【作者】 李凤；

【导师】 甘犁； 董艳；

【作者基本信息】 西南财经大学 ， 产业经济学， 2010， 博士

【摘要】 高考一直是社会各界高度关注的焦点话题。在高校招生录取这个庞大系统工程中,志愿填报和录取机制属于其中一个关键而复杂的子系统,同时也是高考制度改革的重点和难点。当前我国志愿填报和录取机制还是存在一些突出的弊端,比如考生志愿偏好表达不完全,信息不充分,博弈成分较大,“高分低录”和“高分落榜”现象无法避免,等等。本文基于经济学的效率和公平概念,试图从资源配置角度对这些问题进行梳理和分析,并结合国外相关文献的研究成果和实践经验,探索我国高考志愿填报和录取机制改革的方向。概括而论,本文的学术贡献主要体现在三点：(1)研究视角新颖。与其他大部分文献不同,本文基于经济学的效率和公平概念,试图从资源配置角度研究志愿填报和录取机制。(2)研究方法独特。本文采用的方法主要包括文献资料法、博弈理论、蒙特卡洛模拟法、实证研究法和调查访谈法,是至今为止第一篇从经济学角度对该问题进行深刻分析的博士论文。(3)研究结论具有可实践性。通过严谨的理论推导和深入的数据挖掘,本文全面剖析了我国高考志愿填报和录取机制改革取得的成绩以及存在的弊端,并结合国外相关文献的研究成果和实践经验,为我国高考招生制度改革提出了一些切实可行的建议。本文总共分为六章。第一章是绪论,主要阐述了选题的意义和研究的价值,概括了本文研究的方法和贡献。第二章主要回顾和总结了我国高考志愿填报和录取机制的历史变革,包括志愿填报和录取实现手段、志愿填报方式、投档方式和志愿设置情况。通过对高考制度恢复以后高考志愿填报和录取机制发展历程的梳理,探索其演变规律,分析各地采取不同政策的原因以及未来发展的方向。第三章从资源配置的角度分析考前填报、考后估分填报、考后知分填报和实时动态志愿等四种志愿填报方式的优劣。第四章讨论的内容围绕高考录取机制展开。首先,本文分析了传统梯度志愿固有的弊端,然后针对目前教育部拟在全国推行的平行志愿提出了改进的途径和方式。最后,本文根据国外文献在学校录取问题方面的研究成果以及实践经验提出了一种既能让考生真实表达其偏好,又能体现公平和效率,还能化解统一高考与自主招生之间矛盾的一种机制——Gale-Shapley学生最优机制。第五章首先通过对我校学生微观数据的计量分析,从高校人才培养和大学生就业与升学角度探讨了高校招生录取评价标准,为将来志愿填报和录取机制改革的必要性提供了实证支持。然后,本文还从受教育机会的公平性角度探讨了我国高考录取制度中另一个敏感问题——文理分科。最后一章是结论,综合前面的分析结果,归纳本文研究的基本结论,对高考志愿填报和录取机制改革的基本走向提出可行性建议。本文主要的研究结论和政策建议如下：(1)考前、估分和知分填报各有利弊,但都无法实现效率与公平的统一。从整体考生的录取率考虑,知分填报优于估分填报,估分填报优于考前填报,从资源配置的效率考虑,随着分数信息的增多,整体考生的福利能够得到更大的改善,但其中一部分考生的福利改善是以其他一些考生的福利损失为代价。考前填报虽然在公平性和筛选人才效率方面具有优势,但资源浪费较大；知分填报的考生整体福利最高,但不公平现象严重；如果估分误差控制在一定范围之内,虽然能够提高公平性,但也只是一种中庸之道,它既不是资源利用率最高的,也不是公平程度最高的。因此,在不完全信息条件下,无法实现效率与公平的统一。(2)实时动态志愿能够兼顾公平与效率,因此,它可能将成为我国志愿填报改革的方向和目标。实时动态志愿消除了由于不完全信息所导致的志愿填报博弈。以考生成绩为衡量标准,这种方式能够提高所有上线考生的录取率,并且能够提高在知分填报时所有考生的效用。实时动态志愿虽然相较于估分填报和考前填报没有实现“帕累托改进”,但如果从维护大多数考生,尤其是好成绩考生的利益时,这种方式比前三种方式都更优。虽然如此,对于高质量考生而言,这种“光鲜亮丽”的志愿填报方式仍然不及考前填报,因为这部分考生的录取概率和效用在考前填报模式下依旧是最高的。(3)实行多次高考制度,让考生高考成绩能够体现真实水平。给予更多的考生偏好信息固然能够解决“高分低就”或“高分落榜”的弊端,但依然没有真正维护到好质量考生的利益。要实现这一终极目标,必须减少甚至消除考生质量与高考成绩之间的误差。解决办法可以考虑采取多次高考制度,或者将考生能力测评更加客观、科学、公平地体现在高考分数之上,让考生高考成绩能够体现其真实水平。(4)梯度志愿弊端严重,应该逐步淘汰于高考录取机制。梯度志愿投档方式(即波士顿机制)算法最简单,在计算机还未普及之前被广泛运用于解决学校录取问题。但是,波士顿机制既不是公平的,也不是帕累托最优。最重要的是,它不具有抗策略性。由于真实申报偏好会导致不公平的匹配结果,因此学生有很强的激励去“操纵”偏好。但对每个家长和考生来说,填好志愿并非易事,特别是在信息不充分的情况下。(5)平行志愿推进了公平与效率,但还存在继续改进的空间。如果每个学校都以分数来评价学生,平行志愿就是一个公平、无浪费、个人理性、抗策略,并且帕累托最优的录取机制。但是,当考生填报志愿的数目受到限制时,平行志愿就达不到这种理想效果,并且也不能完全解决填报志愿难的问题。不过,我们的模拟分析表明,在知分填报或者竞争增大时,志愿数目受限的平行志愿也优于梯度志愿,大大缓解了“高分低就”或“高分落榜”现象,而部分平行志愿收效甚微。从福利角度考虑,平行志愿相对于梯度志愿并非“帕累托改进”,不过从分数公平角度说,这就是一种进步。而增加考生填报志愿数目,减小高校投档比例,按大类招生或打破校际次序以专业志愿为准,这些措施可以进一步提高公平和效率。(6)Gale-Shapley机制是一个更好的录取机制选择。无偏好限制的平行志愿和实时动态志愿都能达成公平与效率,不过,其前提假设是高校都以分数评价考生。如果允许高校对考生有不同偏好,即高校不再“唯分论人”,都根据各自的评价标准录取考生,那么,Gale-Shapley机制就是一个很好的选择,因为这种算法既是公平的,也是抗策略的,并且其匹配结果相对于其他任何公平的匹配结果都是“帕累托改进”。它能够化解统一高考与自主招生之间的矛盾,赋予高校更充分的自主招生权力,促进“多元化评价”的实施与推行。(7)取消文理分科能够提高文科考生录取概率,促进教育公平。历年来,文科考生录取率都远远低于理科考生,这意味着文科考生没有享受到公平的受教育机会。取消分科之后上线文科考生的录取概率将增加,这是由于文科考生录取比例的增加(竞争程度的减少)以及整体排名上升的结果。因此,取消分科可以改善文理考生录取率差异较大的现象,这在维护考生利益、推进教育公平方面具有深刻的价值和意义。(8)建立和完善科学、客观的综合评价机制,实行统一考试与自主招生制度。通过对我校学生微观数据的计量分析发现,当前的高考分数在一定程度能反映出学生掌握和运用知识的能力,为高校选拔新生提供了比较科学的评价依据,但考生的综合素质也应成为评价体系中的一个重要组成部分,而且不同的院校应有不同的评价标准。这为将来志愿填报和录取机制改革的必要性提供了实证支持。综上所述,在现行全国高考制度下,分数是评价考生能力、平等享有受高等教育机会的最客观的依据。在这种评价标准下,实时动态志愿和偏好不受限制的平行志愿最能体现公平与效率。随着“多元化评价”、“多样化选拔”成为高校招生考试制度改革的重要目标取向,纯粹“按分分配”的录取规则俨然已经不能跟上改革步伐。在这种大环境和大趋势下,必然要求推出一种崭新的录取机制。这种机制就是本文所建议的Gale-Shapley机制。更多还原

