Temperament and emotional overeating: the mediating role of caregiver response to children's negative emotions

Emotional overeating (EOE) is the tendency to consume food to regulate negative emotions and is linked to obesogenic eating and pediatric obesity risk. Early childhood is critical for developing self-regulation and eating patterns, yet mechanisms driving early EOE are unclear. Grounded in the psychobiological model of temperament (negative affectivity, orienting/regulation, surgency/extraversion), the study examines whether infant temperament (at 3 months) predicts preschool EOE (at 36 months) and whether caregiver responses to children’s negative emotions (at 18 months) mediate these associations. The authors hypothesize indirect effects such that higher negative affectivity and lower regulation increase EOE risk via less supportive or more non-supportive caregiver responses; surgency was expected to relate to higher EOE, with no specific hypothesis about its link to caregiver responses due to limited prior evidence.

- Emotion and eating: Beyond homeostatic drives, hedonic motivations can lead to eating for pleasure or to downregulate negative affect. EOE is considered a learned response shaped by individual, relational, and environmental factors, including caregiver feeding practices. The exact developmental mechanisms in early childhood remain underexplored. - Child temperament and EOE: Temperament involves biologically based reactivity and regulation. Higher negative affectivity has been associated with EOE, potentially reflecting greater need for regulation. Early orienting/regulation (precursor to effortful control) is typically inversely related to dysregulated eating, but links from infancy to later EOE warrant further study. Surgency/extraversion has been associated with food approach behaviors (e.g., eating in absence of hunger), though its relation to EOE is less clear. - Caregiving and EOE: Feeding practices such as using food to regulate emotions or as reward predict later EOE. Broader caregiving qualities (warmth, emotional responsiveness) are associated with better-regulated eating; low parental emotional responsiveness during mealtimes predicts higher EOE, especially under household chaos. - Emotion socialization: Supportive caregiver responses (validating emotions, problem-solving) foster children’s emotion regulation; non-supportive responses (punitive, minimizing, caregiver distress) can escalate distress and impede regulation. Few studies have tested how these responses relate to EOE in young children. - Child temperament and caregiving interactions: Caregiver behaviors and child characteristics are transactional. Caregivers may tailor strategies based on child temperamental reactivity/regulation, including use of food to soothe. However, the role of infant temperament with caregiver responses to negative emotions in shaping EOE is not well understood. - Present study: Tests a mediation model where infant temperament predicts EOE via caregiver supportive and non-supportive responses to children’s negative emotions, controlling for demographic covariates.

Design and participants: Longitudinal analysis of 358 caregiver–child dyads from the STRONG Kids 2 (SK2) birth cohort in the United States (parent cohort N=468; 50.6% male). Data waves at child ages: T1=3 months, T2=18 months, T3=36 months. Attrition: ~7.93% from T1 to T2; ~15.20% from T1 to T3. Sample demographics: majority White (85.8%), with smaller proportions Asian (8.1%), Black (7.5%), Hispanic/Latino (5.0%), and Native American (1.4%). IRB approval: University of Illinois (#13448). Measures: - Dependent variable (T3): Emotional overeating (EOE) via Child Eating Behavior Questionnaire (CEBQ) EOE subscale (4 items; e.g., “eats more when worried”; 1=never to 5=always), α=0.75; composite score (M=1.68, SD=0.58). - Independent variables (T1): Infant temperament via Infant Behavior Questionnaire–Revised Very Short Form (IBQ-R VSF; 37 items): negative affectivity (12 items; α=0.72), orienting/regulation (12 items; α=0.66), surgency (13 items; α=0.63). Items rated 1–7; higher scores indicate higher trait levels. - Mediators (T2): Caregiver responses to children’s negative emotions via Coping with Children’s Negative Emotions Scale (CCNES; 12 scenarios; 6 response types per scenario; 1–7 scale). Six subscales were grouped into two latent constructs: supportive (Problem-Focused Responses, Emotion-Focused Responses, Expressive Encouragement) and non-supportive (Punitive, Minimizing, Caregiver Distress). - Covariates: Child gender, race/ethnicity, caregiver education, household income. Analytic approach: Structural Equation Modeling (SEM) using lavaan 0.6-12 in R 4.1.2. Steps: (1) Confirmatory Factor Analysis (CFA) for supportive and non-supportive latent constructs; model fit evaluated with CFI, TLI, RMSEA, SRMR. Modifications added residual covariances between certain indicators (EER–MR, EER–PR, EFR–MR, EFR–DR) to improve fit. (2) Structural regression with 5,000 bootstrap samples to test direct and indirect effects from temperament to EOE via caregiver responses. Missing data assessed with Little’s MCAR test; FIML used. Model fit: CFA initial poor fit improved to good fit after modifications: χ²(3)=7.05, p=0.070; CFI=0.996; TLI=0.978; RMSEA=0.061; SRMR=0.035. Final SEM fit: χ²(104)=144.84, p=0.005; CFI=0.959; TLI=0.945; RMSEA=0.033; SRMR=0.043.

