“易地教育扶贫”的理论与路径研究

That the education in the deep poverty areas has short boards and has difficult to achieve leapfrog development in the short term restricts the role of education to block the inter-generational transmission of poverty. Migration is an important factor in human capital investment. In the high school and the University level, a variety of forms of " educational migration" can help the poor people to improve their human capital and help to achieve the precision of poverty alleviation. The project focuses on the theory and the implementation path of educational migration for poverty alleviation. This project is to explore the transmission mechanism of educational migration on getting rid of poverty, to reveal the characteristics of educational migration and its influence factors of poor students, to evaluate the effect and restriction factors of migration poverty alleviation policy, and to put forward the improvement plan for future policy. The features of the study is : first, through the big data of poverty households registration and education student information , and through the machine learning method, regression method, the study is in the lead in the data and methods; second, using random field experiment methods, the study can achieve the quantitative evaluation and reveals the causal relationship; third, through analyzing the typical case, the study has important value for summarizing Chinese poverty alleviation plan.

深度贫困地区的基础教育存在明显短板且短期内难以实现跨越式发展，制约了教育阻断贫困代际传递的作用。迁移是人力资本投资的重要要素。在高中及大学阶段开展多种形式的“易地教育”，有助于贫困人群的人力资本提升，有助于精准扶贫的实现。本课题聚焦“易地教育扶贫”的理论与实现路径，探讨通过教育迁移摆脱贫困的传导机制，揭示贫困学生的教育迁移特征及其影响因素，评价易地教育扶贫政策的效果及制约因素，提出未来政策的改进方案。本研究的重要特色在于：第一、利用贫困户建档立卡数据和教育学籍信息数据等行政大数据，通过机器学习方法、计量回归方法对迁移的特征、影响因素及作用进行精细化描述，在数据和方法上具有领先性；第二、采用随机实地试验方法对促进教育迁移的因素做出定量评价，对因果关系揭示清晰；第三、对易地教育典型案例进行案例分析，对提炼中国扶贫方案有重要价值。

贫困地区的教育存在明显短板且短期内难以实现跨越式发展，制约了教育阻断贫困代际传递的作用。迁移是人力资本投资的重要要素。在高中及大学阶段开展多种形式的“易地教育”，有助于贫困人群的人力资本提升，有助于精准扶贫的实现和乡村振兴的接续。本课题聚焦“易地教育扶贫”的理论与实现路径，描述了国内、国际对贫困学生进行教育扶贫的政策经验，对中国建档立卡户信息、中国贫困学生就学信息进行了整理、展示与分析，对研究易地教育扶贫所需的机器学习算法进行了开发和使用。本项目识别出了对贫困学生从学前期干预的必要性和有效性，描述了英国对贫困学生社会流动性监测的有效做法；结合经费投入数据论证了我国深度贫困地区实施贫困学生迁移就学的必要性与可行性；发现了我国建档立卡学生高等教育异地就学数量比例偏低，需加强政策支持；采用了机器学习算法构建了中国大学生就业选择的有效预测模型；发现了经济因素对地方吸引大学生强度的影响机制。本研究的结果对理解贫困学生就学、就业特征，理解贫困对学生的影响机制，提高减贫政策的精准性有重要价值。