Can Children Distinguish Between Irrelevant and Contradictory Evidence for Claims?

• In Event: When Children Navigate Misinformation and Disagreement, how far Does Evidence Matter?

Abstract

Using evidence to evaluate a new scientific claim is a critical skill; however, the relevance and the support of the available evidence can vary. Children have some sense of what makes evidence relevant or irrelevant, although there is significant improvement in children’s ability to evaluate relevance between ages 5 and 10 (Danovitch, Mills, Duncan, Williams, & Girouard, 2020; Johnston, Sheskin, & Keil, 2019). The current project extends the existing literature by examining how children identify and distinguish evidence that supports, contradicts, or is irrelevant in the evaluation of scientific claims.

In Study 1, children ages 7-8 (N = 28) and ages 9-10 (N = 24) were told that they were going to be detectives and sort “clues” (i.e., evidence) about animal behavior (see Table 1 for full examples). For each of the 6 animals, children heard a character make a “guess” (i.e., hypothesis) about why an animal behaves in a certain way. Children then heard 3 different types of evidence that related to the animal: supporting, irrelevant, and contradictory. For each piece of evidence, children indicated whether it was related to the hypothesis, and if so, the degree to which it supported or contradicted the hypothesis. Children’s responses were then recoded into a 1-5 scale: really contradicts, sort of contradicts, not related, sort of supports, and really supports. After the primary experimental task, children completed the Picture Vocabulary Task (PVT; a measure of children’s verbal intelligence), a 15-item Biological Knowledge Test, and the Cognitive Reflection Test – Developmental (Young & Shtulman, 2020).

Overall, participants gave supporting evidence ratings that were above chance, and significantly higher than their ratings of the irrelevant and contradictory evidence (see Figure 1). However, children’s ratings of irrelevant and contradictory evidence did not significantly differ, and neither varied significantly from the midpoint of the scale. After controlling for age, PVT and Biological Knowledge Test Scores related to higher ratings for relevant evidence (rs > .34, ps < .02), but no measures related to ratings of contradictory evidence (rs < .20, ps > .16).

These results suggest that children of this age are able to recognize supporting evidence, but they struggle to distinguish contradictory evidence from irrelevant evidence. To examine whether experimental design issues influenced children’s evaluation of contradictory evidence, we conducted a follow-up study that is currently ongoing.

In Study 2, children ages 7-10 are completing the same paradigm as Study 1 with three alterations. First, instead of initially sorting evidence as relevant or irrelevant, children now initially choose supporting, irrelevant, or contradictory, then choose the degree the evidence supports or contradicts the hypothesis (resulting in the same 5-point scale). Second, for each observation, children now have one more type of evidence to sort: an “Uber” contradictory item that directly contradicts the hypothesis. Finally, the total number of animal items was reduced from six to five to minimize possible fatigue from adding the Uber contradictory items. For SRCD, the results of this ongoing data collection will be presented and individual difference measures will be discussed.

Group Authors

N/A

Authors

• Anthony J Monroe, The University of Texas at Dallas
  Presenting Author

• Areeba Khan

• Alina Yoo, University of Texas at Dallas

• Judith H Danovitch, University of Louisville

• Candice Mills, The University of Texas at Dallas