- Missingness: Little’s MCAR test indicated data missing completely at random, χ²(63)=67.59, p=0.32; handled via FIML. - CFA of caregiver responses: After adding theoretically consistent residual covariances (EER–MR, EER–PR, EFR–MR, EFR–DR), measurement model fit was good: χ²(3)=7.05, p=0.070; CFI=0.996; TLI=0.978; RMSEA=0.061; SRMR=0.035. Supportive and non-supportive latent variables were not significantly correlated (r=-0.09, p=0.22). - Direct effects (SEM): - Infant surgency at T1 positively predicted EOE at T3: c=0.09, SE=0.04, p=0.03. - Infant orienting/regulation positively predicted supportive caregiver responses at T2: a2=3.50, SE=1.11, p=0.002. - Supportive caregiver responses at T2 negatively predicted EOE at T3: b1=-0.10, SE=0.003, p<0.001. - Non-supportive caregiver responses at T2 positively predicted EOE at T3: b2=0.01, SE=0.003, p=0.04. - Early temperament did not significantly predict non-supportive responses. - Indirect (mediation) effects: - Supportive responses significantly mediated the link from T1 orienting/regulation to T3 EOE: a2*b1=-0.04, SE=0.01, p=0.02, 95% CI [-0.067, -0.010]. - Non-supportive responses did not mediate temperament–EOE associations. - Overall: Lower infant orienting/regulation was associated with greater EOE via reduced supportive caregiver responses; surgency had a direct positive effect on later EOE. Both supportive and non-supportive caregiver responses had significant direct effects on EOE.

Findings support a mechanistic pathway whereby infant regulatory capacities (orienting/regulation) shape caregiver emotion socialization, which in turn influences children’s tendency to overeat in response to negative emotions. Higher infant orienting/regulation elicited more supportive caregiver responses, fostering better emotion regulation strategies and reducing reliance on food for affect regulation. In contrast, non-supportive responses directly increased EOE, independent of infant temperament links. Surgency predicted EOE directly, suggesting temperament-driven approach/impulsivity pathways to overeating that operate outside the measured caregiving responses. Contrary to expectations, negative affectivity did not predict EOE directly or indirectly in the full model, potentially due to shared variance with other temperament dimensions or the specific role of caregiver strategies that frame food as a regulatory tool. Overall, results underscore that how emotions are regulated—via caregiver support—may be more critical for EOE development than the mere presence of negative emotions.

EOE in early childhood is shaped by both biologically based temperament and caregiver emotion socialization. Infant orienting/regulation reduces later EOE through supportive caregiver responses, whereas higher infant surgency directly increases EOE risk. Both supportive and non-supportive caregiver responses have distinct direct effects on EOE. The study highlights caregiver emotion socialization as a key pathway linking early temperament to eating regulation, pointing to opportunities for early prevention and caregiver-focused interventions. Future research should trace these pathways across development and disentangle specific temperament facets and caregiving strategies that foster or mitigate EOE.

- Sample composition: Predominantly White, well-educated families; most primary caregivers were mothers, limiting generalizability across diverse racial/ethnic, socioeconomic, and caregiver-role groups. - Measurement method: Reliance on caregiver self-reports may introduce shared method variance and common-informant bias; need for multimethod (observational, lab tasks) and multi-informant designs. - Psychometrics: Internal consistency for IBQ-R VSF orienting/regulation and surgency subscales was slightly below 0.70 in this sample. - Conceptual framing: Use of the term “negative emotions” may obscure adaptive functions of such emotions; interpretation should consider nuance. - Causal inference: Despite longitudinal design, unmeasured confounding and alternative explanations cannot be ruled out; replication is needed. - Cultural context: Emotion socialization practices vary by culture and context; findings may not apply to populations where suppression or other strategies are adaptive.