Poster #94 - The Emotion Matching Task (EMT): Measurement Invariance and Longitudinal Growth in Preschoolers

• In Event: Poster Session 02

Abstract

Emotion knowledge is a critical aspect of young children's social and emotional development (Trentacosta & Fine, 2010) that is associated with school readiness and academic achievement (Harrington et al., 2020; Ursache et al., 2020). Given the importance of emotion knowledge in early childhood and beyond, it is crucial for measures to accurately and equitably assess multiple aspects of emotion knowledge across children’s demographic characteristics and to index developmental changes in children’s skills accurately. In the current study, we will examine multi-group and longitudinal invariance of the Emotion Matching Task (EMT; Morgan, Izard, & King, 2010), a popular measure of emotion knowledge designed for preschool-age children. We administered the EMT to a sample of 863 racially/ethnically (48% Asian, 26% Latinx, 11% Black, 10% White) and linguistically (42% English; 34% Cantonese, 19% Spanish) diverse preschoolers in San Francisco (Age at start of fall: M = 4.01, SD = 0.55; 47% female) in the fall and again in the spring. The EMT assesses emotion knowledge using photos of children’s faces. It consists of three blocks that measure situational emotion knowledge (matching a specific emotion-eliciting situation to an emotional expression; e.g., “show me the child who just got a nice new toy, just what they wanted”), expressive emotion knowledge (labeling emotion expressions), and receptive emotion knowledge (matching emotion expressions to a label; e.g., “show me the child who feels sad”). Each item is scored as incorrect (0) or correct (1). For each of the three blocks, an accuracy composite is calculated by averaging scores on all items in that block. To ensure that we measure emotion knowledge in an ecologically valid and culturally and linguistically sensitive manner, we selected racially diverse facial stimuli of preschool-age boys and girls (Table 1). Additionally, we administered the task in the child's preferred language (English, Cantonese, or Spanish). Table 2 provides an overview of accuracy composites across the fall and spring and for a subset of demographic groups. Confirmatory factor analysis (CFA) will be used to evaluate the fit of a three-factor model and assess measurement invariance across key demographic groups (child’s gender, race/ethnicity, and language groups) and time points (fall vs. spring). We will follow this up with an analysis of group mean differences in the fall and spring. Finally, we will describe growth trends across the three blocks and demographic groups. This study is one of the few that captures emotion knowledge through an anti-bias lens by adapting our measurement approach to attend to ethnic-racial and linguistic diversity. Additionally, it will advance our understanding of how different facets of emotion knowledge differ across groups and develop over time.

Authors

• Javier Omar, Stanford University
  Presenting Author

• Michael J. Sulik, Stanford University

• Jelena Obradovic, Stanford University