Who Teaches in Rural Schools in Underdeveloped Areas? An Investigation Based on a Survey of 5,554 Teachers from 117 Towns in H Province in Wuling Mountains Zone, China Tianjiao Zhang, 1 Weiping Wang, 2 Li Yi 3 1. Party School of Shanghai Xuhui District Committee of C. P. C., Shanghai 200030, China 2. Jingxiangyi Education Institute, Hangzhou Normal University, Hangzhou 310000, Zhejiang, China 3. Hunan First Normal University, Changsha 410000, Hunan, China Abstract. Teacher shortage is a major hindrance to China’s rural education growth in underdeveloped areas, as well as one of the main causes of educational injustice. We conducted a survey of 5,554 teachers from 117 towns in H province in the Wuling Mountains Zone to investigate the issue of rural school teacher supply. From geographical, emotional, and institutional perspectives, we used a polynomial logit model to examine the validity of the “hometown effects” hypothesis. The findings showed that hometown effects exist in China in all three dimensions. The institutional hometown effects are the most pronounced; when compared to open recruitment, teachers sourced through teacher supply augmentation programs (such as the Secondary Normal Graduates Program, Special Position Program, and Targeted Position Program) are more likely to teach in rural schools, particularly more disadvantaged village primary schools or teaching sites. China’s policy of increasing teacher supply has had a considerable positive influence on rural school staffing. Students from rural areas make better teacher candidates; feelings for hometowns should be encouraged among normal school or university students in pre-service education; and the implementation of teacher support policies should be emphasized to retain rural teachers and improve their teaching quality. Best Evidence in Chinese Education 2022; 11(1):1429-1448. Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No1, 2022 1430 Doi: 10.15354/bece.22.or060. How to Cite: Zhang, T., Wang, W., & Yi, L. (2022). Who teaches in rural schools in underdeveloped areas? An investigation based on a survey of 5,554 teachers from 117 towns in h province in Wuling mountains zone, China. Best Evidence in Chinese Education, 11(1):1429-1448. Keywords: Rural Teachers, Hometown Effects, Teacher Supply Augmentation Policy, Teacher Shortage About the Authors: Weiping Wang, Hangzhou Normal University, Hangzhou 310000, China. E-mail: weipingzhang@hznu.edu.cn Li Yi, Hunan First Normal University, Changsha 410000, China. E-mail: lily_yi2011@163.com Correspondence to: Tianjiao Zhang, Party School of Shanghai Xuhui District Committee of C. P. C., Shanghai 200030, China. E-mail: tjzhang17@fudan.edu.cn Funding: This study is supported by 1) the project of Research on the Strategies for Improving Basic Education After-school Service Quality in the Context of Double Reduction Policy, sponsored by Party School of Shanghai Committee of C. P. C. (School of Administration) (2021SHB015); 2) the project of Research on the Strategies for Enhancing the IT-supported Instruction Skills of Local Public-funded Normal School Students in the Context of New Normal Education, sponsored by Hunan Provincial Department of Education (20C0411). Conflict of Interests: None. © 2022 Insights Publisher. All rights reserved. Creative Commons Non Commercial CC BY-NC: This article is distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 License (http://www.creativecommons.org/licenses/by-nc/4.0/) which permits non-commercial use, reproduction and distribution of the work without further permission provided the original work is attributed by the Insights Publisher. mailto:weipingzhang@hznu.edu.cn mailto:lily_yi2011@163.com mailto:tjzhang17@fudan.edu.cn Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1431 Introduction As per the teacher-to-class ratios in compulsory education schools released in the Re- port on China’s Rural Education Development 2019, China’s rural schools are experi- encing a serious teacher shortage (Wu & Qin, 2019). South Reviews (2021) published a report entitled “China Needs More Rural Teachers” in 2021, which emphasized the critical issue of rural teacher shortage. It was widely shared on social media sites like Phoenix and Tencent, and it drew a lot of attention and discussion. Rural teachers face a variety of issues, including low pay, insufficient social support, limited professional development, a lack of incentive mechanisms, a higher number of problem students, difficult class management, low social status, poor living conditions, potential cultural conflicts, and so on (Neeshatlen, 1985; Yarrow, 1999). These disadvantages in rural school working and living situations make teacher recruitment and retention challeng- ing in the growth of rural education in impoverished areas around the world. As a result, the question of “who teaches in rural schools” has become a hot topic in academia. Since its foundation, the People’s Republic of China has implemented a num- ber of teacher supply augmentation measures to alleviate the rural teacher shortage. The central government began implementing the Secondary Normal Graduates Program in several provinces in the early 1980s, in which graduates from junior colleges and sec- ondary normal schools were selected and assigned to teach in local rural schools. For the next two decades, and until the early 21st century, this initiative was a major source of rural teachers. The influence of urbanization on the countryside as well as the structural changes in rural schools has made it more difficult to recruit and retain rural teachers, as well as to improve the quality of rural education at the turn of the 21st century. The Ministries of Education and the Finance collaborated in 2006 to develop the Special Position Program, which used central government financing to hire college graduates to work in rural compulsory education institutions (Ministry of Education, 2006). The ini- tiative began in western China and was eventually expanded to the central and eastern parts of the country. In the meantime, the State Council issued the Measures for the Im- plementation of Free Education for Students of Normal Universities Affiliated to the Ministry of Education (Ministry of Education, 2007), requiring the six normal universi- ties affiliated to the Ministry of Education to provide free education to their students. In 2013 and 2015, Jiangxi Normal University and Fujian Normal University, both joint establishments of provincial governments and the Ministry of Education, joined the Free Education Program. The Measures to Implement Public-funded Education for Stu- dents of Normal Universities Affiliated to the Ministry of Education were published by the State Council in 2018, and the “Free Education Program” was renamed the “Pub- lic-funded Teacher Education Program” (also known as the Targeted Position Pro- gram). The Special Position Program and the Public-funded Teacher Education Pro- gram have gained widespread acclaim in society for significantly reducing staffing shortages in rural parts of western China, and also for helping to attract and retain rural teachers and improve compulsory education quality (Pu et al., 2019). Preferential poli- cies such as the Particular Position Program and the Public-funded Teacher