Examining Relations Between Self-Regulation and Temperament: A Multivariate Approach

• In Event: Poster Session 13
  In Poster Session: PS 13 Topic - Social, Emotional, Personality

Abstract

Self-regulation, a multidimensional construct involving an interplay between children’s internal resources and socialization experiences (Kopp, 1982), emerges during early childhood (Calkins, 2007). Theoretical models of temperamental reactivity and regulation (Rothbart et al., 2011) describe individual differences in children’s dispositions to enact prepotent tendencies, i.e., react strongly to situations (e.g., Negative Affectivity, Surgency) and use internal resources to inhibit such action (Effortful Control; Rothbart & Bates, 2006). These individual differences should account for differences in children’s actual behavior during tasks known to elicit prepotent responses. Yet, evidence of associations between parent-reported child temperament and children’s observed behaviors is mixed, perhaps because the evidence is mostly generated from analysis of one or few behavioral indices, e.g., anger or a particular strategy (e.g., Hayden et al., 2005). We use an explicitly multivariate approach, examining whether temperament manifests in multidimensional profiles of emotions and strategy use during challenging tasks.

Our person-oriented approach uses modern mixture (e.g., Latent Profile Analysis) models to identify groups of children (N=158; 49% female; age 30 to 60 months) based on their emotions (happy, sad, angry, fear) and behaviors (approach, withdrawal, engagement of executive processes [EP]) during a frustrating wait and a novel situation. We examined if and how these groups differ with respect to parent-reported Negative Affectivity, Surgency, Effortful Control; CBQ-VSF; Putnam & Rothbart, 2006). In parallel, we used a variable-oriented approach with modern data mining (e.g., regression trees) to identify specific combinations of emotions and behaviors that are most predictive of each dimension of temperament.

The simple aggregates (Ms, SDs) of children’s emotions and behavior (Figure 1a) provided identification of 5 distinct profiles (Figure 1b). However, generating latent profiles based on simple mean-level aggregates we found no hypothesized relations; greater surgency is not differentiated by higher positive emotion / higher approach / lower withdrawal (Dollar & Stifter, 2012); greater negative affectivity is not differentiated by higher negative emotion / higher approach to restricted objects / higher withdrawal from novel objects / lower engagement of executive process / lower positive emotion (Moran et al., 2013); and effortful control is not differentiated by lower negative emotion, lower withdrawal (distraction) from restricted objects / lower approach to novel objects / higher engagement of executive process (Tan et al., 2013); all Fs < 1.13, R2s < .03. See optimized regression tree model (Figure 1c for Negative Affectivity) provided only slight improvements.

The absence of results from analyzing typical emotion and behavior aggregates underscores the need for more temporally and change-oriented variables when mapping differences in all three dimensions of temperament to actual behavior. For this poster, we will conduct analyses using more dynamic temporal variables (e.g., latency and duration of approach, inertia of emotion, peak executive effort) that may better represent the relation of temperament to actual behavior. Merging data science and developmental science, we will expand our analysis by calculating 100+ more features of each child’s behavior (Roque & Ram, 2019) and identifying configurations of the emotion and behavioral dynamics involved in self-regulation and how they are actually related to temperament.

Authors

• Jaclyn Yuro, Pennsylvania State University, University Park
  Presenting Author

• Breana Genaro, Pennsylvania State University
  Non-Presenting Author

• Tobi Quadri, Pennsylvania State University, University Park
  Non-Presenting Author

• Pamela M Cole, Pennsylvania State University, University Park
  Non-Presenting Author

• Nilam Ram, Stanford University
  Non-Presenting Author