【Abstract】 This paper discusses application and admission mechanism in the National College Entrance Exam. Application and admission mechanism is a key and difficult part during the NCEE reform. It is associcated closely with the populace interest, and plays an important role in promoting education fairness and education for all-around development.The contributions this paper makes are summaried in three aspects:(1) the research perspective is new. Different from most other studies, this paper trys to analyze application and admission machenism from the perspective of resource allocation based on the efficiency and fairness concept in economics. (2) The study methods are unique. The methods used in this paper include literature review, game theory, montecarlo simulation, empirical study and interview.This is the first doctoral dissertation studying from the economics perpective. (3) The suggestions this paper proposes are practical. Through theory and impirical study, this paper analyzes deeply the achievement of application and admission mechanism reform and the disadvantages exist in it. Based on the research findings and practice experiences in developed contries, this paper gives some useful and practical advices on the application and admission mechanism reform.This paper is arranged in six chapters, which are summaried as follows:The first chapter discusses the importance, contribution and framework of this paper. The second chapter reviews the historical trend of application and admission mechanism in NCEE. (1) Application and admission methods. With the development of internet technology and educational informatization, application methods changed from ’application in paper form’ to ’application on line’. As for the admission methods, it developed in three stages: ’manual operation’,’computer assistance’ and ’admission on line’. (2) Application timing. There are three kinds of application timing:’application before exam (ABE)’,’application after exam and before scores (ABS)’ and’ application after scores (AAS)’. Each kind has its own advantages and disadvanteges, and different provinces adopt different timing. In general, more and more provinces began to apply AAS. No matter what kind of application timing, students are involved in an application game, and’ real time dynamics application’ could solve this dilemma. (3) Admission mechanism. Admission mechanism is divided into three modes:’tiers choices’,’patial parallel choices’ and ’complete parallel choices’. More and more provinces would change its admission mechanism into the last one. The number of school choices and speciality choices are also different among provinces and admission stage. Anyway, ’the second time application’ is a good way to protect student from failing.The third chapter compares the application timing. We propose a theoretical model on the application and admission process in the NCEE, the parameter values of the model are calibrated based on the observed aggregate data and the model is solved by simulation. We find that AAS has the highest admission rate, and ABE has the lowest one. However, there will be more high-mark students failing under ABS or AAS, especially under more intense competition. If assessment error is controlled in a certain scope, ABS could relief the phenomenon of ’high-mark students failing’. As for the colleges, ABE would result in least number of admitted students, but the quality of admitted students is higher than the other two modes. With more information on score, students are better off in a whole, but some students are worse off, even including the high-mark students when competition becomes more intense. In general, AAS is the best way to protect most students’ interest; however, it is not the best way to screen high-quality students. Using the student-level data from the NCEE, we find that the admission rate of their