Education Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1432 Program were first adopted in the western regions due to regional special demands and the state’s limited financial capability. However, there are impoverished and underde- veloped areas in the central provinces that confront similar issues, such as local budget deficits, poverty-stricken rural areas, stymied rural school growth, and a rural teacher shortage. Despite little funding from the central government, several provinces strive to address staffing issues in rural schools by increasing teacher labor supply, improving teacher quality, and reducing structural human resource imbalances. Hunan Province in central China, for example, drafted the Opinions on Strengthening Primary and Sec- ondary Teacher Education in 2005 and pioneered the Targeted Position Program for village primary teacher education in China. The provincial government and lower levels of government contribute to the funding of this initiative. The provincial department of education uses the joint money to commission local normal institutions and universities to train suitable teachers for specific rural schools. As a result, open recruitment, the Secondary Normal Graduates Program, the Targeted Position Program, the Special Position Program, temporary teachers, and other sources of rural teacher supply exist in present China. This study applied a poly- nomial logit model to assess the validity of the “hometown impacts” hypothesis in the geographical, emotional, and institutional dimensions, based on the findings of a survey of 5,554 teachers from 117 towns in H Province in the Wuling Mountain Zone. A Literature Review and Research Hypotheses Some studies claimed that “hometown effects” (also known as “draw of home”) exist in the teacher labor market, which are not common in other professions (Reininger, 2012; Cannata, 2010; Goff & Bruecker, 2017; Engel et al. 2014). According to the “hometown effects” hypothesis, teachers prefer to teach in the place where they grew up or in an area that is similar to their home environment. In their research of teacher distribution at New York State, Boyd et al. (2005; 2013) find that 61% of teachers start their careers in schools within 15 miles of their homes, and more than 80% in schools within 40 miles. In a survey conducted in Gansu Province, China, Wei (2016) discov- ered that 93% of teachers begin their careers in native countries, 51% in native towns, and 23% in native villages. Furthermore, Wei et al. (2020) find that rural teachers who start their careers in the same region as their birthplace have much lower employment mobility than those who start their careers in a different region. A survey of 40 rural schools in Guangdong and other provinces found that localization is an important geo- graphical aspect of rural teacher employment, with many teachers working in their hometowns (Liu, 2019). According to Monk (2007), people who are interested in rural education and have strong feelings about rural living will opt to teach in rural areas, despite the fact that interests and feelings are vulnerable to spatial linkages. As a result, hometown effects, which are a distinctive feature of the teacher la- bor market in China and other countries, are universal. Governments all over the world are forming rural teacher education and selection programs based on the localization tendency of the teacher labor market, such as the Grow-your-Own Program in the Unit- ed States, the Indigenous Teacher Education Program in Australia, and the Public- funded Teacher Education Program in China, to name a few. Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1433 Localized selection of teachers in rural or distant locations is the most success- ful strategy to attract and retain rural teachers, according to academia and governments. Teachers with local “registered residences” (also referred to as “huji” or “huko” in Chi- na) are more likely to work in local town schools or in more remote rural schools and teaching sites in China, according to a well-known occurrence. There are discrepancies in how Chinese researchers and their international colleagues define the variable “re- turning to hometown”. When assessing the variable “returning to hometown,” foreign scholars typically mention the distance between the school and the home or calculate the distance using the school’s and home’s zip codes (Boyd et al, 2013). In China, “re- turning to hometown” refers to returning to work in the area of one’s registered resi- dence or birthplace; Scholars commonly utilize the county-level division of administra- tive regions (in China, a county is a lower-level administrative unit than a city) as a cri- terion to separate natives and foreigners, i.e., people from outside a certain county (Liu, 2019). As a result, we suggest the first hypothesis (H1): Teachers with a local registered residency are more likely to teach in rural schools, implying that rural teachers have geographical hometown effects. Researchers also pointed out that many front-line teachers in rural areas spent part of their childhood or schooling in the area (Monk, 2007). According to surveys, 72.3% of teachers who have worked in rural schools for a long time were raised in rural areas, and 60% acquired their education in rural schools. In contrast, Chinese research focuses on geographic hometown preferences associated with registered residence, but there is little talk about emotional hometown impacts deriving from teachers’ early life or education experiences in local places. Our second hypothesis (H2) is that: Based on geographical hometown effects related to the registered household, there is an emotion- al hometown preference based on teachers’ rural life experience, i.e., teachers who grew up in rural areas or attended rural schools are more likely to teach in rural schools. China’s education administration system differs from that of various Western countries. In China, the county-level government is primarily responsible for funding primary and secondary schools in towns and villages. As a result, the county-level gov- ernment is also responsible for teacher recruitment and appointment based on the staff- ing needs reported by the towns and villages under its jurisdiction. Government policies have a greater impact on teachers recruited through teacher supply augmentation initia- tives. The Targeted Position Program in H Province requires program participants to agree to teach in the areas of their registered residency for 5-7 years after graduation, and the county bureau of education sets their compensation and other perks, such as Bianzhi membership (a system of officially budgeted posts in China). The county gov- ernment may take into account teachers’ preferred working locations as well as their registered residences when making teacher appointments. Teachers, on the other hand, are typically assigned to areas where teaching staff is in limited supply throughout the year (Wei, 2016). As a result, while examining teachers’ hometown preferences within a county region, we must consider this institutional aspect. The third hypothesis (H3) we propose is that: after accounting for geographical and emotional hometown effects, institutional hometown effects exist, as evidenced by the fact that students educated