生成研究数据来源清单

以下数据来源在本科生留存率、缺课预警与课程通过率研究中具有较高权威性与实用价值。各来源涵盖的粒度与可用指标不同，建议根据研究问题组合使用。

1) 美国国家教育统计中心（NCES）—综合高等教育数据系统（IPEDS）及追踪类调查（BPS/NPSAS）
- 主管与覆盖：NCES；覆盖美国Title IV授予资助资格的高等教育机构。
- 适用研究：院校层面的留存率（秋到秋）、毕业/完成率；学生层面持久在读/辍学轨迹（BPS）。
- 关键变量/粒度：IPEDS为机构层面；BPS为学生层面纵向（需受限使用许可）。不提供课程/模块通过率或出勤日志。
- 访问与限制：IPEDS年度公开数据与工具；BPS/NPSAS提供公开汇编表与受限微观数据访问。
- 官方入口：nces.ed.gov/ipeds；nces.ed.gov/surveys/bps；nces.ed.gov/surveys/npsas

2) National Student Clearinghouse Research Center（美国全国学生信息中心研究院，NSCRC）
- 主管与覆盖：非营利数据汇聚机构；覆盖美国绝大多数高校的注册与学位记录（参与机构上报）。
- 适用研究：全国/州际留存与“持续在读”率、转学与跨校路径，补足单校数据局限。
- 关键变量/粒度：学生跨机构纵向注册与完成信息；不含课程通过率与细粒度出勤数据。
- 访问与限制：发布Signature Reports等汇总报告；StudentTracker为机构服务（合约与合规要求）。
- 官方入口：nscresearchcenter.org

3) 英国高等教育统计局（HESA）与监管公开数据（含OfS）
- 主管与覆盖：HESA（全英）；与英国学生事务办公室（OfS）相关指标。
- 适用研究：本科非持续率（non-continuation）、进程/完成、分组对比（性别、年龄、背景等）。
- 关键变量/粒度：机构与学科层面年度指标；课程/模块通过率通常不公开，需机构内部数据支持。
- 访问与限制：HESA Open Data与定制数据；受限微观数据需数据共享协议；OfS提供英格兰院校监管指标。
- 官方入口：hesa.ac.uk；officeforstudents.org.uk/data-and-analysis

4) 澳大利亚教育部—高等教育统计（含“学生进程/成功/留存”指标）
- 主管与覆盖：Australian Government Department of Education；覆盖澳大利亚高等教育提供者。
- 适用研究：本科留存、流失与“单位（unit）成功率/通过率”（课程通过的可评估学分占比），以及院校与学科分组。
- 关键变量/粒度：机构与学科层面年度进程（progress）、成功（success）、留存（retention）、流失（attrition）指标。
- 访问与限制：年度统计报表与数据下载；对课程级别原始记录通常不公开。
- 官方入口：education.gov.au（Higher Education Statistics；Student progress/success/attrition 数据表）

5) Open University Learning Analytics Dataset（OULAD）
- 主管与覆盖：英国开放大学（The Open University）；公开研究数据集。
- 适用研究：缺课/缺勤预警与早期学业风险识别、课程成绩预测、通过率相关建模。
- 关键变量/粒度：学生人口学与评测数据、VLE（虚拟学习环境）交互日志（资源访问、点击次数、时间序列），课程与呈现期信息；可构建出勤代理变量与行为特征。
- 访问与限制：公开可下载（研究用途）；为单一机构样本，外推需谨慎并进行情境化验证。
- 官方入口：https://analyse.kmi.open.ac.uk/open_dataset

补充说明与使用建议
- 课程通过率与缺课预警通常需要细粒度教学过程数据（SIS/EMS/LMS、课堂考勤、作业与测验时间戳、登录/点击日志）。此类微观数据多由机构内部持有，国家级公开数据很少提供。可将上述宏观权威来源（留存/完成）与校内微观数据（出勤/互动/成绩）结合，开展多层次分析与早预警建模。
- 隐私与合规：使用学生层面数据时应遵循适用法规（如FERPA、GDPR）与伦理审查要求；跨来源链接需进行去标识化与最小化处理，并严格控制访问权限与审计记录。

- OECD UOE Data and Education at a Glance (OECD.Stat)
  - Scope and comparability: Harmonized cross-national indicators based on the joint UNESCO-OECD-Eurostat (UOE) data collection, ensuring methodological consistency across participating countries over time.
  - Coverage and unit: Country-level (with some subnational modules in select OECD databases); annual or biennial updates.
  - Relevant indicators: Education expenditure by level and source, expenditure per student, teacher numbers and pupil–teacher ratios, teacher compensation, instructional time, and staffing structure.
  - Use for this topic: Quantifies funding policies and staffing patterns across countries; can be linked to external sources providing urban–rural context.
  - Limitations: Urban–rural breakdowns are generally not included; analyses of intra-country disparities require merging with other datasets.

