Abstract:  The knowledge of statistical information, experimental results and informant’s testimonies can provide important help for children to understand science world. In the past decade, the psychological mechanism of the interaction of the three kinds of information in children's scientific thinking has not been elucidated. In this study, research paradigm from Bridgers et al. (2016) had been used to explore the interaction of independent exploration and informant testimony on physical causal reasoning of 5-year-old children. Moreover, children's perception of testimony's reliability and informant's self-awareness had been analyzed simultaneously. The proposed approach is supported by two experiments which can be summarized as follows. Experiment 1 adopted a 2×2 completely randomized design. The independent variable is the confidence of testimony (two levels: self-confidence, not-confidence) and the number of children’s independent exploration (two levels: no-exploration, nine-exploration). The dependent variable is the frequency of children's choice of candidate cause. Informant's testimony (such as A is more likely to cause an effect than B) was given by an adult female in a laptop, and children played with a music box to get evidences contradicting testimony (B is more likely to cause an effect than A). In this paper, 84 preschool children (50 boys and 34 girls, mean age 63.7 months) took part in experiment 1. The results show that there are more children choose cause contradicting testimony in nine-exploration level than in no-exploration level, but the difference between treatments was not statistically significant. Moreover, children are not sensitive to the testimony's reliability and informant's self-awareness. Experiment 1 found that the evidences support cause B did not significantly improved child's choice of cause B, in condition that testimony support cause A. The possible reason is the amount of evidence that support cause B is too small. To test this hypothesis, Experiment 2 used a big amount of evidence that support cause B to explore the interaction of independent exploration and informant’s testimony. A single factor completely random experimental design was adopted in experiment 2, and the independent variable is the confidence of testimony (two levels: self-confidence, not-confidence). Experimental procedure is the same as the nine-exploration level in experiment 1. There were 54 preschool children (28 boys and 26 girls, mean age 63.9 months) took part in experiment 2, and each child independently played with a music box eighteen times. The results show that children are more likely to choose cause contradicting testimony in the eighteen-exploration level than in no-exploration level (experiment 1). Once again, children are not sensitive to testimony's reliability and informant's self-awareness. The results of two experiments indicate that 5-year-old children can integrate the evidence from independent exploration and the testimony from informant to infer causality relationship. The number of children who choose cause contradicting testimony is increased by the increase in the number of evidence contradicting testimony. The follow-up research should focus on the interaction effect between the exploration of preschool child, the results of adult’s experiment and the testimony from informants, and focus on how to motivate children to evaluate adult’s self-awareness and credibility.

Key words: preschool, independent exploration, confidence of testimony, conflict evidence, causal inference