Identifying Students At-Risk in Online STEM Courses

Project No.
PI Name
Claire Wladis
Borough of Manhattan Community College at the City University of New York

EHR Core Research

Abstract 1

Identifying Students At-Risk in Online STEM Courses

Presentation Type
Claire Wladis, Borough of Manhattan Community College and The Graduate Center at the City University of New York Katherine Conway, Borough of Manhattan Community College at the City University of New York Alyse C. Hachey, Borough of Manhattan Community College at the City University of New York


In 2013, over 40 million college students took online classes worldwide; by 2017, that number should triple (Atkins, 2013). Yet little research on the factors affecting retention in online STEM courses exists. Online courses may increase college access, but higher attrition (not yet well understood), could be an obstacle to degree completion. This project investigates which factors predict poorer outcomes online vs. face-to-face for STEM students, particularly for traditionally underrepresented groups. This research departs from previous studies by focusing on online STEM outcomes specifically and by utilizing national and, institutional level data along with qualitative data from faculty and students; and employing rigorous methodological and statistical techniques to adjust for factors that influence selection bias.


Ultimately the goal is to develop a model identifying those students at highest risk of dropping out of online STEM courses (or college subsequently), so that support services can be effectively targeted. We aim to predict differential online versus face-to-face outcomes, and to identify factors relevant to the online environment specifically (as opposed to factors, such as GPA, that may predict outcomes generally across all mediums).


This project draws on general, and online specific, conceptual retention models (Anderson & Kim, 2006; Bean & Metzner, 1985; Kember, 1989; Seymour & Hewitt, 1994; Tinto 1975; 1986; 1993), alongside Bourdieuï¾’s theories (1984; 1986) of capital and habitus, to posit a model of online STEM student retention. Refinements are made to the model through surveys of students and faculty, as well as in-depth interviews with online students. Both quantitative (e.g. multi-level modeling, propensity score matching, sensitivity analysis) and qualitative approaches (e.g. constant comparison analysis, domain analysis, classical content analysis) are employed.


At the symposium we will present preliminary results; the end result of this research will be identification of factors that consistently predict differential online versus face-to-face STEM course (and college) outcomes across a variety of samples, plus an assessment of the effect size of these various factors on online course outcomes. Upon completion, we anticipate issuing recommendations for identifying at-risk students in online courses, so that they can be targeted for support and interventions, including mentoring, tutoring, technical support, advisement, or skills and behavioral training.

Broader Impacts

This research will potentially impact students considering online courses, faculty designing and teaching online courses, staff implementing online student support structures, administrators determining online education guidelines, and policymakers seeking to broaden student access to, and success in, STEM disciplines via online education.

Results are being widely disseminated to several different audiences. Articles and conference presentations aimed at researchers, practitioners and administrators are being submitted to audience-specific peer-reviewed journals and organizations. Policy briefs, written to appeal to a broad audience, will be freely available on the project website. Locally, results are being shared with e-learning directors and institutional research departments across our multi-campus university system, and presentations to faculty in academic departments are planned, in order to educate all faculty about factors that impact student enrollment and success in online STEM courses.

Unexpected Challenges

So far we have not encountered any particularly unexpected challenges.


Wladis, C., Hachey, A. C. & Conway, K. Who Succeeds Online? Using student characteristics to predict online course and subsequent college attrition, submitted for publication

Wladis, C., Conway, K. & Hachey, A.C. Time Poverty and Parenthood: Who Has Time for College?, submitted for publication