选大学、选专业还是选地方:基于匹配理论分析学生对高校的偏好
|
摘要
匹配理论证明,在公平的匹配结果下,学校的录取分数线反映了学生对学校的偏好。基于这一理论,本文对2014年110所211高校的录取分数线和学校特征进行回归,估计学生对高校的偏好,重点考察了学校的等级(质量)、专业类型和所在区域这三类特征的影响。研究结果发现,学生对这三类特征均表现出明显的偏好。其中,对学校等级的偏好强度大于所在区域,而对所在区域的偏好强度又大于学校的专业类型。文科生(而非理科生)对于本地大学存在显著偏好。
Matching theory proves that,under a fair matching outcome,the score cutoffs in college admission reflect studentspreferences for colleges.Based on this theory,we use the cutoff scores from 110 "two-two-one"colleges in 2014 and perform a regression on college characters,to reveal student preferences over colleges.We focus on three categories of college characters:college ranks(or qualities),college majors and college locations. We find that students have significant preferences for all three characters.Preference intensity is strongest on college ranks,and then on locations,and least on majors.Art and humanity students(as opposite to science and technology students)also have a significant preference of local colleges.
Matching theory proves that,under a fair matching outcome,the score cutoffs in college admission reflect studentspreferences for colleges.Based on this theory,we use the cutoff scores from 110 "two-two-one"colleges in 2014 and perform a regression on college characters,to reveal student preferences over colleges.We focus on three categories of college characters:college ranks(or qualities),college majors and college locations. We find that students have significant preferences for all three characters.Preference intensity is strongest on college ranks,and then on locations,and least on majors.Art and humanity students(as opposite to science and technology students)also have a significant preference of local colleges.
引文
陈小龙.2015.中国城市统计年鉴2015[M].北京:中国统计出版社.
邱均平.2016.高考分数线与专业排名2016—2017[M].北京:科学出版社.
文祺.2015.全国高校录取分数线统计(本科)[M].北京:北京理工大学出版社.
武书连.2013.挑大学选专业——2014高考志愿填报指南[M].北京:中国统计出版社.
Abdulkadirog(lu A,S9nmez T.2003.School choice:A mechanism design approach[J].American Economic Review,93(3):729-747.
Agarwal N,Somaini P.2014.Demand analysis using strategic reports:An application to a school choice mechanism[R].NBER Working Paper No.20775.
Agarwal N.2015.An empirical model of the medical match[J].American Economic Review,105(7):1939-1978.
Akyol S P,Frisancho V,Krishna K,et al.2013.Preferences,selection,and value added:A structural approach applied to Turkish exam high schools[R].CESifo Working Paper Series No.4302.
Azevedo E M,Leshno J D.2016.A supply and demand framework for two-sided matching markets[J].Journal of Political Economy,124(5):1235-1268.
Boyd D,Lankford H,Loeb S,et al.2013.Analyzing the determinants of the matching of public school teachers to jobs:Disentangling the preferences of teachers and employers[J].Journal of Labor Economics,31(1):83-117.
Burgess S,Greaves E,Vignoles A,et al.2009.What parents want:School preferences and School choice[R].Working Paper No.09/222.
Haeringer G,Klijn F.2009.Constrained school choice[J].Journal of Economic Theory,144(5):1921-1947.
Hastings J S,Kane T J,Staiger D O.2005.Parental preferences and school competition:Evidence from a public school choice program[R].NBER Working Paper No.11805.
He Y H.2012.Gaming the Boston school choice mechanism in Beijing[R].Working Paper.
Lien J W,Zheng J,Zhong X H.2016.Preference submission timing in school choice matching:Testing fairness and efficiency in the laboratory[J].Experimental Economics,19(1):116-150.
Lien J W,Zheng J,Zhong X H.2017.Ex-ante fairness in the Boston and serial dictatorship mechanisms under pre-exam and post-exam preference submission[J].Games and Economic Behavior,101:98-120.
Wu B Z,Zhong X H.2014.Matching mechanisms and matching quality:Evidence from a top university in China[J].Games and Economic Behavior,84:196-215.
Wu B Z,Zhong X H.2016.Fairness of the Boston matching mechanism in Chinas college admission[R].Working Paper.
(1)搜狐教育:http://learning.sohu.com/20150619/n415348128.shtml。
(2)中国教育在线:《2015高招调查报告》http://www.eol.cn/html/g/report/2015/report1.shtml#baogao1-1;《2014年高招调查报告》http://www.eol.cn/html/g/report/2014/report1.shtml#baogao1-1。
(1)反映专业偏好的更好的方式是直接考察各高校各专业的录取分数线,但收集这一数据困难很大,本文暂时不采取这一方式。
(1)对于Azevedo and Leshno(2016)而言,既不要求学校对学生的优先序相同,因此不同学校可以考察学生不同的成绩维度(比如只看语文,只看数学等),也不要求学生对学校的偏好相同,因此是更为一般的结论。
(1)计划招生人数不包含定向国防生和贫困生等单独设置的招生专业或名额,因为录取分数线通常也不包含这些学生---他们通常都是单独录取的(如在提前批录取)。
(1)根据命题2,某一高校的招生人数变化还可能引起其他高校的分数线变化,但究竟哪些高校受到影响,如何引入计量模型中,本身是个复杂问题,本文不打算进行深入处理。本文主要采取在回归残差项中考虑同一匹配系统内的残差关联予以处理,当然这一处理也许并不完全。考虑到绝大部分样本分处不同的匹配系统,这一招生人数之间的关联对于回归结果的影响应该不是很大。
(1)注意:我们无法同时使用所在区域和所在城市两个特征变量效应的加总。因为两个变量不是完美正相关,处于其中一个变量最小(大)值的样本,未必在另一个变量下也是最(小)和最(大)值。不过考虑所在区域和所在城市GDP大体是正相关的,即发达地区通常对应发达城市,我们估算的地区变量的偏好强度有可能是低估的。
(1)见《国务院关于深化考试招生制度改革的实施意见》(国发[2014]35号),2014年9月3日,新浪教育,http://edu.sina.com.cn/gaokao/2014-09-04/1016432710.shtml。
(2)比较异常的结果是招生名额在个别回归中显著为正,与理论预期不符。一个可能的解释是招生名额可能具有某种内生性,会随招生质量变化而调整。
(1)《教育部:不再新设211和985工程学校》,新浪教育,2011年12月31日,http://edu.sina.com.cn/gaokao/2011-12-31/1140322608.shtml。
邱均平.2016.高考分数线与专业排名2016—2017[M].北京:科学出版社.
文祺.2015.全国高校录取分数线统计(本科)[M].北京:北京理工大学出版社.
武书连.2013.挑大学选专业——2014高考志愿填报指南[M].北京:中国统计出版社.
Abdulkadirog(lu A,S9nmez T.2003.School choice:A mechanism design approach[J].American Economic Review,93(3):729-747.
Agarwal N,Somaini P.2014.Demand analysis using strategic reports:An application to a school choice mechanism[R].NBER Working Paper No.20775.
Agarwal N.2015.An empirical model of the medical match[J].American Economic Review,105(7):1939-1978.
Akyol S P,Frisancho V,Krishna K,et al.2013.Preferences,selection,and value added:A structural approach applied to Turkish exam high schools[R].CESifo Working Paper Series No.4302.
Azevedo E M,Leshno J D.2016.A supply and demand framework for two-sided matching markets[J].Journal of Political Economy,124(5):1235-1268.
Boyd D,Lankford H,Loeb S,et al.2013.Analyzing the determinants of the matching of public school teachers to jobs:Disentangling the preferences of teachers and employers[J].Journal of Labor Economics,31(1):83-117.
Burgess S,Greaves E,Vignoles A,et al.2009.What parents want:School preferences and School choice[R].Working Paper No.09/222.
Haeringer G,Klijn F.2009.Constrained school choice[J].Journal of Economic Theory,144(5):1921-1947.
Hastings J S,Kane T J,Staiger D O.2005.Parental preferences and school competition:Evidence from a public school choice program[R].NBER Working Paper No.11805.
He Y H.2012.Gaming the Boston school choice mechanism in Beijing[R].Working Paper.
Lien J W,Zheng J,Zhong X H.2016.Preference submission timing in school choice matching:Testing fairness and efficiency in the laboratory[J].Experimental Economics,19(1):116-150.
Lien J W,Zheng J,Zhong X H.2017.Ex-ante fairness in the Boston and serial dictatorship mechanisms under pre-exam and post-exam preference submission[J].Games and Economic Behavior,101:98-120.
Wu B Z,Zhong X H.2014.Matching mechanisms and matching quality:Evidence from a top university in China[J].Games and Economic Behavior,84:196-215.
Wu B Z,Zhong X H.2016.Fairness of the Boston matching mechanism in Chinas college admission[R].Working Paper.
(1)搜狐教育:http://learning.sohu.com/20150619/n415348128.shtml。
(2)中国教育在线:《2015高招调查报告》http://www.eol.cn/html/g/report/2015/report1.shtml#baogao1-1;《2014年高招调查报告》http://www.eol.cn/html/g/report/2014/report1.shtml#baogao1-1。
(1)反映专业偏好的更好的方式是直接考察各高校各专业的录取分数线,但收集这一数据困难很大,本文暂时不采取这一方式。
(1)对于Azevedo and Leshno(2016)而言,既不要求学校对学生的优先序相同,因此不同学校可以考察学生不同的成绩维度(比如只看语文,只看数学等),也不要求学生对学校的偏好相同,因此是更为一般的结论。
(1)计划招生人数不包含定向国防生和贫困生等单独设置的招生专业或名额,因为录取分数线通常也不包含这些学生---他们通常都是单独录取的(如在提前批录取)。
(1)根据命题2,某一高校的招生人数变化还可能引起其他高校的分数线变化,但究竟哪些高校受到影响,如何引入计量模型中,本身是个复杂问题,本文不打算进行深入处理。本文主要采取在回归残差项中考虑同一匹配系统内的残差关联予以处理,当然这一处理也许并不完全。考虑到绝大部分样本分处不同的匹配系统,这一招生人数之间的关联对于回归结果的影响应该不是很大。
(1)注意:我们无法同时使用所在区域和所在城市两个特征变量效应的加总。因为两个变量不是完美正相关,处于其中一个变量最小(大)值的样本,未必在另一个变量下也是最(小)和最(大)值。不过考虑所在区域和所在城市GDP大体是正相关的,即发达地区通常对应发达城市,我们估算的地区变量的偏好强度有可能是低估的。
(1)见《国务院关于深化考试招生制度改革的实施意见》(国发[2014]35号),2014年9月3日,新浪教育,http://edu.sina.com.cn/gaokao/2014-09-04/1016432710.shtml。
(2)比较异常的结果是招生名额在个别回归中显著为正,与理论预期不符。一个可能的解释是招生名额可能具有某种内生性,会随招生质量变化而调整。
(1)《教育部:不再新设211和985工程学校》,新浪教育,2011年12月31日,http://edu.sina.com.cn/gaokao/2011-12-31/1140322608.shtml。