This systematic review provides a meta-analytical look at five years of quantitative MOOC research. By synthesizing 126 studies, the authors map out the psychological and technical drivers of learner success and introduce "Factors-Goals (F-G) graphs" as a new framework for designing future empirical studies.
The MOOC Paradox: High Access, High Attrition
Massive Open Online Courses (MOOCs) represent a paradox in modern education: they offer unprecedented access to knowledge but suffer from notoriously high dropout rates. While thousands of researchers have investigated why students quit or how they perform, the literature is fragmented across various theoretical models and methodologies. This study addresses the lack of a unified "map" for questionnaire-based research, identifying the most effective models and the specific factors—such as self-efficacy or system quality—that correlate with learner outcomes.
Methodology and Theoretical Framework
The authors conducted a systematic review of 126 articles retrieved from the Science Citation Index (SCI) and Social Sciences Citation Index (SSCI) published between 2015 and 2020. A central finding is that researchers rarely use "off-the-shelf" models. Instead, they extend classical frameworks like the Technology Acceptance Model (TAM) or Self-Determination Theory (SDT) by adding niche factors such as "network externalities" or "social recognition."
Visualizing Research: The Factors-Goals (F-G) Graph
The paper’s primary methodological contribution is the introduction of Factors-Goals (F-G) graphs. These graphical summaries visualize the complex web of correlations between research goals (e.g., reducing dropout rates) and influencing factors (e.g., perceived usefulness). This approach moves beyond simple lists of variables to show the structural hypotheses common in the field.
As shown in the F-G graph above, "Continuance Intention" is driven by a mix of subjective factors (learner preference, expectancy) and objective factors (course quality, system stability), drawing from multiple theories like UTAUT and SDT.
Key Insights and Results
The review identifies three primary research clusters:
- Learner Performance (35.71%): What drives engagement and grades?
- Dropout/Continuance Intention (28.57%): Why do students stay or leave?
- Assessment (23.02%): How do students and teachers evaluate the platform?
Methodologically, while descriptive statistics are common (48.41%), Structural Equation Modeling (SEM) has become the gold standard for testing complex hypotheses, appearing in over 42% of the reviewed studies. Software like IBM SPSS and AMOS remain the dominant tools for these analyses.
Limitations and Future Directions
Despite the growth in research, several "bottlenecks" persist:
- Small Sample Sizes: Many studies lack the statistical power for universal conclusions.
- Traditional Statistics Bias: Reliance on "snapshot" surveys may miss the dynamic nature of learning.
- Lack of Diversity: Participants are often clustered in specific geographic or institutional silos.
To overcome these, the authors recommend integrating "big data" (clickstream data), standardizing data across platforms, and applying AI and deep learning to move from describing behavior to predicting it with personalized interventions.
Summary updated