Poster #43 - Young Children Use More Unconfounded Evidence in Response to False Than Correct Causal Claims

• In Event: 2-110 - Poster Session 07
  In Poster Session: PS 07 Section - Cognitive Processes

Integrative Statement

Young children selectively interact with more unconfounded than confounded evidence in response to false causal claims (Köksal-Tuncer & Sodian, in press), yet it is unclear whether they accurately distinguish correct and false causal claims and use unconfounded or confounded evidence selectively in response to false and correct claims. The present study investigated this question with 45 preschoolers (24 females, Mage = 5;2, range = 3;10–6;8) recruited from kindergartens in a large German city. In a “blicket detector” paradigm (Gopnik & Sobel, 2000), children were presented with a music box which was activated when certain objects were put on it. Objects had two salient features, namely bigness (big vs. small) and having a black dot on them or not. In the learning phase, children interacted with informative evidence which suggested a certain hypothesis (e.g., big objects activate the music). To make sure that children formed the hypothesis, we asked them to predict the effects of novel objects, and only the children who correctly predicted (n = 44) were included in the study. In two within-subjects conditions with the order of the conditions counterbalanced, children were presented with a false and a correct claim. In each condition, a different puppet who was ignorant to the causal relations presented either the correct (e.g., big objects activate the music) or the false claim (e.g., objects with black dot activate the music) and the experimenter asked children whether the puppet was right and why so. Children had two confounded and two unconfounded objects in front of them with which they could freely interact. Eighty-nine percent of the children (39/44) disagreed with the false claim and agreed with the correct claim. The proportion of disagreement and agreement responses were significantly different in the false and correct conditions, McNemar’s test χ2(1) = 37.03, two-sided, p < .001. We coded children’s “use” of individual objects including both their verbal referrals to and behavioral interactions with the objects (e.g., putting on the box, pointing). An Evidence Type (unconfounded, confounded) x Condition (false, correct) repeated measures ANOVA with the number of objects children used as dependent variable revealed significant main effects for the Evidence Type, F(1,43) = 8.69, p = .005, ηp2 = .17; and Condition, F(1,43) = 30.52, p < .001, ηp2 =.42; and a significant interaction effect, F(1,43) = 10.06, p = .003, , ηp2 =.19. Children used significantly more unconfounded than confounded evidence in the false condition, p < .001. Moreover, they used significantly more unconfounded evidence in the false condition than the correct condition, p < .001 (See Figure 1). Young children successfully differentiate correct and false causal claims, use more pieces of evidence in response to false than correct claims and selectively pinpoint more unconfounded than confounded evidence in order to falsify claims. The present findings demonstrate that young children adjust the informativeness of evidence they use in relation to the veracity of the claims that they encounter during communication.

Authors

• Özgün Köksal-Tuncer, LMU Munich
  Presenting Author

• Andrea Saffran, LMU Munich
  Non-Presenting Author

• Beate Sodian, LMU Munich
  Non-Presenting Author