Outcome Predictions of Students in Massive Open Online Courses (OPSMOOC)

这一探索性项目旨在开发评估大规模在线公开课(MOOC)课程的方法。MOOC专门为网上互动学习而设计，在学生的时间承诺和时间表方面具有相当大的灵活性，在改善接受高等教育的机会方面具有巨大的潜力。然而，到目前为止，MOOC往往表现出很高的流失率。由于入学人数有数万或数十万，即使是提高毕业率的微小变化，也可能对受益于这种新教学形式的学生的绝对数量产生重大影响。这个项目涉及对一门非常受欢迎的哈佛MOOC的深入研究，该课程通过edX平台提供。它将原始数据(关于参与课程的学生的特征和他们参加课程的动机)与与学生课程活动有关的数据结合在一起进行分析(包括描述性统计、危险建模和多元线性回归)，以探索传统课程结果(例如，学生成绩的衡量标准)和持久性/自然减退。通过在开发MOOC的早期关注课程完成的衡量标准，该项目有可能引导人们关注MOOC的那些方面，这些方面似乎对学生的成功和他们对STEM领域的持续兴趣至关重要。预计研究结果和分析方法都可以纳入教学平台，以帮助教师监测困难和快速改进课程，而且在更统一的数据收集方面的进展将允许对各种MOOC进行比较，特别是那些具有相似目标或内容领域的MOOC。

This exploratory project seeks to develop methods to evaluate Massive Open Online Course (MOOC) offerings. Designed specifically for interactive study over the web, with considerable flexibility in terms of students' time commitment and scheduling, MOOCs have tremendous potential for improving access to higher education. To date, however, MOOCs have tended to exhibit high attrition rates. With enrollments in the tens or hundreds of thousands, even modest changes that increase completion rates could have a major impact on the absolute numbers of students benefiting from this new form of instruction. This project involves in-depth examination of a highly popular Harvard MOOC, offered through the edX platform. It combines primary data (on participating students' characteristics and their motivations for taking the course) with data pertaining to students' course activities in analyses (including descriptive statistics, hazard modeling, and multiple linear regression) to explore both traditional course outcomes (e.g., measures of student achievement) and persistence/attrition. By focusing on the metric of course completion early in the development of MOOCs, this project has potential to direct attention to those aspects of MOOCs that appear critical to student success and their continuing interest in STEM fields. It is anticipated that both research findings and methods of analyses can be incorporated into the instructional platform to aid instructors in monitoring difficulties and in rapid course improvement and that progress toward more uniform data collection will allow comparisons across a variety of MOOCs, particularly those with similar goals or content domains.