Limited social modeling of familiar and unfamiliar foods in infant feedings at 6/9/12 months

• In Event: Towards a Holistic Developmental Investigation of Children’s Eating Behaviors

Abstract

Under their “natural pedagogy” framework, Csibra and Gergely (2009) posit that communicative tools such as joint attention, demonstrations, and gestures support young learners to quickly acquire “cognitively opaque cultural knowledge,” information that would be difficult to learn from observation or interaction with their environment alone. Starting around 6 months of age, infants can learn from other people’s facial expressions, vocal cues, and gestures (e.g., Csibra & Gergely, 2009; Mumme et al., 1996; Vaish & Striano, 2004). Along this same timeline, infants encounter one type of new stimuli frequently: Foods they have not eaten before. Given that infants and young children are willing to ingest substances that are not food or could be dangerous to their health (Cashdan, 1994; DeJesus et al., 2015; Rozin et al., 1986), food seems like an ideal candidate to study in terms of natural pedagogy, as what constitutes a food versus a non-food may represent cognitively opaque cultural knowledge.
In this study, mothers in a longitudinal study of eating in infancy were asked to feed their infants familiar and unfamiliar foods when infants were 6, 9, and 12 months old (n = 195). Mothers were given video cameras to record feedings at home with any foods they viewed as familiar and unfamiliar to their infant at that age. A primary aim of this observational study was to investigate spontaneous social modeling in a naturalistic feeding context (i.e., mothers taking a bite of the food they were offering their infants, or miming taking a bite of food). We hypothesized that maternal social modeling would be especially common for unfamiliar foods (as compared to familiar foods) as an opportunity for learning.
We observed a wide range of modeling behavior: Instances of modeling ranged from 0 to 117 for familiar foods (mean = 3.03, median = 0, SD = 11.70) and from 0 to 71 for unfamiliar foods (mean = 2.78, median = 0, SD = 8.07); 27% of observations included at least one instance of modeling (73% did not include any modeling) and the modal response was 0.
Multilevel models reveal that modeling was more common during family meals (67% of family meals) than individual meals (19% of individual meals), b = -2.65, SE = 0.35, z = -7.64, p < .001, and was higher with infant age, b = 0.21, SE = 0.12, z = 4.08, p < .001. Modeling was not related to infant food acceptance, b = 2.78, SE = 3.07, t = 0.90, p = .366, but we did observe significant effects of food type (familiar > unfamiliar), meal setup (family > individual), and feeding modality (spoon-fed foods > finger foods), see Figure 1.
Despite evidence from the natural pedagogy literature that modeling may help transmit opaque cultural knowledge, we rarely observed mothers spontaneously taking a bite of food when offering it to their infants, including when offering unfamiliar foods. Social modeling is discussed as a potential intervention target for future studies given that we rarely observed it spontaneously in this study.

Authors

• Jasmine M. DeJesus, University of North Carolina at Greensboro
  Presenting Author

• Shruthi Venkatesh, University of North Carolina at Greensboro

• Alison Leslie Miller, University of Michigan - Ann Arbor

• Ashley Gearhardt

• Katherine L Rosenblum, University of Michigan - Ann Arbor

• Danielle Appugliese

• Niko Kaciroti

• Julie Lumeng, University of Michigan - Ann Arbor