Poster #52 - Method Variance and the Direct and Indirect Contribution of Emotional Knowledge to Early Academic Success

• In Event: Poster Session 16

Abstract

Introduction. Kindergarten readiness is marked by social-emotional and cognitive competencies with research showing emotional competence–success with identifying, expressing, and regulating emotions–contributes to early math, literacy, and language skills in complex ways (Denham et al., 2012; Franco et al., 2017; Trentacosta & Izard, 2006). Several studies show Social and Emotional Learning Behaviors (i.e., SELB; e.g., classroom-based engagement, persistence, motivation, social competence, aggression, anxiety) mediate associations between Emotion Knowledge (EK) and early academic success (Denham & Brown, 2010; Denham et al., 2013; Franco et al., 2017; Voltmer & von Salisch, 2017). Although fundamental to construct validity (Campbell & Fiske, 1959), no studies of EK have partialed out systematic method variance despite estimates suggesting it accounting for over 50% of the total variance (Cote & Buckley, 1987). We hypothesize coefficients will be weaker when EK method variance is partilated out with corresponding weaker evidence for SELBs as mediators. Methods. Children (n=1,424) in 209 classrooms at 65 preschools were assessed in either English or Spanish by research assistants using three measures of EK (Affect Knowledge Test-Shortened [AKT-S], Denham et al., 2013; Emotion Matching Task [EMT], Morgan et al., 2009; adapted Box Task [A-Box], Russell & Widen, 2002) as well as language proficiency (i.e., preLAS; Duncan & De Avila, 1998) and early math and literacy (Woodcock Johnson Applied Problems and Letter-word Identification, Schrank & Wendling, 2018). Teachers reported on children’s SELBs (i.e., Preschool Learning Behaviors Scales, McDermott et al., 2002; Social Competence and Behavior Evaluation-30, LaFreniere & Dumas, 1996). Children were between 32 and 68 months and were identified by parents as 45% Hispanic/Latine, 49% White, 28% Black/African American, and 20% Multiracial or other background. Median family income was $42,000, and about one third of families held a high school diploma or less. Factor scores were output from a multi-trait, multi-method confirmatory factor analysis where latent variables representing EK methods were included alongside EK variables representing emotion recognition, understanding, and labeling (see Authors et al., 2022; Koch et al., 2020). Measurement models for EK (⍵=.59), SELBs (⍵=.87), and early academic success (⍵=.78) were created. Paths among these constructs were explored while controlling for covariates. Two mediation models were run: one where EK methodology was partialled out and one where EK methodology was included. Results. All models fit the data well. When EK methodology was partiled out, coefficients were between 46 and 1.10% weaker. Regarding mediation, the direct association between EK and early academic success decreased from β=0.31 to β=0.29 (ps<.001) when SELBs were included. Thus, the association between EK and early academic success was partially mediated by SELBs with a small indirect effect of β=0.03, p=.006. The association between EK and SELBs was weak, β=0.14, p<.001, as was the association between SELBs and early academic success, β=0.19, p<.001. Discussion. Results raise concerns when studying psychological constructs in young children with coefficients being weaker than previously thought in areas other than between EK and early academic success.

Authors

• Craig S. Bailey, Yale School of Medicine
  Presenting Author

• Maya Ingram

• Navsheen Kaur, Yale University