Early Detection of At-risk Students – Then What?

According to the National Center for Education Statistics (NCES, 2022), only 64% of undergraduate students starting at a four-year institution in 2014 completed their degree within six years. Temple University, like all higher eduction institutions, is continually looking for ways to boost student retention, academic progress, and graduation rates.

Temple has a long-standing policy of issuing midterm progress ratings for all undergraduates. Students receive feedback from their instructors on whether their work is Satisfactory (S), or Unsatisfactory (U) for reasons of Attendance (A), poor Grades (G), Missing assignments (M), or insufficient Participation (P), or a combination of factors. Thus, a student receiving a U-AM has poor attendance and missing assignments. Colleges follow up with emails to struggling student to encourage them to seek assistance from instructors, advisors and the many support services available to Temple undergraduates.

A logistic regression analysis showed that midterm progress ratings are highly predictive of final grades. Students with receiving a U rating for one or more reasons were 14 times more likely to receive an unsatisfactory final grade or withdraw from the course. Are midterm progress ratings and subsequent messaging ineffective, or do the warnings simply come to late for struggling students to turn it around?

In a poster I will present at Temple University’s 22nd Annual Faculty Conference on Teaching Excellence (January 10 & 11, 2024), I will discuss a study conducted while I was Director of General Education at Temple to see: (1) whether Canvas LMS data could be used to identify at-risk students earlier in the semester, and if so, (2) whether issuing the same warning earlier would improve student outcomes. Spoiler alert: using Canvas data it proved possible to detect at-risk students as early as weeks 3-4, however, reaching out to these students earlier did not improve outcomes. The literature on the use of big data, machine learning algorithms, and AI to detect as-risk students is growing exponentially, while research on effective intervention lags sadly behind (Motz et al, 2023).


Motz, B., Bergner, Y., Brooks, C., Gladden, A., Gray, G., Lang, C., … & Quick, J. (2023). Lak of direction. Journal of Learning Analytics, 1-13. https://doi.org/10.18608/jla.2023.7913

National Center for Education Statistics. (2022). Undergraduate Retention and Graduation Rates. Condition of Education. U.S. Department of Education, Institute of Education Sciences. Retrieved [19 Dec. 2023], from https://nces.ed.gov/programs/coe/indicator/ctr.

This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply