A Review of Children's Use of Causal Inference Principles

Abstract

SEDLAK, ANDREA J., and KurTZ, SUSAN T. A Review of Children's Use of Causal Inference Principles. CHILD DEVELOPMENT, 1981, 52, 759-784. Evidence for the development of principles guiding causal inferences was critically reviewed. The principles examined included: (1) the logical requirements suggested by Hume, that is, the temporal precedence of causes, and causeeffect covariation and contiguity; (2) special circumstances affecting inferences, that is, causeeffect similarity, causal facilitation versus inhibition, and the particular causal configuration relevant to behavioral effects; and (3) principles associated with causal schemata, that is, the extremity principle/magnitude covariation, direct compensation/augmentation, and inverse compensation/discounting. Although discrepant findings characterize work in a number of areas, some general patterns were apparent. Even preschoolers could use many of these principles when demands on their memory or verbal skills were minimized. Nevertheless, their performance was unreliable and they were surprisingly willing to accept imperfect cause-effect covariation. They had difficulty in identifying inhibitory covariation and in applying magnitude covariation to inhibitory causes. Taken together with the comparatively later (7-9 years) emergence of discounting, these results suggested a possibly general delay or difficulty with inverse inference functions. Consistent with this, even preschoolers were able to use direct compensation or augmentation. In fact, the younger children overextended this direct inference function and systematically violated discounting by offering augmentation errors. 2 integrative theoretical accounts for the appearance and differential distributions of these errors were proposed. Finally, discounting seemed to be preferred to inference uncertainty, since the latter was only occasionally evident even among adults.