The area of learning analytics is often viewed as a tool for supporting quantitative analysis. Based on previous research, this association between quantitative analysis and learning analytics does seem to be the trend. However, certain researchers have proposed the use of multimodal learning analytic techniques as a viable and valuable contribution to more qualitative research methodologies. This paper examines that idea by trying to use the output from an algorithm that learns discriminating features, as the starting point for video observations. Ultimately, the analysis suggests that there is utility in leaning on machine learning to help identify important patterns in the data, provided that those patterns are contextualized and studied using the original video data. Additionally, the work makes clear the need for better tools for conducting these types of multimodal analyses.
|Original language||English (US)|
|Title of host publication||Proceedings of the 2018 International Conference of the Learning Sciences|
|Number of pages||4|
|State||Published - 2018|