The Implications of a Subtle Difference in the Calculation of Affect Dynamics

Abstract

D'Mello and Graesser’s (2012) highly-cited model of affect dynamics proposes a sequence of theoretically-grounded transitions between affective states during learning. However, empirical studies in a range of contexts have not produced the predicted results. Several factors may explain this lack of replication, including the demographics of the populations studied, the degree of authenticity of the learning setting (e.g. classrooms versus laboratory studies), the grain-size of observation, affect data collection procedure, duration of research sessions, and the methodological choices used to analyze transitional patterns. In particular, whereas D'Mello and Graesser (2012) exclude self-transitions (when a student remains in the same affective state across two observations) in calculations of transition probabilities, most other published works in this area, including some of their earlier publications, do not. This paper investigates the impact of this seemingly minor methodological choice by applying both analyses to previously collected data from a study of the Physics Playground system. In particular, this paper investigates whether this difference is sufficient to produce (or suppress) the transitions theorized in D'Mello and Graesser's theoretical model.