Rational variability in children's causal inferences: the Sampling Hypothesis.

Area of Science:

• Cognitive Development
• Developmental Psychology
• Bayesian Inference

Background:

• Young children's responses often show variability.
• This variability is frequently interpreted as irrational or unsophisticated.

Purpose of the Study:

• To propose and test the "Sampling Hypothesis" for inductive inference in young children.
• To investigate if response variability reflects a rational strategy for learning.

Main Methods:

• Four experiments were conducted with 4- and 5-year-olds.
• Children observed block distributions and events, then made probabilistic guesses about hidden information.
• Experiments manipulated probability distributions and selection processes.

Main Results:

• Children's guesses accurately reflected the probability distributions of possible hypotheses.
• Guessing patterns were consistent with random sampling from subjective probabilities.
• Performance indicated sophisticated probabilistic reasoning, not just frequency tabulation.

Conclusions:

• The "Sampling Hypothesis" provides a rational explanation for response variability in young children's inductive inference.
• Children may employ sophisticated probabilistic strategies in learning and causal inference.