The Stanford-Binet Intelligence Scale: Fourth Edition is a conceptually new version of this traditional intelligence scale. The new scale has a solid basis in theory, but there is little evidence that the Binet matches its intended theory. This study was designed to determine whether the Binet corresponds to the theory that guided its construction. First- and higher-order confirmatory factor analysis was conducted using the entire standardization sample (N=3,354) and three age groups from the standardization sample--12-23 years (n=910), 7-11 years (n=960), and 2-6 years (n=1,484)--to determine the extent to which the Binet measures verbal reasoning, quantitative reasoning, abstract/visual reasoning, short-term memory, crystallized versus fluid intelligence, and general intelligence. The results generally support the four first-order factors as reflecting the underlying structure of the new Binet. The biggest deviation from the theoretical structure occurred for ages 2-6 years; for this age group, it was difficult to separate memory from reasoning. The hierarchical analyses support the test authors' contention that there is a strong "g" component underlying the Binet, but did not support the second level of the Binet theory (which combines the first-order factors into crystallized versus fluid intelligence). Five tables present sample data, and six figures diagram theoretical relationships among factors. (Author/SLD)