through teacher supply augmentation programs are more likely to teach in rural schools. Research Design Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1434 Sample Summary The Wuling Mountains Zone is a theoretical economic notion employed in the state’s efforts to alleviate poverty, and it spans most of the Wuling Mountains. The Wuling Mountains Zone is economically poor and has a large underprivileged population due to its physical location. It is one of the most important beneficiaries in the country’s pov- erty alleviation program. The participants in this study are rural teachers from 117 towns in seven counties in H province in central China’s Wuling Mountains Zone. The following factors influenced the sample selection: (i) According to national criteria, the seven counties studied are destitute. The Wuling Mountains Zone is one of China’s 14 most impoverished areas, with insufficient basic public services, severely outdated in- frastructure in education, culture, health, and sports, and per capita education and health spending of only 51% of the national average. Recruiting and keeping talented rural teachers in this area is a common problem in socioeconomically underdeveloped areas. (ii) H Province was one of China’s first provinces to adopt the Targeted Position Pro- gram. In 2006, H Province launched a special program to recruit junior secondary school graduates for 6-year pre-service education in order to provide targeted training of multi-disciplinary primary school teachers for towns and villages, one year ahead of the Ministry of Education’s and its six affiliated normal universities’ Free Teacher Ed- ucation Program (Hunan First Normal University, 2017). H Province has acquired a large number of teachers as a result of various sorts of teacher supply augmentation schemes during the last 10 or more years, ensuring the sample’s representativeness. The Wuling Mountains Zone encompasses the counties and cities of H Prov- ince’s western, northwestern, and central regions, with the western half of the province having the most counties. We chose four counties at random from the province’s west, two from the northwest, and one from the central region based on geographical distribu- tion, and conducted a two-month field study in each of the seven counties in 2020. The seven counties have a total of 117 towns and nearly 15,000 primary and secondary school teachers, as shown in Table 1. A total of 5,554 questionnaires were obtained for this study, all of which were legitimate. Variables and Measurement The Dependent Variable The school type, which is a categorical variable, is the study’s dependent variable. County-level schools (referring to primary and secondary schools in the county’s met- ropolitan areas), town-level nine- and 12-year schools, town junior secondary schools, town primary schools, village primary schools, and teaching sites are all covered. This paper groups town-level nine- and 12-year schools, town junior secondary schools, and town primary schools into one category in the regression model for statistical conven- ience, making school type a categorical variable with three categories: county-level schools, town primary and secondary schools, and village primary schools and teaching sites, which are coded as 0, 1, and 2, respectively. Independent Variables Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1435 Table 1. Sampled Counties. Sampled Counties Numbers of Towns Under Jurisdiction Numbers of Teachers Retrieved Questionnaires Valid Questionnaires Validity Rate A 12 1,641 718 718 100% B 19 2,244 1,005 1,005 100% C 17 3,417 1,358 1,358 100% D 26 3,405 1,832 1,832 100% E 20 5,730 596 596 100% F 23 n/a 0 0 n/a Total 117 5554 5554 100% Notes: 1. Respondents are grades 1-9 teachers (no secondary vocational and high school teachers surveyed); 2. Due to some emergencies, the investigation team failed to distribute the questionnaires in F County, and only conducted interviews in this county. The teacher’s registered residence, rural living and education experience, and teacher supply sources are the three independent factors in this study. For the teacher’s regis- tered residence, the inquiry provides five options: the town or village where the school is located; other towns or villages within the local county; the local county’s urban re- gion; other counties within the province; and other provinces. There is just one option for this question. We combine “other counties within the province” and “other provinc- es” into one category: registered residence outside the local county, and come up with a variable with four categories of registered residence, with 1, 2, 3, and 4 representing registered residence in the town or village where the school is located, other towns or villages within the local county, the local county’s urban area, and outside the local county, respectively. Furthermore, we collect information about teachers’ birthplaces, determine whether the teacher is a local based on their response to the question “Whether your birthplace is the same as the location of the current school” in the ques- tionnaire, and create a dummy variable for teachers’ birthplaces, with 0 indicating “no” and 1 indicating “yes”. In addition, with the variable of five or more years of early rural life or educa- tion experience, the questionnaire evaluates emotional hometown impacts. Teachers’ rural experience or schooling experience is treated as a dummy variable, with 0 indicat- ing “no” and 1 indicating “yes.” The source of teacher supply is a categorical variable, with 0, 1, 2, 3, 4, 5 representing open recruitment, temporary teachers, the Targeted Position Program, the Special Position Program, the Secondary Normal Graduates Pro- gram, and other supply sources (such as inter-school teacher rotation), respectively. Control Variables Individual characteristics of teachers, such as gender, age, ethnicity, and proximity to the school, serve as control variables in this study. Gender is a dummy variable (0 = female, 1 = male); age is a continuous variable; proximity to the school is a dummy variable (0 = no, 1 = yes); and ethnicity is a categorical variable (0 = Han nationality, 1 = Miao nationality, 2 = Tujia nationality, 3 = other ethnic minorities). Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1436 The Model and Analytical Approach This study employs polynomial logit regression to analyze the dependent variable “school type,” which is a multi-category variable. The model is as follows:     ln ln iij i j ib i P y j x x P y b x                where b signifies the baseline group and j specifies the number of school types in the categorical variable, i.e., j =1, 2, 3... The anticipated probability of each category can be computed by solving the equation below (Yang & Zhang, 2020):      1 exp = = exp i j ij i J i mm x P y j x x        Thus, we obtain the frequency ratio of each category relative to the baseline group. In the examination of regression findings, the odds ratio is utilized to calculate and evaluate the model’s regression coefficients. To further address the research issue, the study employs stepwise regression to examine the hometown effects of rural Chi- na’s teachers. Research Results Sample Descriptions Table 2 provides a summary of the subjects. (i) Female teachers account for 16.46% more than their male counterparts, which is in line with China’s current gender structure in elementary and secondary schools. (ii) The proportions of young and older teachers are greater than those of middle-aged teachers, according to the age distribution pattern. (iii) Currently, 72.04% live within walking distance of a school. (iv) The ethnicity of the subjects in the sample is quite diverse. More than half of the subjects are Miao and Tujia ethnic groups, which corresponds to the existing population situation in the Wuling Mountains Zone). (v) In terms of professional titles, the teaching staff is primar- ily made up of first- and second-rank teachers, with senior teachers accounting for ap- proximately 11.76%. (vi) Pre-service education was provided to over 60% of teachers through teacher supply augmentation initiatives. (vii) The vast majority of study partic- ipants were students from rural elementary and secondary schools. Therefore, sample distribution is consistent with general features of participants, indicating that the sample is ideal and representative. Localization of Rural School Staffing Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1437 Table 2. A Sample Summary. Variable Sample Classification Frequency Percentage (%) Gender Female 3,234 58.23 Male 2,320 41.77 Age 18-27 1,485 26.74 28-38 1,295 23.32 39-55 2,311 41.61 56-77 435 7.83 others 28 0.50 Living around School Locations or Not Yes 1,553 27.96 No 4,001 72.04 Nationality Han 2,066 37.39 Miao 1,763 31.91 Tujia 1,567 28.36 Other ethnic minorities 129 8.19 Professional Title The third-rank teacher and others 1,240 22.32 The second-rank teacher 1,503 27.06 The first-rank teacher 2,158 38.85 Senior teacher 653 11.76 Sources of Teacher Supply Open recruitment 1,158 20.85 Targeted Position Program 1,078 19.41 Special Position Program 1,076 19.37 Secondary Normal Graduates Program 956 17.21 Temporary teachers 470 8.46 Others 816 14.69 School Type Village primary schools and teaching sites 949 17.09 Town-level 9- and 12- year schools 667 12.01 Town primary and secondary schools 2,644 47.61 County-level schools 1,294 23.30 Notes: 1. Percentiles 1/4, 1/2, and 3/4 are employed to divide age groups 2. The Targeted Position Program includes the Ministry of Education’s public-funded teacher education Program and local level targeted position programs. 3. The Special Position Program includes nation level and local level programs. The data in Table 3 show that rural teachers in the county’s territory have a high rate of localized employment. More than half of the teachers working in rural primary schools or teaching sites (475/949 = 50.5%) have a registered residence in the local towns or villages. Teachers, who live in other towns or villages in the county, or beyond the county, are less likely to work at rural primary schools or teaching sites; instead, they prefer to teach in town primary and secondary schools, or at the county level. Further- more, teachers who live in the county’s urban areas are the least likely to work in vil- lage primary schools or teaching sites. A close link exists between rural teachers’ regis- tered residence and their school category (p < 0.001), indicating that teachers with local registered residency are more likely to accept work in rural schools. As a result, H1 has been certified. Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1438 Table 3. School Types of Teachers with Different Registered Residence. Variable Names # of Teachers from the Local Town and Village # of Teachers From Other Towns or Villages of the Local County. # of Teaches From Urban Areas of the County # of Teachers from Outside the County Total Village Primary Schools and Teaching Sites 475 193 112 169 949 Town Primary and Secondary Schools 1460 856 447 548 3,311 County-Level Schools 225 394 548 127 1,294 Total 2,160 1,443 1,107 844 5,554 chi2 688.2650 P-value 0.000 (***) Note: Levels of significance: ***p < 0.001, **p < 0.01, *p < 0.05, +p < 0.1 Table 4. School Types of Teachers from Different Teacher Supply Sources. Variable Open Recruitment Temporary Teachers Targeted Position Program Special Position Program Normal College or University Student Program Other Sources Total Village Primary Schools and Teaching Sites 178 158 155 221 133 104 949 Town Primary and Secondary Schools 595 287 704 719 593 413 3,311 County-Level Schools 385 25 219 136 230 299 1,294 Total 1,158 470 1,078 1,076 956 816 5,554 chi2 378.1099 P-value 0.000 Note: levels of significance: ***p < 0.001, **p < 0.01, *p < 0.05, +p < 0.1 The Influence of Teacher Supply Augmentation Pro- grams on Rural School Staffing As mentioned before in this study, due to China’s unique educational administration system, rural teachers may experience local effects as a result of institutional arrange- Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1439 ments. So, how do different teacher supply programs affect teachers’ hometown prefer- ences? There are substantial disparities in school categories among teachers produced by different programs (p < 0.001), according to data from Table 4. Teachers in village primary schools or teaching sites are mostly from the Special Position Program (221 teachers), which is part of the policy to increase teacher supply. The majority of Target- ed Position Program and Special Position Program teachers (704 and 719, respectively) chose to work in town primary and secondary schools, which offer superior working conditions, accounting for about half of the entire staff. The study results from 117 towns in H Province suggest that educational poli- cies may have a significant impact on the hometown effects of rural China teachers. Due to the fact that the Special Position Program recruits teachers based on the staffing needs of local schools as opposed to the distribution plan of county-level educational authorities, the influence of teachers’ independent decisions on hometown effects can- not be neglected. Therefore, additional research is required to explore institutional hometown impacts. The Effects of Geographical Influence, Emotional Preference, and Policy Guidance on Rural Teachers’ Employment The results of a regression study of the impacts of geographical influence, emotional preference, and policy advice on teachers’ choice of school types are presented in Table 5. The registered residence of teachers is used to assess geographical hometown impacts, whereas their early rural life or schooling experiences are utilized to measure emotional hometown effects. The scope of Model 1 is limited to evaluating the effect of teachers’ registered residence on their employment and school kinds. In Model 2, the early rural life or education experience of teachers is included in Model 1. The source of teacher supply is added as a variable in Model 3. Due to the inability to conduct direct compari- sons between the Mlogit model’s regression coefficients, we must convert them into odds ratios in order to evaluate and present the results of the analysis (Hong, 2015). The Relationship between Geographical Hometown Ef- fects and Teachers’ Choice of School Types Using teachers employed by county-level schools as the baseline group, Model 1 demonstrated that after controlling for variables such as gender, age, ethnicity, and whether or not they live around the school location or not, teachers from the local town or village are more likely to work in village primary schools or teaching sites than teachers with registered residence outside the county, with a 2.147 times higher odds ratio (e 1.147 -1 ), which is statistically significant (p < 0.001) than teachers with registered residence within the county. In a related manner, teachers who were native to the town in which the school is located were 1.254% more likely to teach in the town’s primary and secondary schools than teachers whose registered residences were outside of the county (e 1.006 -1, p < 0.001). In a fundamental sense, this is in line with the statistical descriptions made in the past. Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1440 Table 5. Results of the Regression Analysis of Hometown Effects on Rural Teachers’ Employment. (1) (2) (3) (4) (5) (6) Model 1 Model 2 Model 3 Variable School types (County-level schools as the baseline group) I II I II I II Gender 0.747*** 0.813*** 0.747*** 0.811*** 0.808*** 0.843*** (0.108) (0.085) (0.109) (0.085) (0.111) (0.086) Age -0.035*** -0.039*** -0.036*** -0.041*** -0.046*** -0.042*** (0.005) (0.004) (0.005) (0.004) (0.006) (0.005) Living around the School location or not -0.838*** -0.654*** -0.843*** -0.700*** -0.836*** -0.711*** (0.115) (0.092) (0.119) (0.095) (0.121) (0.096) Nationality (Han nationality as the baseline group) Miao -0.723*** -0.508*** -0.727*** -0.510*** -0.748*** -0.539*** (0.112) (0.088) (0.112) (0.088) (0.115) (0.090) Tujia -0.506*** -0.254** -0.512*** -0.261** -0.552*** -0.301** (0.115) (0.092) (0.115) (0.092) (0.119) (0.094) Other ethnic minorities -1.582*** -0.516* -1.589*** -0.521* -1.631*** -0.564* (0.383) (0.235) (0.383) (0.235) (0.389) (0.240) Registered Residence (Registered residence outside the local county as the baseline group) Local town Or village 1.147*** 1.006*** 1.149*** 0.999*** 1.137*** 0.886*** (0.163) (0.137) (0.164) (0.137) (0.172) (0.142) Other towns or Villages in the Local county -0.610*** -0.349** -0.610*** -0.363** -0.482** -0.332** (0.156) (0.124) (0.156) (0.124) (0.162) (0.129) Urban areas of the Local county -1.318*** -1.166*** -1.318*** -1.156*** -1.137*** -1.108*** (0.168) (0.127) (0.168) (0.127) (0.173) (0.132) Early rural life or Education experience 0.035 0.230* 0.168 0.311** (0.135) (0.108) (0.138) (0.111) Teacher Supply Sources (Open recruitment as the baseline group) Temporary teachers 2.656*** 2.039*** (0.264) (0.240) Targeted Position Program 0.308* 0.646*** (0.154) (0.114) Special Position Program 1.032*** 0.988*** (0.155) (0.126) Secondary Normal Graduates Program 0.363* 0.604*** (0.168) (0.123) Other sources -0.105 -0.015 (0.162) (0.116) Constant terms 1.918*** 2.971*** 1.915*** 2.917*** 1.649*** 2.434*** (0.195) (0.162) (0.197) (0.164) -0.242 -0.194 Pseudo R2 0.092 0.093 0.118 Log likelihood -4746.895 -4743.728 -4610.350 AIC 1.739 1.740 1.697 BIC -37564.929 -3754543 -37657.16 Sample size 5,500 5,500 5,500 5,500 5,500 5,500 Notes: I: Village Primary Schools & Teaching Sites; II: Town Primary & Secondary Schools. 1. Values in parentheses are stand- ard errors. 2. Levels of significance: ***p<0.001, **p<0.01, *p<0.05, +p<0.1 Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1441 In the meantime, we observed that teachers with registered residences in other towns or villages within the county do not demonstrate a positive hometown preference for the local rural schools compared to teachers with registered residences outside the county, and the likelihood is even significantly reduced (the coefficients in Model 1 are -0.610, and -0.349, respectively, and the p-values are both less than 0.001). As a result, we can conclude that the geographical characteristic of hometown effects is not posi- tively correlated with teachers’ registered residence in the local county (as was assumed in prior research), but rather with teachers’ registered residence in the local town, a smaller administrative area. This cautions us that, in the context of China, a more com- prehensive examination of the consequences of geographical hometown is required. The survey results from Hubei, Jiangsu, and Henan confirmed that teachers from local towns had the lowest turnover rates (Liu, 2019). The Relation between Emotional Hometown Effects and Teachers’ Choice of School Types The coefficient of determination of Model 2 is improved, and the model’s significance is maintained, when the variable of early rural life or teacher education experience is introduced. Teachers with five or more years of early rural life or education experience are more likely to work in village primary schools or teaching sites (with a positive co- efficient) than in county-level schools after controlling for variables such as gender, age, ethnicity, registered residence, and whether or not they live near the school location, though the difference is not statistically significant. Nonetheless, compared to those with no such experience, teachers with five or more years of early rural life or education experience had a 26% greater odd ratio of working in town primary and secondary schools (e 0.230 -1), which is statistically significant (p < 0.01). The findings of Stiut (2010) and Boylan (2010), who both believed that teachers’ prior rural living or job experience had a beneficial effect on their extended stay at rural schools. As a result, H2 is only partially accepted. The Relationship between Policy-guided Hometown Ef- fects and Teachers’ Choice of School Types In order to investigate the impact that the teacher supply augmentation programs have on hometown effects, Model 3 contains the variable of teacher supply sources as an independent variable. As a consequence of this, the model’s coefficient of determina- tion rose from 0.093 to 0.118, which indicates that the model’s explanatory power has improved and that this improvement is statistically significant. Taking teachers em- ployed at county-level schools as the baseline group, Model 3 demonstrates that after controlling for the variables of gender, age, ethnicity, registered address, and whether or not the teacher lives near the school location, teachers produced through the Targeted Position Program by the Ministry of Education and local governments have a 36.10% higher odds ratio (e 0.308 -1) of working in village primary schools and teaching sites, and a 90.81% higher odds ratio (e 0.646 -1) of working in town primary and secondary schools, compared to those recruited through open recruitment, and these differences are statisti- cally significant (p < 0.1 and p < 0.001, respectively). Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1442 There is no question that the hometown effects associated with the Targeted Position Program are partially attributable to the rigorous and forceful implementation of pertinent policies, which state that graduates who benefit from the program must re- turn to work in their original region of residence. Moreover, the majority of students trained under local targeted position programs come from towns or villages. This sub- stantially increases the likelihood of teachers returning to the town or village. Then, what about graduates of the Specific Position Program, which does not restrict the sup- ply of students but instead creates special posts based on the needs of local schools? Under the same control conditions, in comparison to the baseline group (con- sisting teachers from county-level schools), teachers recruited through the Special Posi- tion Program are much more likely to work in village primary schools and town schools than those recruited through open recruitment. The odds ratios are 1.801 times (e 1.032 -1) and 1.689 times (e 0.988 -1) higher, respectively, which is statistically significant at the 0.001 level. Similarly, teachers recruited through the Secondary Normal Graduates Program in the early stages are far more likely to work in rural primary schools and urban schools than those recruited through open recruitment (with positive coefficients, signif- icant at levels of 0.05 and 0.001, respectively). Since the 1980s, the implementation of teacher supply augmentation programs in China has significantly reduced the teacher shortage in rural schools, particularly in remote and disadvantaged village primary schools and teaching sites, as indicated by the data presented above. Effective educational policy direction generates considerable institutional impacts in the local community. As such, H3 is confirmed. In addition to this, the impact of the control factors should not be disregarded. When compared to the baseline group of county-level school teachers, the likelihood that male teachers will work in town primary and secondary schools, as well as village primary schools and teaching sites, is significantly higher than the likelihood that fe- male teachers will do the same (p < 0.001) in Models 1, 2, and 3. The likelihood of a teacher working in a rural school decreases significantly with the teacher’s age (with negative coefficients, p < 0.001). Teachers who live in close proximity to schools are less likely to work in town schools and village primary schools (with negative coeffi- cients, p < 0.001). It is statistically significant that teachers of Han nationality are more likely to work in town schools and village primary schools than teachers of ethnic mi- nority backgrounds (such as Miao, Tujia, and other ethnic minorities). This suggests that Han educators have a greater propensity to work in rural schools. Robustness Check In the robustness test, teachers’ registered residence is replaced by birthplace, sources of teacher supply are by means of joining Bianzhi, and the Mlogit model is used for regression analysis. As shown in Table 6, the model’s analytical results remain signifi- cant despite slightly reduced values of coefficients of determination, suggesting the model’s remarkable robustness. In this study, the impacts of control variables such as gender, age, ethnicity, and proximity to the school are consistent with previous analyti- cal findings. Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1443 Table 6. Robustness Check on Hometown Effects on Rural Teachers’ Em- ployment. (7) (8) (9) (10) (11) (12) Model 1 Model 2 Model 3 Variable School Types (County-level schools as the baseline group) I II I II I II Gender 0.862*** 0.933*** 0.862*** 0.930*** 0.972*** 0.954*** (0.105) (0.081) (0.105) (0.081) (0.116) (0.087) Age -0.029*** -0.035*** -0.032*** -0.039*** -0.043*** -0.044*** (0.004) (0.003) (0.005) (0.004) (0.008) (0.006) Living around the school location or not -0.895*** -0.669*** -0.944*** -0.758*** -0.912*** -0.686*** (0.113) (0.090) (0.117) (0.093) (0.131) (0.101) Nationality (Han nationality as the baseline group) Miao -0.811*** -0.576*** -0.815*** -0.581*** -0.825*** -0.640*** (0.107) (0.083) (0.107) (0.083) (0.120) (0.091) Tujia -0.554*** -0.295*** -0.560*** -0.304*** -0.561*** -0.363*** (0.111) (0.087) (0.111) (0.087) (0.125) (0.095) Other ethnic minorities -1.432*** -0.403+ -1.437*** -0.407+ -1.395** -0.212 (0.373) (0.222) (0.373) (0.222) (0.445) (0.253) Birthplace 0.848*** 0.593*** 0.852*** 0.603*** 0.718*** 0.535*** (0.096) (0.074) (0.096) (0.074) (0.109) (0.081) Early rural life or education experience 0.239+ 0.414*** 0.377* 0.505*** (0.130) (0.103) (0.153) (0.116) Ways of Joining Bianzhi (Joining Bianzhi through examinations as the baseline group) From temporary positions 2.127*** 1.730*** (0.240) (0.202) From Targeted Position Program 0.285 0.539*** (0.179) (0.136) From Special Position Program -0.198 -0.081 (0.159) (0.118) From Secondary Normal Graduates Program -0.115 0.202+ (0.150) (0.106) Constant terms 1.259*** 2.474*** 1.198*** 2.364*** 1.222*** 2.157*** (0.173) (0.140) (0.175) (0.141) (0.282) (0.215) Pseudo R2 0.040 0.041 0.052 Log likelihood -5020.523 -5012.341 -4066.613 AIC 1.835 1.834 1.805 BIC -37095.184 -37085.711 -29858.928 Sample size 5,500 5,500 5,500 5,500 5,500 5,500 Notes: I: Village Primary Schools & Teaching Sites; II: Town Primary & Secondary Schools. 1. Values in paren- theses are standard errors. 2. Levels of significance: ***p<0.001, **p<0.01, *p<0.05, +p<0.1 Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1444 The key explanatory variable, birthplace, is a dummy variable with a value of 0/1 (0 = non-local birth, 1 = local birth). It has a positive regression coefficient of p0.001, indicating that, after controlling for variables such as gender, teachers born lo- cally are more likely to work in village primary schools and teaching sites, or town pri- mary and secondary schools (than county-level schools), and the difference is statisti- cally significant. The findings of the emotional hometown effects tests also show that the impact of teachers’ early years of rural living or education is favorable (all coeffi- cients were positive) and statistically significant (p < 0.05). As a result, the geograph- ical and emotional influences of home are still present. The impact of the variable “the way of joining Bianzhi” confirms that institutional hometown impacts are also signifi- cantly good. Teachers who join Bianzhi as temporary teachers are more likely to work at village primary schools and town schools than at county-level schools, compared to the baseline group (teachers who join Bianzhi through examinations). Joining Bianzhi from the Targeted Position Program and the Secondary Normal Graduates Program had both positive and statistically significant regression coefficients. Nonetheless, there is no statistically significant difference (p > 0.1) in rural school employment between teachers who join Bianzhi through the Special Position Program and those who join through examinations. After substituting teachers’ registered location and sources of teacher supply with other key explanatory variables, the impact of geographical, emotional, and institu- tional hometown effects on rural teachers’ employment remains significant, according to the results of the robustness test. The modified model yields essentially identical ana- lytical results to the original model. Consequently, we believe the model estimation results of this study to be reliable. Conclusions and Suggestions This study applies the polynomial logit model to assess hometown effects from geo- graphical, emotional, and institutional aspects based on a survey of 5,554 teachers from 117 towns in H province in the Wuling Mountains Zone. The conclusions are as follows. (i) Teachers with local registered residences are more likely to work in rural schools in China, proving the importance of geographical hometown effects. However, based