A Mixed-Method Approach to Investigating Difficulty in Data Science Education

The purpose of this study was to define a methodology to identify any disconnect between students and instructors in data science classrooms through analyzing qualitative data. A combined qualitative and quantitative approach was used for analysis of survey data from students, faculty/instructors, and teaching assistants across three institutions. Using a manual content analysis paired with a TF-IDF analysis, researchers were able to pull out frequently used terms within responses and encode them into categories and subcategories. Trends were identified from these categories and subcategories to examine general areas of disconnect within the data science classroom. Additionally, a quality analysis was run to determine the effectiveness of the phrasing of the questions posed during the survey. As a whole, the methods used throughout this research process provide direction for researchers in interpretation and analysis of the survey data in an efficient and time-sensitive manner. Furthermore, it allows researchers to analyze the quality of responses to give insight towards rephrasing of survey questions in future analyses. Although the research was applied to data science classrooms, this method has the potential to be applied into other fields and areas of study when performed with coordination between a field expert and a data scientist