@article{Matayoshi_Karumbaiah_2024, title={Analyzing Transitions in Sequential Data with Marginal Models}, volume={16}, url={https://jedm.educationaldatamining.org/index.php/JEDM/article/view/570}, DOI={10.5281/zenodo.12179681}, abstractNote={<p>Various areas of educational research are interested in the transitions between different states—or events—<br />in sequential data, with the goal of understanding the significance of these transitions; one notable example<br />is affect dynamics, which aims to identify important transitions between affective states. Unfortunately,<br />several works have uncovered issues with the metrics and procedures commonly used to analyze<br />these transitions. As such, our goal in this work is to address these issues by outlining an alternative<br />procedure that is based on the use of marginal models. We begin by looking at the specific mechanisms<br />responsible for a recently discovered statistical bias with several metrics used in sequential data analysis.<br />After giving a theoretical explanation for the issue, we show that the marginal model procedure appears<br />to adjust for this bias. Next, a related problem is that the common practice of removing transitions to<br />repeated states has been shown to have unintended side-effects—to account for this issue, we develop<br />a method for extending the marginal model procedure to this specific type of analysis. Finally, in a<br />recent study evaluating the problem of multiple comparisons and sequential data analysis, the Benjamini-<br />Hochberg (BH) procedure, a commonly used approach to control for false discoveries, did not perform<br />as expected. By applying a technique from the biostatistics and epidemiology literature, we show that the<br />performance of the BH procedure, when used with the marginal model method, can be brought back to its<br />expected level. In all of our analyses, we evaluate the proposed method by both running simulations and<br />using actual student data. The results indicate that the marginal model procedure seemingly compensates<br />for the problems observed with other transition metrics, thus resulting in more accurate estimates of the<br />importance of transitions between states.</p>}, number={1}, journal={Journal of Educational Data Mining}, author={Matayoshi, Jeffrey and Karumbaiah, Shamya}, year={2024}, month={Jun.}, pages={197–232} }