Testing the limits of children’s selective preference for generalisable information transmission

Bazhydai, Marina and Karadag, Didar (2025) Testing the limits of children’s selective preference for generalisable information transmission. In: Budapest CEU Conference on Cognitive Development, 2025-01-08 - 2025-01-11, CEU.

Abstract

Emerging literature on children’s teaching shows that they select information for transmission based on information type and its recipients. Children preferentially share information that is generalisable to a kind (“dogs have fur”), rather than specific (“this dog has spots”) (Gelman et al., 2008, 2013; Baer & Friedman, 2018; Cimpian & Scott, 2012). However, prior research focused on neutral information and systematic investigations into the extent and limits of such preference are lacking. Specifically, it is unknown whether generalisability preference is retained when information is affectively arousing, potentially threat-inducing, or health-related. Such saliency dimensions have been shown to affect children’s selective social learning (Ronfard & Harris, 2018) but have not been explored in children’s selective teaching. In four studies, using an interactive online paradigm, children aged 6-9 (N = 144, n = 36/study) were presented with generic and specific facts which were: neutral about animals (study 1), presented an implicit threat of evolutionary (e.g., snakes) or culturally acquired (e.g., germs) nature (study 2), health-related about animals (study 3), and health-related about humans (study 4). Following learning, children selected facts for transmission to a naïve agent. We coded the first choice and the total number of generic and specific facts shared. We expected a disruption of generalisability preference in non-neutral domains (studies 2-4). Study 1 conceptually replicated prior findings that children preferentially share generic information about neutral topics. Robust preference across both measures was also evident in studies 3 and 4, while study 2 instead found a null effect (Figure 1). These results challenge the assumption about the uniform nature of the value of information selected for transmission. Specifically, the implicit threat present in information seems to disrupt the generalisability preference in information transmission, while evidential health-relevant information is treated with preferential sharing of most socially valuable - generalisable - facts.