on the subdivisions of registered residence, we find that the scope of local registered residence pertaining to this topic is limited to the local town, a smaller administrative division, as opposed to the local county (as- sumed by preview studies). Wei’s (2016) research in Gansu Province and Liu’s (2019) research in Guangdong, Zhejiang, Shanxi, Sichuan, and Guizhou Prov- inces supports the conclusion. (ii) Teachers with at least five years of rural life or education experience are more likely to work in village schools or teaching sites, demonstrating a favorable ef- fect of teachers’ hometown attachments on their employment in rural schools. In recruiting and educating rural teachers, consideration should be given to the motivational effect of teachers’ attachment to their hometown. Existing studies have also indicated that it is difficult to attract a sufficient number of qualified teachers using only market-based strategies in underdeveloped regions. Incor- porating spiritual components such as emotions, wills, and beliefs into teacher Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1445 education and recruiting is not only helpful in enhancing the retention rate of competent employees, but also in instilling a feeling of civic responsibility in teachers towards their hometowns (Li, 2020). (iii) The results of regression analysis show strong institutional hometown effects after controlling for geographical and emotional hometown preferences. Those educated by teacher supply augmentation programs are more likely to work in rural schools, notably village primary schools and teaching sites in more back- ward conditions, than those employed through open recruiting. It demonstrates how educational programs have aided in the expansion of China’s rural work- force. Preliminary studies have not paid enough attention to the policy-induced local consequences. Because of the severe lack of available human resources in rural schools, both the central and local governments have been making efforts to increase the number of available teachers (Wang & Wu, 2019). The central and local authorities have come to the conclusion that selecting, training, and recruiting students of local origins as rural teacher candidates under the logic of hometown effects is an effective way to ensure a sufficient supply of rural teaching staff. This conclusion has led to the formation of a consensus among the central and local authorities. However, an excessive reliance on institutional hometown effects may lead to some negative consequences, such as a high rate of default on the part of normal school students educated under the Free Teacher Education Program (Ren & Chen, 2020), and a high turnover rate of teachers produced by the Special Position Program who only use rural school positions as a transition in their careers (Wang & Feng, 2020). On the basis of this study’s empirical findings and results from prior research, we provide some policy recommendations for attracting and keeping rural teachers. First, the geographical quality of hometown effects is most prominent within a constrained administrative region, i.e., the local town. Its conclusion for policymakers is that more students from native towns should be recruited into future rural teacher edu- cation programs, and that when staffing rural schools, graduates should be encouraged to teach in their native towns’ rural schools. It is important to note that this recommen- dation may be more applicable in developing countries and regions, particularly those with large disparities in educational resources between urban and rural areas or with distinct racial or ethnic cultural characteristics, where teachers have stronger ties to their native lands and are obligated to care for their families and relatives, and where work- ing close to home helps them achieve work-life balance (An & Cao, 2017). Second, in the pre-service education of rural teachers, their attachment to their hometowns should be fostered (Zheng & Hu, 2018). According to existing research, 72.3% of teachers with extensive experience in rural schools were raised in rural areas, and 60% have schooling experience in rural settings (Boylan & McSwan, 1998). In a longitudinal research study, Hernan (2018) evaluates the effects of two distinct ap- proaches to rural teacher training. The first program engaged pre-service teachers in a variety of classroom-, school-, and/or community-based professional activities; the se- cond program directly recruits and trains students from rural backgrounds. It is deter- mined that the second strategy is superior to the first because those with a background in rural life are better able to comprehend rural culture and integrate into rural life. Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1446 Third, to maintain an appropriate supply of skilled rural teachers, it is crucial to invest more in the Targeted Position Program through partnerships with secondary normal schools, normal junior colleges, and normal universities, and to improve the educational quality of the program. However, strengthening the execution of rural teacher support policies is more crucial. According to Merrow (1999), addressing the teacher shortage solely through recruitment is a “wrong diagnosis” and a “phony cure.” Increasing rural teachers’ compensation, elevating their social status, and providing them with more accolades may be more effective solutions to the rural school staffing issue. In conclusion, the present study attempts to experimentally answer the question, “Who teaches in rural schools?” based on the hometown effects hypothesis. It has broadened the scope and dimensions of the concept by sparking discussion on the insti- tutional aspect of hometown effects. However, the limitations of this study should be acknowledged. First, there are significant gaps between central and western China, as well as between poor regions in various provinces. This study focused on the origins of rural teachers in the Wuling Mountains Zone in central China, and the sample is some- what unrepresentative, which hinders the generalizability of the findings. Second, the hometown effects hypothesis is a Western academic theory. To confirm the explanatory power and scope of applicability of the theory in China, it is necessary to conduct addi- tional empirical studies based on evidence. Furthermore, this study identified the insti- tutional impact of hometown effects on rural teachers in China, but the influence mech- anism of pertinent policies on rural teachers’ employment decisions remains to be in- vestigated. References An, X. M., & Cao, X. M. (2017). Who is more willing to teach in rural schools? -Based on the investigation and analysis of rural teach- ers’ intention of mobility. Journal of Educa- tional Science of Hunan Normal University, 16(4):12-15. DOI: https://doi.org/10.19503/j.cnki.1671- 6124.2017.04.003 Boyd, D., Lankford, H., Loeb, S., & Wyckoff, J. (2005). The draw of home: How teachers’ preferences for proximity disadvantage ur- ban schools. Journal of Policy Analysis and Management: The Journal of the Association for Public Policy Analysis and Management, 24(1):113-132. DOI: https://doi.org/10.1002/pam.20072 Boyd, D., Lankford, H., Loeb, S., & Wyckoff, J. (2013). Analyzing the determinants of