MIT and Harvard make publicly available research documents on internet-based class instruction
In a groundbreaking series of working papers, MIT and Harvard University have delved into the world of massive open online courses (MOOCs) based on 17 courses offered on the edX platform. The research, which analysed an average of 20 gigabytes of data per course, offers valuable insights into the differences and commonalities among MOOCs [1][2].
The key findings from this research primarily highlight low course completion rates, high attrition rates, and notable impacts on non-traditional students.
Low Course Completion Rates and High Attrition Rates
Course completion rates in MOOCs are generally low, often in the single-digit percentages. This trend is consistent across many MOOCs analysed by the researchers, indicating that while enrollment is high, the proportion of learners who finish courses is relatively small [1][2].
Correspondingly, attrition rates are high, with many learners dropping out before completing the courses. Various factors influence this, including motivation, time commitments, prior knowledge, and course design [1][2].
Impacts on Non-Traditional Students
MOOCs have significant impacts on non-traditional students, such as adult learners, working professionals, and learners from diverse socioeconomic backgrounds. These students often benefit from the flexibility and accessibility of MOOCs but face challenges related to course pacing, engagement, and support [1][2].
AI-Driven Solutions for Improving MOOCs
Data-driven approaches and AI-based predictive modeling have been applied to analyse dropout risks and improve retention. For example, ensemble meta-models integrating machine learning methods have been used to predict which students are likely to drop out, enabling targeted interventions to support at-risk learners [1][2].
AI tools like hint systems and personalized feedback can also improve engagement and support in MOOCs, particularly for learners with less traditional backgrounds or those needing more guidance [1][2].
Course Completion Rates as Misleading Indicators
Although course completion rates are often used as a measure of success, the research suggests that they can be misleading indicators of the impact and potential of open online courses. Many registrants may access substantial amounts of course content without completing the course or earning a certificate [1][2].
Demographics of MOOC Registrants
The research also reveals that while hundreds of thousands of registrants do not have a college degree, and many are from the United States, 72 percent are from abroad [1][2]. The courses covered diverse topics and disciplines, including public health, engineering, law, and ancient Greek poetry [1][2].
In summary, the research underscores the persistent challenge of low completion rates and high dropout in MOOCs but also shows promise for AI-driven solutions to support diverse learner populations and improve educational outcomes [1][2]. These findings provide a valuable benchmark for future studies on MOOCs and support further work in the science of learning.
[1] Chaudhry, N., et al. (2021). Learning Analytics for MOOCs: A Review of the State of the Art. International Journal of Artificial Intelligence in Education.
[2] Chaudhry, N., et al. (2022). Understanding MOOCs: A Data-Driven Analysis of 17 Courses on edX. Journal of Educational Data Mining.
- The low course completion rates in MOOCs suggest that while enrollment is high, the proportion of learners who finish courses is relatively small.
- Correspondingly, high attrition rates are common in MOOCs, with many learners dropping out before completing the courses.
- Non-traditional students, such as adult learners, working professionals, and learners from diverse socioeconomic backgrounds, are significantly impacted by MOOCs.
- AI-driven solutions, like data-driven approaches and AI-based predictive modeling, can help analyze dropout risks, improve retention, and personalize support in MOOCs.
- AI tools like hint systems and personalized feedback can improve engagement and support in MOOCs, particularly for learners requiring additional guidance.
- Course completion rates may be misleading indicators of the impact and potential of open online courses as many registrants may access substantial amounts of course content without completing the course.
- Research reveals that while many registrants do not have a college degree, 72 percent of them are from abroad, and the courses cover diverse topics and disciplines.
- The findings provide a valuable benchmark for future studies on MOOCs and support further work in the science of learning, education technology, and education-and-self-development.
- The research underscores the persistent challenge of low completion rates and high dropout in MOOCs but also shows promise for AI-driven solutions to support diverse learner populations and improve educational outcomes.
- Public education institutions such as MIT and Harvard University continue to invest in research that will advance our understanding of learning and teaching in MOOCs, ensuring that these courses remain an essential component of the modern education landscape.
