Replication files for "The Impact of Money on Science."

This dataset replicates tables and figures in the article "The Impact of Money on Science." The paper measures the productivity of university research investments and asks whether the resulting pools of knowledge create valuable, downstream technology (or simply accumulate in the “ivory tower”). It uses unexpected NCAA athletic outcomes to vary research support to university faculty and estimate knowledge productivity. We find positive, significant effects of research expenditures on articles published and patents filed. Then, using data on university technology licensing income, we show that these investments produce large returns in real terms. The data consists of an instrument, inputs, and outputs (from the knowledge production process). -The instrument consists of Associated Press college football team rankings. -The inputs the knowledge production function consist of the following: * research support expenditures (downloaded from the National Science Foundation) * faculty headcounts and expenditures (also downloaded from the National Science Foundation) * physical capital, i.e. facility expenditures (downloaded from IPEDS) - The outputs measure patenting, publishing, and licensing activity * Patent data comes from PatentsView.org (http://www.patentsview.org) and the USPTO Office of the Chief Economist. * Publishing activity is measured by academic publications, mainly in the ``STEM'' fields * University technology licensing revenues are surveyed and published by Association of University Technology Managers (``AUTM'') More details on the construction of these measures can be found in the associated article. Name correspondences are provided in an Excel file (nameMerge.xlsx). Code that combines the data from the subfolders and replicates the relevant tables and figures of the paper is provided in a STATA "do" file (master_final.do). Output files are exported to the "output" subfolder.本数据集复刻了论文"The Impact of Money on Science"（《金钱对科学的影响》）中的表格与图表。该论文旨在衡量大学科研投入的产出效率，并探究由此产生的知识储备是否能催生具有实际价值的下游技术，抑或仅仅停留在"象牙塔"之中。研究利用意外出现的全国大学体育协会（National Collegiate Athletic Association, NCAA）赛事结果作为外生冲击，以此调整大学教员获得的科研支持额度，并据此估算知识生产的效率。研究发现，科研支出对学术论文发表与专利申请均存在显著的正向影响。随后，借助大学技术许可收入数据，本研究进一步证明，此类科研投入能够带来可观的实际收益。本数据集包含工具变量、投入要素与产出结果（均来自知识生产流程）。 - 工具变量为美联社（Associated Press, AP）大学橄榄球排名。 - 知识生产函数中的投入要素包括以下三类： * 科研支持经费（数据取自美国国家科学基金会（National Science Foundation, NSF）） * 教员人数与薪酬支出（同样取自美国国家科学基金会） * 实物资本，即设施经费（数据取自高等教育数据集成系统（Integrated Postsecondary Education Data System, IPEDS）） - 产出指标涵盖专利、学术出版与技术许可活动： * 专利数据来自PatentsView.org（网址：http://www.patentsview.org）与美国专利商标局（United States Patent and Trademark Office, USPTO）首席经济学家办公室。 * 学术出版活动以主要集中在科学、技术、工程与数学（Science, Technology, Engineering, Mathematics, STEM）领域的学术出版物数量衡量。 * 大学技术许可收入由大学技术管理者协会（Association of University Technology Managers, AUTM）开展调查并发布。 有关这些指标构建方式的更多细节可参阅相关论文。变量名称的对应关系可在Excel文件nameMerge.xlsx中查阅。用于合并子文件夹内数据并复刻论文相关表格与图表的代码已封装为STATA格式的"do"文件master_final.do。生成的输出文件均导出至"output"子文件夹中。