the matching of public school teachers to jobs: Disentangling the preferences of teachers and employers. Journal of Labor Economics, 31(1):83-117. DOI: https://doi.org/10.1086/666725 Boylan, C. (2010). Staffing Rural Schools: A New Perspective. Demographic Change in Australia’s Rural Landscapes, 2010(12):329-351. DOI: https://doi.org/10.1007/978-90-481-9654- 8_14 Boylan, C., & McSwan, D. (1998). Long-staying rural teachers: Who are they? Australian Journal of Education, 42(1):49-65. DOI: https://doi.org/10.19503/j.cnki.1671-6124.2017.04.003 https://doi.org/10.19503/j.cnki.1671-6124.2017.04.003 https://doi.org/10.1002/pam.20072 https://doi.org/10.1086/666725 https://doi.org/10.1007/978-90-481-9654-8_14 https://doi.org/10.1007/978-90-481-9654-8_14 Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1447 https://doi.org/10.1177/00049441980420010 4 Cannata, M. (2010). Understanding the teacher job search process: Espoused preferences and preferences in use. Teachers College Record, 112(12): 2889-2934. DOI: https://doi.org/10.1177/01614681101120120 5 Cuervo, H., & Acquaro, D. (2018). Exploring metropolitan university pre-service teacher motivations and barriers to teaching in rural schools. Asia-Pacific Journal of Teacher Education, 2018(4):384-398. DOI: https://doi.org/10.1080/1359866X.2018.143 8586 Engel, M., Jacob, B. A., & Curran, F. C. (2014). New evidence on teacher labor supply. American Educational Research Journal, 51(1):36-72. DOI: https://doi.org/10.3102/0002831213503031 Goff, P. T., & Bruecker, E. M. (2017). The role of place: Labor market dynamics in rural and non-rural school districts (WCER Working Paper No. 2017-4). Retrieved from University of Wisconsin– Madison, Wiscon- sin Center for Education Research website: http://www.wcer.wisc.edu/publications/work ing-papers Hong, Y. B. (2015). Comparison of logistic model coefficients and solutions: a literature review. Chinese Journal of Sociology, 35(4):220-241. Hunan First Normal University. (2017, May 10). The Graduation of first 6-yearpublic-funded normal students. Retrieved from: https://www.hnfnu.edu.cn/info/1044/3818.ht m. 2021-10-20. Li, J. M. (2020). A discussion on the core issues of rural teacher supply in China. Theory and Practice of Education, 40(4):50-54. Liu, X. Q. (2019). Geographical structure char- acteristics of rural teachers in the context of spatial division of teacher resources and their influencing mechanism. Education and Economy, 35(3):73-77. DOI: http://dx.chinadoi.cn/10.3969/j.issn.1003- 4870.2019.03.009 Merrow, J. (1999). The teacher shortage: Wrong diagnosis, phony cures. Education Week, 1999(10):38-64 Ministry of Education. (2006). Notice on the implementation of the Special Position Pro- gram for rural compulsory education teach- ers in 2006. Retrieved from: http://www.moe.gov.cn/srcsite/A10/s7058/2 00605/t20060519_81622.html Ministry of Education. (2007). The Measures for the Implementation of Free Education for Students of Normal Universities Affiliated to the Ministry of Education. Retrieved from: http://www.moe.gov.cn/srcsite/A10/s7058/2 00605/t20060519_81622.html Monk, D. H. (2007). Recruiting and retaining high-quality teachers in rural areas. Future of Children, 17(1):155-174. DOI: https://doi.org/10.1353/foc.2007.0009 Neeshatlen, V. (1985). Collaborating on Writing Assignments: A Workshop with Theoretical Implications. Journal of Teaching Writing, 4(2):234-246. Pu, D. Y., Wang, L. J., Ren, X. L., & Xu, H. P. (2019). Special-position teachers in western China: Based on the survey and analysis of a large sample of 3468 special-position teach- ers who have fulfilled their service in the western regions. School Administration, 2019(2):40-43. DOI: https://doi.org/10.3969/j.issn.1002- 2384.2019.02.013 Reininger, M. (2012). Hometown disadvantage? It depends on where you’re from: Teachers’ location preferences and the implications for staffing schools. Educational Evaluation and Policy Analysis, 34(2):127-145. DOI: https://doi.org/10.3102/0162373711420864 South Reviews. (2021, January 11). China Needs More Rural Teachers in 2021. Re- trieved from: https://mp.weixin.qq.com/s/boBwUw8SMZ9 UvoGbMO18tw. 2021-10-20. Stuit, D., & Smith, T. M. (2010). Teacher Turn- over in Charter Schools. Research Brief. Na- tional Center on School Choice, Vanderbilt University, 2010(4) Wang, D. F., Feng, X. X. (2020). Why special positions become a career springboard -- a narrative study based on the experience of a special-position teacher. Educational Re- search Monthly, 2020(1):72-78. Wang, X. S., & Wu, Z. H. (2019). Review and improvement of the mechanism for retaining https://doi.org/10.1177/000494419804200104 https://doi.org/10.1177/000494419804200104 https://doi.org/10.1177/016146811011201205 https://doi.org/10.1177/016146811011201205 https://doi.org/10.1080/1359866X.2018.1438586 https://doi.org/10.1080/1359866X.2018.1438586 https://doi.org/10.3102/0002831213503031 http://www.wcer.wisc.edu/publications/working-papers http://www.wcer.wisc.edu/publications/working-papers http://dx.chinadoi.cn/10.3969/j.issn.1003-4870.2019.03.009 http://dx.chinadoi.cn/10.3969/j.issn.1003-4870.2019.03.009 http://www.moe.gov.cn/srcsite/A10/s7058/200605/t20060519_81622.html http://www.moe.gov.cn/srcsite/A10/s7058/200605/t20060519_81622.html http://www.moe.gov.cn/srcsite/A10/s7058/200605/t20060519_81622.html http://www.moe.gov.cn/srcsite/A10/s7058/200605/t20060519_81622.html https://doi.org/10.3969/j.issn.1002-2384.2019.02.013 https://doi.org/10.3969/j.issn.1002-2384.2019.02.013 https://doi.org/10.3102/0162373711420864 https://mp.weixin.qq.com/s/boBwUw8SMZ9UvoGbMO18tw.%202021-10-20 https://mp.weixin.qq.com/s/boBwUw8SMZ9UvoGbMO18tw.%202021-10-20 Zhang et al. Who Teaches in Underdeveloped Rural Chinese Schools? BECE, Vol.11, No. 1, 2022 1448 rural teachers. Education Science, 35(6):71- 77. Wei, Y. (2016). Teacher mobility in rural China: Evidence from Northwest China. Michigan State University. Wei, Y., Zhou, S., & Liu, Y. (2020). The draw of home: How does teacher’s initial job placement relate to teacher mobility in rural China? PLoS ONE, 15(1):e0227137. https://doi.org/10.1371/journal.pone.022713 7 Wu, Z. H., Qin, Y. Y. (2015). A Report on Chi- na Rural Education Development 2013-2014. Beijing: Beijing Normal University Press. Yang, L. & Zhang, T. (2020). Family back- ground, key middle schools, and educational attainment. Education and Economy, 36(5):33-44. Yang, W. A. (2019). Evolution of rural primary teacher supply supplement policy: A review of the past 70 years and prospect. Educa- tional Research, 40(7):16-25. Yarrow, A., Herschell, P., & Millwater, J. (1999). Listening to Country Voices: Prepar- ing, Attracting and Retaining Teachers for Rural and Remote Areas. Education in Rural Australia, 9(2):1-12. Zheng, X. R. & Hu, Y. (2018). Return and hope: The oral history of young rural teachers. Nanning: Guangxi Education Publishing House. Zhao, Y. & Liu, J. (2017). Implementation of free education policy for normal students in China and strategies for improvement. Edu- cation Exploration, 2017(4):89-93. Received: 19 April 2022 Revised: 29 April 2022 Accepted: 10 May 2022 https://doi.org/10.1371/journal.pone.0227137 https://doi.org/10.1371/journal.pone.0227137 OriginalArticle-TianjiaoZhang-BECE_Title_May2022 OriginalArticle-TianjiaoZhang--BECE_MaintextMay2022