ABSTRACT
In recent years, a number of insights have been gained into the cognitive processes that explain how individuals overcome misconceptions and revise their previously acquired incorrect knowledge. The current study complements this line of research by investigating the moment-by-moment emotion processes that occur during knowledge revision using a think-aloud methodology. Undergraduate students read both refutation and nonrefutation texts and reported out loud their thoughts, which were coded along valence and activation dimensions of emotions. Results showed that at key points during reading, emotions differed within and between experimental text conditions. Further, exploratory mediational analysis showed that surprise was an influential emotion for learning. Findings are discussed in terms of theoretical contributions to our basic understanding of the role of emotions during knowledge revision.
Notes
1 We had initially tested the plausibility of adopting a coding scheme of discrete emotions to fit our data, but finding relatively uncommon occurrences of various discrete emotions (e.g., curiosity), which rendered extremely skewed distributions, we opted instead to use a circumplex coding scheme that we determined to be a better fit to the current data.
2 Only one utterance could not be categorized with the current circumplex coding scheme. One participant verbalized, “[It] makes me miss being in Europe, I would like to go back again soon.” We interpreted this to be an expression of nostalgia, which we define as “a feeling of pleasure and also slight sadness when you think about things that happened in the past” (Cambridge Dictionary, 2017). Without additional information, this emotional expression could reasonably be categorized as both positive and negative; thus, we excluded this instance from our analyses.
3 All segments were tested, but only the theoretically relevant segments and baseline are reported (i.e., Segments 1, 2, 3, and 5).
4 Because of some subsections containing zeros for NDEs and surprise, to avoid Hessian matrix singularity, we transformed the data by adding a small positive constant (1) to all values for analysis only; descriptive statistics, planned comparisons, and figures represent original, untransformed data.
5 We thank the editor and reviewers for the suggestion to conduct and incorporate these analyses.