first choice is highest under AAS and lowest under ABE, which provides evidence for some theory conclusion.’Real time dynamics application (RTDA) can be seen as a game with complete information. This mode solves application dilemma for students and elimites the phenomenon of ’high-mark students failing’. Though RTDA is fair and Pareto Efficient, this mode is also not the best one for high quality students. ’Multi-NCEE’ system is one of the solutions to dispose of this conflict.The fourth chapter discusses the admission methanism. Before 2008, nealy all provinces adopted ’tier choice’, which is know as ’boston mechanism’ in foreign literatures. The algorithm of Boston mechanism is simple and efficient, thus it is applied widely in school admission problem. However, Boston mechanism is neither fair nor Pareto Efficient. Most importantly, it is not strategy-proof. And finding a good school choice strategy is not an easy thing, especially with incomplete information.In 2008, the Ministry of Education began to promote ’parellel choices’, the admission rule was changed from ’choices first’ into ’scores first, follow choices’. The algorithem efficient of this ’score absoluta mechanism’ is also high, and it is a fair, strategy-proof and Pareto Efficient mechanism when college value students based on their scores. However, when the school choices are limited in number, ’parellel choices’ would not achieve an ideal effect and the application game is also difficult for students and parents. Yet our analysis suggests that ’parellel choices’ with constraint is better than ’tier choices’ under AAS, especially under intense competition, because it protects high-mark students, but ’partial parallel choices’ is not significantly different from ’tier choices’. Moreover, by increasing the choice number or decreasing the submission ratio or applying in speciality choice, we could improve education fairness and efficiency futher more.If we allow colleges assess students according to their own standards other than scores,’parallel choices’ is no longer a good way to sovle the admission problem. In that case, Gale-Shapley student optimal table mechasim (SOSM) is a good choice instead. SOSM is a fair, strategy-proof mechanism, and it is Perato Efficient than any other fair mechanism.The fifth chapter contains two special topics. The first topic empirically analyzes the assessment standard in the NCEE. We find that the scores in current NCEE reflect students’ academical ability in certain degree, but comprehensive quality should also be considered when assessing a student’s all-around ability. We also find that different schools should take different assess standards to meet their own requirments. These conclusions provide an empirical support to the necessity of application and admission reform.The second topic is about division of discipline. We analyze the consequence and the fairness of rescinding the discipline-dependent exam and admission system. We estimate an empirical model on college admissions. Based on this model, we find that rescinding the discipline-dependent testing and admission system would increase the admission rate of liberal arts students, largely because of an increase of liberal arts students’ relative ranking comparing with the science students. And this would play an important role in protecting student interest and promting education fairness.The last chapter makes a conclusion about the whole paper, and gives some advices on the reform of national college entrance exam.更多还原