Kidd Lab at Cognitive Development Center Conference 2023
Celeste Kidd's lab presented at the Cognitive Development Center Conference 2023. Check out the research they presented below!
"Children increase their evidentiary standards in more unreliable informational environments" by Evan Orticio, Martin Meyer, and Celeste Kidd.
From their research paper:
Children are increasingly learning from online media sources, not all of which are reliable and few of which contain clear indicators of reliability. We present empirical evidence that children track the reliability of incoming information and become more diligent about fact-checking in less reliable informational environments. We asked 4-6 year-old children (N=60) to listen to a set of statements about jungle animals that were either entirely accurate (reliable condition) or contained some misinformation (unreliable condition), and judge their accuracy. Following this exposure phase, we asked them to judge the veracity of a novel claim about an alien species in a context with the opportunity to sample information to check the claim. Participants could choose to sample anywhere from 1 to 20 aliens, each of which provided positive evidence in support of the claim. Participants in the unreliable condition sampled significantly more aliens before accepting the claim than those in the reliable condition. This result shows that children sensibly adjusted the amount of evidence they required for validating a novel claim according to the reliability of the given informational environment. Greater engagement with the material in the unreliable context further supports existing developmental theories of uncertainty-driven learning. Work in progress investigates (1) whether children’s level of skepticism shows a graded sensitivity to prior information quality, and (2) whether higher skepticism manifests not only in higher amounts of information sampled, but also in higher discernment between strong and weak evidence.
"Visual engagement as a predictor of learning in young children" by Sarah Stolp and Celeste Kidd.
From their research article:
Visual engagement has become a major driver behind media produced for children. Algorithms on YouTube Kids prioritize offering material expected to maintain children’s attention for longer, as do those on other platforms designed for children like Roblox and Twitch. Visual engagement is also used in testing children’s entertainment materials for suitability in promoting language development and literacy (e.g., Anggraini et al., 2022). In this literature, attention is commonly interpreted as resulting from the learning value of the given material. In fact, decades of work on children’s attention has demonstrated that there are at least two sets of factors that contribute to visual engagement, one of one of which indicates informational utility (e.g., Kidd & Hayden, 2015). Children may visually engage because they are learning (e.g., Montessori, 1917; Kidd et al., 2012)—or they may visually engage because of low-level, perceptual attentional attractors like movement, high contrast, or saturation (Aslin, 2007). If we conflate high visual engagement with learning, we run the risk of concluding that children’s television shows purported to be educational on the basis of high visual engagement metrics could actually be scoring high because of their stimulating perceptual content. Likewise, if edutainment production companies manufacture content designed to score high on visual engagement metrics without specifically testing whether children learn from said material, they have done children a disservice. Despite the vast relevant literature on these topics, no work to our knowledge specifically tests whether increasing visual engagement through the use of perceptual attentional attractors like color, motion, and contrast could increase learning. Thus, we designed a study for this purpose. In our study, we present children (ages 2-5 years) with novel objects (toys) introduced with novel spoken labels (e.g., children see a small monster toy and hear “Oh, look at the biffle! What a nice biffle”). We expose children to 6 such word-object associations, in one of two randomized conditions. Either children hear the word-object associations with the object presented in an austere setting (on a gray background), or they hear them in a highly visually engaging one (on top of a colorful, attractive, moving video). Data collection with children is in progress, and will be complete within the next 6 weeks. Based on our piloting work, we anticipate that visual engagement as defined as total time of visual fixation on screen will be highest in the engaging condition, but that learning in this condition will be diminished as compared to the presentations with austere backgrounds. Even if we do not observe this expected trend in the full sample, however, we expect the outcome to be informative about the merits or potential perils of using visual engagement as a proxy for learning utility and educational value.
To learn more about this conference, please visit here.