About this srcd poster session

---

Panel information
Panel 11. Language, Communication

Abstract

---

The ability to develop emotion categorization is important for children to understand not only themselves but other human beings (Izard et. al, 2001). Learning emotions involves, in part, understanding facial expressions and the words that represent them. Research shows that by 3, children can link emotion words to faces (Ogren & Sandhofer, 2022) and rely on language mechanisms like word labels. Although the role of word labels has been studied, little work examines other language tools that may also impact emotion categorization. Tone of voice was selected as research suggests children may rely more on tone than facial expressions when asked to categorize emotions (Ross et. al, 2021). This study builds on prior work by examining how both emotion word labels and tone of voice interact to help children understand emotion facial expressions.

In this study, we told 4-year-olds (N=27, 12 Female, 16 Male; Mean age=4.52 years; 46.43% White, 17.85% Asian, 3.57% Black or African American, 28.57% Multiple Ethnicities, and 7.14 % Decline to State) 20 short scenarios and asked them to select a face that represents the story. Each scenario was one of four emotion contexts (happy, sad, angry, or afraid) as well as an associated emotion word label and tone of voice. We had 8 filler stories where both word and tone matched the context. Each of the other 12 experimental stories fell into one of two conditions, where either the word label or tone matched the story’s context (Partial) or both were mismatched to the story’s context (Mismatched).

As expected, kids selected the face that matched the word, tone, and context on filler trials at a high proportion of the time (Mean = 0.778, p < 2.2e-16). Across the experimental conditions, there was an effect of word over tone (figure 1), as children tended to select facial expressions that matched the word label heard more (z=2.901, p = 0.00371). Children chose the face matching the word label more often than would be expected by chance (t = 13.832, p < 2.2e-16), and when comparing the Partial and Mismatch conditions, we saw children selecting the face matching the word more in the Mismatch condition (z=2.226, p=0.026 ). When examining the interaction of the emotion context and the condition, we see faces matching the word selected at a high amount (figure 2). Across emotion contexts, we see a preference for word-matching faces (ps < .0001).

Overall, we see that children were categorizing emotion faces based on the word label they heard and appeared to disregard tone of voice as a tool. These results suggest that 4-year-olds may rely on the word labels heard more than the tone of voice, which can inform how parents and educators decide to teach and label emotions to children.

Author information

Author	Role
Idella Smolyar, University of Southern California, Los Angeles	Presenting author
Catherine Sandhofer, Ph.D., University of California, Los Angeles	Non-presenting author
Marissa Ogren, Ph.D., Rutgers University	Non-presenting author

⇦ Back to session