An evaluation of evidence supporting fact-retrieval models

Abstract This literature review compares fact-retrieval models with a schema-based view, which proposes that relational knowledge is a critical component in learning and representing basic number facts such as 8 × 3 = 24. It then critically analyzes evidence adduced to support fact-retrieval models and points out plausible alternative explanations for these results. For example, data collected on mental-arithmetic errors frequently confound retrieved responses with nonretrieved responses. Thus, factor-related errors (e.g., 8 × 3 = 21) and close-miss errors (e.g., 8 × 3 = 23) may, at least in part, be the by-product of fast back-up strategies and may not accurately reflect the mental representation of number facts. Moreover, contrary to Siegler's (1988) fact-retrieval model and consistent with the schema-based view, such errors may increase on unpracticed combinations. The article concludes with recommendations for more clearly determining how the basic number facts are learned and represented.