- UNESCO Institute for Statistics (UIS) Education Data
  - Scope and comparability: Global coverage using standardized definitions, enabling comparisons across low-, middle-, and high-income countries.
  - Coverage and unit: Primarily national aggregates; annual series with long time spans for many indicators.
  - Relevant indicators: Government education expenditure (total and by level), expenditure as a share of GDP and government outlays, numbers of teachers, pupil–teacher ratios, enrollment, and completion.
  - Use for this topic: Tracks funding trends and staffing capacity globally, facilitating comparative analysis of policy shifts across diverse systems.
  - Limitations: Urban–rural disaggregation and detailed staffing distribution within countries are limited; often requires complementary microdata for within-country disparities.

- PISA Student and School Microdata (OECD)
  - Scope and comparability: Standardized international assessment with harmonized student and school questionnaires across economies every three years.
  - Coverage and unit: Student- and school-level microdata; nationally representative samples, with school weights.
  - Relevant indicators: School location/urbanicity categories, indices of resource and staffing shortages, school governance and autonomy, sources of school funding, and student performance outcomes.
  - Use for this topic: Enables analysis of associations between resource constraints, funding sources, staffing adequacy, and urban–rural disparities in learning outcomes; supports multi-level modeling across countries.
  - Limitations: Does not provide audited finance totals; staffing measures are largely perception- or index-based rather than administrative counts.

- TALIS Teacher and Principal Microdata (OECD)
  - Scope and comparability: International survey of teachers and principals with harmonized instruments across countries and cycles.
  - Coverage and unit: Teacher- and school-level microdata; nationally representative samples.
  - Relevant indicators: Teacher workforce characteristics (experience, qualifications, workload), hiring and retention challenges, leadership and HR practices, professional development, school location categories, and perceived resource needs.
  - Use for this topic: Strong for examining staffing allocation, recruitment difficulties, and working conditions across urban–rural settings; complements PISA by focusing on workforce and management rather than student testing.
  - Limitations: Limited direct financial accounting; linking to expenditure data requires merging with national accounts or administrative data.

- World Bank BOOST Subnational Expenditure Databases
  - Scope and comparability: Standardized, program-classified (COFOG-compatible in many cases) government expenditure records at subnational levels for participating countries, enabling within-country comparisons over time.
  - Coverage and unit: Administrative unit–level finance data (e.g., provinces/districts), often with multi-year panels; country coverage varies.
  - Relevant indicators: Detailed education spending by function and economic category, enabling identification of policy shifts (e.g., grants, formula changes) and tracking of resource flows to localities.
  - Use for this topic: Supports quasi-experimental evaluations of funding policy effects and urban–rural disparities by linking expenditures to local characteristics and education outcomes or staffing from EMIS/household surveys.
  - Limitations: Country participation and years covered differ; requires integration with separate datasets to obtain teacher counts and precise urban–rural classifications.

以下为可用于研究生源结构、就业去向与专业吸引力的五类潜在数据来源（需依据法律法规与数据治理规范依法合规获取与使用）：

1) 校内招生、学籍与教务管理系统数据
- 典型字段：生源地/省份、城乡属性、性别、家庭背景（如是否家庭经济困难）、高考/入学成绩与位次、录检志愿与志愿顺序、调剂情况、专业分流去向、转专业记录、课程修读与成绩等。
- 适用维度：生源结构（画像、结构变动）、专业吸引力（志愿顺序、报录比、调剂率、转专业流向）。

2) 教育主管部门与考试招生公开数据
- 来源示例：教育部阳光高考信息公开、各省级教育考试院发布的招生计划、投档线/位次、选科要求，教育统计年鉴与各省教育年鉴等。
- 适用维度：生源结构（省域生源规模与结构对比）、专业吸引力（投档线差、报录比、区域竞争度），为校内数据提供外部基准。

3) 毕业生去向与就业质量调查数据
- 典型来源与载体：学校年度毕业生就业质量报告、毕业去向调查问卷与访谈、用人单位回访记录、校友跟踪调查。
- 典型字段：去向分类（签约就业、升学、出国、创业、灵活就业等）、行业与职业类别、单位性质与规模、就业地区、起薪与薪酬分位、中位签约时间、与所学专业相关度、满意度。
- 适用维度：就业去向（流向结构与质量）、专业吸引力（毕业生对专业和就业结果的主观评价与客观结果）。

4) 政府劳动力与就业管理数据
- 可能来源：人力资源和社会保障部门发布的高校毕业生就业与招聘统计、国家统计局劳动力与薪酬统计、公共就业服务平台数据；经授权的数据共享如社保参保与税务就业登记信息等。
- 适用维度：就业去向（实名制去向与留岗稳定性、行业与地区迁移）、就业质量（薪酬与用工形式），用于校内调查结果的外部校验与长期追踪。

5) 志愿填报与招生咨询行为数据
- 可能来源：第三方志愿填报与咨询平台的匿名化行为数据、学校官网与招生专题页/新媒体的访问与转化数据、线上线下招生咨询与宣讲活动登记、夏令营与开放日报名记录。
- 典型指标：专业页访问量、咨询转化率、意向专业偏好、地域来源热度、活动参与与转化链路。
- 适用维度：专业吸引力（报考兴趣与转化漏斗）、生源结构（潜在生源地域与学校类型分布的前置信号）。