How to Help Young Children Ask Better Questions?

Abstract

In this paper, we investigate the informativeness of 4- to 6-year-old (N = 125) children’s questions using a combined qualitative and quantitative approach. Children were presented with a hierarchical version of the 20-questions game, in which they were given an array of objects that could be organized into three category levels based on shared features. We then tested whether it is possible to scaffold children’s question-asking abilities without extensive training. In particular, we supported children’s categorization performance by providing the object-related features needed to ask effective constraint-seeking questions. We found that with both age and scaffolding children asked more effective questions, targeting higher category levels and therefore reaching the solution with fewer questions. We discuss the practical and theoretical implications of these results.

Children’s failure to control variables may reflect adaptive decision-making

“…How can children be robust and effective causal learners but also fail dramatically at implementing CVS in scenarios where adult scientists deem it to be the appropriate testing strategy? On the one hand, this is in line with previous work showing it is hard to robustly change children's information search strategies through instruction (e.g., questionasking strategies; see Courage 1989;Denney, Denney & Ziobrowski, 1973;Ruggeri, Walker, Lombrozo, & Gopnik, 2021).…”

Finding the (most efficient) way out of a maze is easier than asking (good) questions.

“…In our question-asking tasks in Experiment 2b, providing children with a visual and conceptual aid (i.e., displaying the spatial organization of the hypothesis space), thereby making it easier for children to understand and thus exploit its structure, was not sufficient to improve search efficiency—even when children were familiarized with the structure of the display (as in Experiment 3). This appears at odds with previous work showing that scaffolding children’s categorization can improve the informativeness of their questions (Ruggeri & Feufel, 2015; Ruggeri et al, 2021; Ruggeri et al, 2019). A possible explanation for this discrepancy is that our visual displays of the hypothesis space (on feature cards or as a hierarchical tree) may have been too complicated for children to parse on their own.…”

“…This process effectively obligated children to think about the different categories (i.e., shape, color, or pattern) before picking a variant, which may have additionally contributed to the cards' role as conceptual aid. Indeed, a recent study has shown that reminding children of the categories that can be used to ask (informative) questions can improve their question-asking efficiency (Ruggeri et al, 2021): In a similar 20-questions game design, children had to identify a target monster from a set of monsters differing in color, shape, and pattern. Before asking a question, the experimenter verbally reminded children of the features they could query ("It could be that the monsters with a specific pattern-dotted or not dotted-turn on the machine.…”

How is the hypothesis space represented? Evidence from young children’s active search and predictions in a multiple-cue inference task.

“…Another proposed explanation for young children's limited efficiency in information search is that they have difficulty going beyond the object level, that is, they fail to spontaneously identify, represent and reason with more abstract task structures. Consistent with this idea, Ruggeri and colleagues (Ruggeri et al, 2021;Ruggeri & Feufel, 2015) found that scaffolding more abstract representations of the hypotheses in the 20-questions game helped 4-to 10-year-olds to ask more informative questions. For example, Ruggeri and Feufel (2015) presented 7-and 10-yearold children and adults with 20 cards, each presenting a word label (e.g., "dog" or "sheep").…”

“…In the current paradigm, representing the hypothesis space in terms of abstracted cue-outcome relationships requires encoding both the cue direction and order (i.e., which color is faster, which shape is faster, and whether color or shape is more important for predicting a monster’s speed; see Table 1). Abstracting the relationship between the cues and the outcome from the pairs of monsters encountered might be rather challenging for children, and in particular for younger children, who have been shown to struggle with such abstractions (Herwig, 1982; Ruggeri et al, 2021; Ruggeri & Feufel, 2015). Abstracting the cue order (i.e., whether color or shape was more important for determining relative speed), a second-order cue, may be particularly difficult for them…”