EAGER: The Use of Institutional Data to Identify Predictors of STEM Degree Completion

摘要：拉美裔美国人现在是四年制学院和大学入学的最大少数群体；然而，学位完成率仍然低于非西班牙裔白人和亚裔美国人。学生在攻读科学、技术、工程和数学(STEM)领域的本科学位方面的坚持性和成功程度，并不是简单地通过先前的学业准备和考试成绩就能很好地预测的，而且，那些提高留住能力的经验或素质可能会因种族或民族和学科的不同而有所不同。这项研究从一所服务于拉美裔的大型公立大学获得了非常详细的机构数据，该大学拥有大量历史上代表性不足的少数民族人口。这些数据有可能通过种族/民族发现在学生意向、留住和完成STEM学位课程的各种因素的相对重要性方面可能存在的差异。这个试点项目探索了利用现有的、大规模的大学机构数据来确定可能有助于代表不足的少数族裔更多地参与STEM领域的因素的可能性。大规模的分析数据集，包括关于大学前准备、经济援助、课程注册和完成情况以及学生和教职员工人口统计数据的如此多样化的数据，是很少见的。该项目将从多个数据源中提取、合并和重新编码与学生和课程相关的信息，以产生新的纵向、学生级别的分析数据文件，允许检查学生特征、课程准备和排序以及教学环境如何相互作用，以促进STEM学位的完成。利用新的分析文件进行探索性描述性分析将评估采用更复杂的计量经济学模型和基于模拟的方法的可行性和适当性，以确定和评估政策干预措施，以增加未被充分代表的少数群体在STEM劳动力中的参与。

Abstract: Hispanics are now the largest minority group enrolling at four-year colleges and universities; however, degree completion rates remain lower than among non-Hispanic White and Asian Americans. Student persistence and success in pursuit of undergraduate degrees in science, technology, engineering and mathematics (STEM) fields are not well-predicted simply by prior academic preparation and test scores, and furthermore, those experiences or qualities which enhance retention may differ both across racial or ethnic groups, and across disciplines. This study develops highly detailed institutional data from a large Hispanic-Serving public university with a substantial population of historically underrepresented minorities. These data have potential to detect possible differences by race/ethnicity in the relative importance of various factors in student intentions, retention, and completion of STEM degree programs.This pilot project explores the potential for using existing, large-scale institutional data at the university level to identify factors that may contribute to greater participation of underrepresented minorities in STEM fields. Large-scale analytic datasets including such diverse data on pre-college preparation, financial aid, course enrollments and completions, and demographics of students and faculty are rare. This project will extract, combine, and recode student- and course-related information from multiple data sources to produce new longitudinal, student-level analytic data files, permitting examination of how student characteristics, course preparation and sequencing, and instructional environments interact to facilitate STEM degree completion. Exploratory descriptive analyses utilizing the new analytic files will assess feasibility and appropriateness of applying more sophisticated econometric modeling and simulation-based approaches, towards identifying and evaluating policy interventions to increase participation of underrepresented minorities in the STEM workforce.