How do children decide what people and what information is trustworthy?

To learn about the world, children can gather first-hand evidence based on their own observations or rely on what other people tell them. Children’s ability to learn from what other people tell them saves them time and effort and allows them to quickly acquire a large amount of information. However, to ensure that what they are learning is true, children must appropriately trust (or mistrust) their informants’ claims. Children typically learn from one informant at a time; often someone whom they interact with over a prolonged period (e.g., a parent, teacher, or sibling). How do children assess the reliability of informants they repeatedly interact with? One strategy children can use is to keep track of the prior accuracy of their informants' claims and place greater trust in the claims of more accurate informants.

How do children keep track of informants' past accuracy. There are at least three possibilities. One possibility is that children track an informant’s number of inaccuracies. This predicts that children’s trust in an informant should be the same when an informant has been incorrect for 1-of-1 trials, 1-of-2 trials, or 1-of-3 trials. Another possibility is that children track an informant’s proportional inaccuracy. This implies that children will be least trusting when the informant has been inaccurate 100 % of the time (1-of-1 trials) compared to 50 % of the time (1-of-2 trials), and most trusting when the informant has been inaccurate 33 % of the time (1-of-3 trials). However, these two strategies make a problematic prediction. They predict equivalent trust in an informant who has been correct for 0-of-0 trials, 1-of-1 trials, or 2-of-2 trials because these three informants do not differ in their number of inaccuracies or in their proportional inaccuracy. However, someone who has provided correct information for 2-of-2 trials should be trusted more than someone who has been correct for 1-of-1 trials, and still more than someone who has been correct for 0-of-0 trials. Thus, it is plausible that children update their assessment of an informant’s future testimony based on all of the information they have at their disposal (i.e., the number of incorrect and correct statements made by an informant). Indeed, children’s trust in an informant increases based on the informant’s proportional accuracy as well as the overall number of correct statements (Ronfard & Lane, 2016). Thus, by age 4, children are equipped to learn from other people as they interact with them.