Children use Causality as a Guide to Question Asking

• In Event: Factors Affecting Reasoning and Problem Solving in Children

Abstract

Gathering information via question asking is an essential part of early learning. However, it also requires learners to choose which questions to ask from a vast space of possibilities. Previous research has demonstrated that by age 5, children recognize that some questions are more informative than others (e.g., Ruggeri et al., 2017). Here, we investigate a novel factor that might influence children’s identification of more informative questions: the relationship between cause and effect. Prior research suggests that 4- and 5-year-olds already have expectations for how the abstract features of an event (proportion, dynamics of movement, etc.) relate to the characteristics of its cause (e.g., Magid et al., 2015; Tsividis et al., 2015), but it is unknown whether they can apply this understanding to inform their question asking. The present study explores whether children can use the connections between cause and effect to identify which questions to ask during information search.

To do this, we assess whether 5- to 7-year-olds preferentially choose to ask questions with the potential to reveal relevant causal information. Participants are presented with a storybook in which they hear a description of an event with an unknown cause (e.g., “What animal made the mess in the kitchen?”). In three trials, participants are given a series of forced-choices between two questions about the event’s possible causes. On each trial, both questions target the same kind of feature (e.g., the animals’ abilities, appearance, or preferences), but only one of the two questions has the potential to reveal information that is relevant to determining the cause of the event (see Table 1). For example, for a story in which an unknown animal made a mess on top of a high surface, participants were given a choice between asking if the candidate culprits could climb (relevant to determining cause) or if they could make noise (irrelevant to determining cause).

With half of the target sample (N = 48 per age group) collected for both 5-year-olds (N = 22, mean age: 66.22 months, range: 60.43-70.96 months) and 7-year-olds (N = 22, mean age: 91.19 months, range: 84.66-95.67 months), we find that children choose the causally relevant question on a significant majority of trials (85.6% of trials, p < 0.001, 95% CI [78.4, 91.1], two-tailed binomial). Furthermore, we see similarly strong performance in both age groups and across all three feature types (see Figure 1). Children were also able to use the causally relevant information. At the end of the task, children were told the answers to both the relevant and the irrelevant questions for three candidate causes: 86.3% of children correctly identified the candidate whose features indicated they could have caused the event (p < 0.001; 95% CI [72.6, 94.8], two-tailed binomial).

Preliminary results suggest that young learners employ their understanding of cause and effect relationships during question asking to achieve effective information search. This work offers novel insight into the strategies that children use to guide their self-directed learning about the world.

Authors

• Elizabeth Lapidow, University of California - San Diego
  Presenting Author

• Amberley R Stein, University of California - San Diego

• Caren M. Walker, University of California - San Diego