According to Cronbach's method of data analysis--which decomposes individual scores into separate components for different levels of aggregation--data collected for an investigation of the development of individual differences in ability test scores from grade 4 through grade 9 were grouped into school district, class, and individual components. Demographic characteristics of neighborhoods, self-selection of academic versus general programs, instructional emphasis, and irregularities in test administration were used to explain the test score variation between schools and districts. First-order factor analyses at the individual level showed the standard primary mental abilities pattern. Second-order analyses grouped the primary factors into one "power" and one "speed" factor. These same factors were also found in the analyses of class means; however, the processes behind the factors at the class level were likely to be different from those operating at the individual level. (Author/MV)