Direct and Indirect Links Between Parent-Child Emotion Conversations and Children’s Emotion Knowledge Across Early Childhood

• In Event: The Role of Input in Early Emotion Learning

Abstract

Children’s emotion knowledge is linked to healthy social outcomes (Eisenberg et al., 2005) and is likely influenced by socialization. However, research is somewhat inconsistent as to how parent-child conversations about emotions are related to children’s emotion talk and emotion knowledge and the roles of parents and children in emotion socialization.

Children (N = 256, 115 girls) and their primary caregiver (62% White, 9% Black, 19% Hispanic, 3% Asian American, 7% ‘Other’) participated in three waves across three years. In waves 1 (Mage = 2.45 years, SD = .26) and 2 (Mage = 3.51 years, SD = .26), parents and children discussed wordless images depicting different cartoon children (i.e., the emoter) experiencing an emotion (anger, disgust, fear, sadness, joy) in response to a referent (e.g., dropping ice cream). A separate adult sample (N = 105) validated each emotion image as expressing the intended emotion (agreement range = 80.95% - 93.81%). Parent-child discussions were transcribed and three indicators of emotion talk for parents and children were extracted using Python: (1) emoter words, (2) referent words, and (3) emotion labels. Parents’ knowledge questions (e.g., “what are they feeling?”) and causal questions (e.g., “why do they feel that way?”) were coded by two independent researchers (interrater reliability ICCs ≥ .99). Children’s emotion knowledge was assessed at waves 2 and 3 (Mage = 4.48 years, SD = .26) using the Affective Knowledge Test (Bassett et al., 2013).

We selected the best fitting measurement model with children’s emotion talk modeled as latent factors with emotion labels, emoter words, and referent words as indicators, χ2 (10) = 3.34, p = .97; RMSEA = .00; CFI = 1.00, TLI = 1.00, SRMR = .02. Parents' emotion talk was modeled as observed variables by aggregating the sum of emotion labels, emoter words, and referent words for each page.

We used structural equation modeling with MLR to examine whether parent causal questions, knowledge questions, and emotion talk predicted children’s emotion talk and emotion knowledge across development (Figure 1). We used bootstrapping with 5000 draws to test for indirect effects (MacKinnon et al., 2007). We also included covariates (children’s expressive language, gender, age, and parental income and education) specified with direct pathways to endogenous variables. This specification demonstrated acceptable model fit, χ2 (81) = 77.35, p = .59, CFI = 1.00, TLI = 1.00, RMSEA = 0.003, SRMR = .03.

The model revealed several noteworthy concurrent and longitudinal direct and indirect relations between predictor and outcome variables (see Table 1). First, parents’ questions at waves 1 and 2 were concurrently related to children’s emotion talk, and parents’ questions at wave 1 were indirectly linked to children’s emotion talk a year later. Second, children’s emotion talk predicted their emotion knowledge 1 year later. Lastly, parents’ questions at both waves and parent emotion talk at wave 1 showed indirect, longitudinal associations with children’s emotion knowledge. The results of the current study highlight the multidimensional nature of emotion socialization and the unique roles parents and children play in the development of children’s emotion knowledge.

Authors

• Peter J Reschke, Brigham Young University
  Presenting Author

• Brandon Neil Clifford, Arizona State University

• Mindy A. Brown, Brigham Young University