. The numbers of valid observations (Ns) presented here include cases with valid observations on all predictors. Model Ns vary depending on the numbers of valid observations in the outcome variables (2004 and 2000 outcomes in this table.) 104 Appendix 2. Components of Engagement Measures (2000) Concept Component Variable(s) N % Composite Measure Mean SD Aspirations (highest level of education you want to complete) Approximate Years 13.08 3.17 primary school 88 4.75 middle school 176 9.5 high school 317 17.11 junior trade school 180 9.71 senior trade school 193 10.42 university or above 899 48.52 1,853 100 Academic (in your opinion, are you a good student?) Standardized Scale 0.01 0.75 Confidence no 230 12.41 so-so 839 45.28 yes 784 42.31 1,853 100 (rate your math ability) very poor 151 8.15 below average 160 8.63 average 745 40.21 above average 433 23.37 excellent 364 19.64 1,853 100 (rate your language ability) very poor 141 7.61 below average 128 6.91 average 888 47.95 above average 383 20.68 excellent 312 16.85 1,852 100 Effort (work hard at math?) Standardized Scale 0.00 0.86 do not study hard 126 6.8 occasionally study hard 634 34.23 generally study hard 1,092 58.96 1,852 100 (work hard at Chinese?) do not study hard 137 7.4 occasionally study hard 624 33.69 generally study hard 1,091 58.91 1,852 100 Standardized Scale 0.00 0.75 Alienation (do not want to attend school most of the time) completely disagree 897 48.36 somewhat disagree 552 29.76 somewhat agree 201 10.84 completely agree 205 11.05 1,855 100 (often feel bored at school) completely disagree 515 27.81 somewhat disagree 842 45.46 somewhat agree 315 17.01 completely agree 180 9.72 1,852 100 (often feel lonely at school) completely disagree 597 32.39 somewhat disagree 747 40.53 somewhat agree 300 16.28 completely agree 199 10.8 1,843 100 105 Working Papers Human Development Sector Unit East Asia and Pacific Region The World Bank No. 2004-1: Disability Issues in East Asia: Review and Ways Forward, May 2004. (Report No. 29299). No. 2004-2: Cambodia Skills and Growth, May 2004. (Report No. 29929). No. 2004-3: Evaluating the Performance of SGP and SIGP: A Review of the Existing Literature and Beyond, May 2004. No. 2004-4: Vietnam Reading and Mathematics Assessment Study ­ Volume 1, August 2004. No. 2004-5: Addressing Inequity in Access to Health Care in Urban China: A Review of Health Care Financing Reform Experiments, July 2003. (Report No. 31340). No. 2004-6: Strengthening World Bank Support for Quality Assurance and Accreditation in Higher Education in East Asia and the Pacific, August 2004. No. 2004-8: Financing Health Care for Poor Filipinos, July 2003. (Report No. 31341). No. 2004-8: Health Care Financing for the Poor in Vietnam, March 2003. (Report No. 31004). No. 2004-9: Pro-Poor Health Financing Schemes in Thailand: A Review of Country Experience, June 2003. (Report No. 31009). No. 2004-10: Health Financing for the Poor in Indonesia No. 2004-11: Making Injections Safe in China: How Much Will it Cost and Who Will Pay? April 2004.. (Report No. 31444). No. 2004-13: Private Sector Involvement in Education: A Review of World Bank Activities in East Asia and Pacific, 1996-2002, July 2004. (Report No. 33527). No. 2004-14: Current Trends, Future Opportunities Human Development Outcomes in the Pacific Islands, November 2004. No. 2005-1: China: Pension Liabilities and Reform Options for Old Age Insurance, May 2005. (Report No. 33116). 106 No. 2005-2: Cambodia: Balancing the Books: Household Financing of Basic Education in Cambodia, January 2005. (Report No. 33810). No. 2005-3: Vietnam Learning to Teach in a Knowledge Society. No. 2005-4: Reforming Health Social Security, June 2005. (Report No. 36394). No. 2006-1: Mongolia: Assessment of the Child Money Program and Properties of Its Targeting Methodology, April 2006. (Report No. 36018). No. 2006-2: Indonesia: Quality Assurance and Accreditation in Higher Education in East Asia and the Pacific. No. 2006-2: Indonesia: Early Childhood Education and Development ­ An Investment for a Better Life, June 2006. (Report No. 37531). No. 2007-1: The Mongolian Health System at a Crossroads: An Incomplete Transition to a Post-Semashko Model, January 2007. (Report No. 39014). No. 2007-2: Critical Discussions on Education: Meeting Emerging Challenges in East Asia, January 2007. 107