Exploring institutional classification based on distance education enrollment: A Latent Profile Analysis approach

• In Event: International Perspectives on Digital Platforms: Challenges, Customization, and Institutional Adaptation

Proposal

Introduction

The purpose of this study is to explore the potential classification of higher education institutions using latent profile analysis, focusing on the composition of students enrolled in distance education. The classification of higher education institutions is a critical consideration in the implementation of various policies. For instance, practices and policies may differ based on whether an institution is public or private, and the evaluation criteria may also vary depending on the degree award level. Meanwhile, one of the areas gaining significant attention in the field of higher education is distance education. However, the majority of research focuses on students' views on distance education, its effectiveness, and its evolution, with limited studies on how higher education institutions are classified based on their distance education practices (Dede, 1996; Ortiz-Rodríguez et al., 2005; Tümen Akyildiz, 2020; Zhao et al., 2005). It has the potential to bring about changes not only in traditional on-campus student engagement and the interaction between instructors and students but also in various aspects of higher education institutions, such as housing. Therefore, this study aims to contribute to the expansion of knowledge on this topic by attempting to classify higher education institutions based on the characteristics of the proportions of students who are exclusively enrolled in distance education and those who are partially enrolled in distance education.

Conceptual Framework

Taylor and Cantwell (2018) utilized latent profile analysis to examine the higher education landscape and argued in their work, unequal higher education in the United States with growing participation and shrinking opportunities. This study also follows a similar conceptual framework and aims to extend beyond traditional higher education classifications by analyzing the higher education landscape in relation to distance education practices. Consequently, the research question is: How can institutions be classified according to the patterns of student enrollment, specifically distinguishing between those enrolled exclusively in distance education courses and those participating in some distance education courses, across the categories of undergraduate non-degree/certificate students, undergraduate degree-seeking students, and graduate students at each institution?

Methodology

To answer the research question, I employed latent profile analysis (LPA) using MPlus software (Geiser 2013; Muthen and Muthen 2011). This statistical method is designed to identify groups within the data that can be conceptually inferred but are not directly observable, making it particularly suitable for analyzing analytically distinct subgroups within a single sample. The U.S. federal government gathers information on higher education institutions via the Integrated Postsecondary Education Data System (IPEDS). I utilized data sourced from IPEDS and the sample included Baccalaureate Colleges or institutions with a higher classification based on Carnegie Classification. To exclude anomalous conditions during the COVID-19 period, the analysis period was defined as fall 2019. The primary variables are the proportions of students enrolled exclusively in distance education courses and those enrolled in some but not all distance education courses, within each group of undergraduate degree-seeking students, undergraduate non-degree/certificate-seeking students, and graduate students.

Preliminary Result

In the latent profile analysis, models with the number of classes ranging from 2 to 5 were compared. While the models with 2 classes showed good fit indices for AIC, BIC, ABIC, and entropy, the LMR LRT had a p-value of 0.15, which was not statistically significant. However, based on the exploratory nature of this study, models with 3, 4, and 5 classes were also analyzed. The results indicated that the model with 4 classes had favorable fit indices for AIC (-6764.036), BIC (-6601.752), ABIC (-6706.562), and Entropy (0.972), and the LMR LRT p-value was 0.021, which was statistically significant. Examples of distinctive cases among the latent profile classes include groups where, within each category of undergraduate degree-seeking students, undergraduate non-degree/certificate-seeking students, and graduate students, there are cases where the proportion of students enrolled exclusively in distance education courses is higher than the proportion of those enrolled in some distance education courses. On the other hand, another class exhibited a pattern where the proportion of non-degree/certificate-seeking students enrolled in some distance education courses exceeded 50%. Another class showed a pattern where, across all student levels, the proportion of students enrolled exclusively in distance education courses and the proportion of students enrolled in some distance education courses were both below 40%.

Conclusion

This study explored the classification of higher education institutions based on student enrollment in distance education using latent profile analysis. By analyzing data from Baccalaureate Colleges and above, the study identified distinct patterns in the proportions of students enrolled exclusively versus partially in distance education courses. The lack of statistical significance in the two-class model and the presence of classes with proportions below 5% are limitations of this study. Future research should address these limitations by diversifying the analysis period and exploring additional methodological refinements. The analysis revealed that a four-class model provided the best fit, highlighting significant differences among institutions. These findings could suggest that higher education institutions can vary in their distance education offerings, which could inform policy decisions and institutional practices. This study contributes to understanding how distance education has potential to impact the institutional classification and the broader landscape of higher education.

Author

• Sohee Park, Michigan State University