How Anxiety Affects Affect: A Quantitative Ethnographic Investigation Using Affect Detectors and Data-Targeted Interviews

This study integrates the analysis of student interaction data with classroom interviews in order to better understand how students’ trait-level anxiety relates to in-the-moment measures of students’ affective experiences (i.e., boredom, confusion, delight, engaged concentration, and frustration). We first use quantitative data to drive the data collection, with the interviews being triggered by previously-developed models of student emotions, allowing us to focus interviews during times where specific emotional experiences of interest have just occurred. We then analyze the log data, finding differences in how students with high and low science anxiety manage their emotions (and subsequent behaviors). Finally, we connect these behaviors back to the strategies students articulate in the interviews, finding that many students with higher anxiety scores have an external locus of control, are less likely to seek evidence that they are wrong, and are more likely to make major changes to their solutions in response to their frustration.