Poster #43 - Short-term Stability of Tasks Measuring Young Children’s Causal and Scientific Reasoning Abilities

• In Event: 3-053 - Poster Session 10
  In Poster Session: PS 10 Section - Cognitive Processes

Integrative Statement

In recent years, various paradigms have been developed to study young children’s causal and scientific reasoning abilities (e.g., Köksal-Tuncer & Sodian, in press; Legare, 2012; Legare & Lombrozo, 2014). The body of research involving these paradigms is constantly growing and points to surprising competencies in young children (for review see Gopnik & Wellman, 2012). However, it is unclear whether these experimental tasks can be used as measures of individual competencies. The present study investigates the short-term stability of two causal reasoning tasks and one scientific reasoning task in a preschool sample (Mage = 4;11 month; age range: 3;8-6;11 month). All tasks were administered twice with a time lag of two weeks between testing 1 (t1) and testing 2 (t2). The Learning Causal Mechanisms Task (Legare & Lambrozo, 2014) uses a gear toy which can be taken apart and rebuilt to test children’s understanding of causal relations. Children are asked to select a missing part to reinstall the function of the toy and to reconstruct the toy after it has been taken apart completely. Measures did not covary between testings, even though there was a tendency for the reconstruction measure (N = 23; part selection: χ²(1) = .04, p = .846, Cramer’s V = .04; reconstruction: χ²(1) = 3.76, p = .052, Cramer’s V = .41). Data patterns point to retest effects: Nine of 10 children who selected an incorrect part in t1, selected the correct part in t2. Similarly, 10 of 16 children who did not reconstruct a causally functional toy in t1 did so in t2. The Explaining and Exploring Inconsistent Evidence Task (Legare, 2012) elicits children’s explanation and exploration behaviors by presenting inconsistent evidence in a blicket detector paradigm (i.e., using boxes which light up when specific objects are placed on them). There were significant correlations between testings for the type of explanation (causal function, causal action, noncausal; N = 18; χ²(2) = 7.68, p = .021, Cramer’s V = .65) and for the duration of exploration (N = 18; r(16) = .87, p < .001). In the Counterargumentation Task (Köksal-Tuncer & Sodian, in press), an ignorant puppet presents a claim about objects making a box play music. From their experience in a learning phase, children know that this claim is wrong. Children can argue against the claim verbally and they can choose between confounded and uncounfounded evidence to support their argument. Counterargumentation behaviors in the first and second testing covaried (N = 22; χ²(4) = 15.59, p = .004, Cramer’s V = .60). The present results tentatively suggest that there is sufficient short-term stability in some tasks to interpret individual performance in terms of competencies. However, the presence of retest effects in one task also indicates that performance on a first encounter with a task may reflect children’s reactions to a novel environment, rather than their level of competence in causal learning or scientific reasoning. The findings provide implications for research practice, especially for using the investigated tasks in longitudinal designs, as well as for early education.

Authors

• Andrea Saffran, LMU Munich
  Presenting Author

• Özgün Köksal-Tuncer, LMU Munich
  Non-Presenting Author

• Beate Sodian, LMU Munich
  Non-Presenting Author