The thesis of this paper is that the decision to use one of three approaches to unit-of-analysis in educational research should be based on substantive considerations, not statistical factors. In addition to the commonly used "total analysis" (regression analysis across individuals), the within and between analyses are inherent in the same data. The within analysis is defined as an individual-based approach in which group differences are held constant, computed by correlation across individuals within each group separately, and averaging or "pooling" correlations if appropriate. The between approach uses the group as the unit-of-analysis by first computing means on the variables across individuals in each group, then using these group means to compute correlations across groups. To illustrate the uses of within and between approaches, a data-based discussion of the conceptual differences is provided in five sections. The first section consists of a review of unit-of-analysis literature, and a presentation of between, within, and total results for the correlations between two variables is included. The second section examines statistical properties of the three approaches when doing internal consistency analyses for constructs assumed to be unidimensional or when exploring the multidimensionality of a set of items using factor or cluster analysis techniques. In the third section, a literature review on methods of measuring organizational climate is presented. The fourth and fifth sections contrast the between and within approaches and discuss total, between, and within psychometric analyses for measuring climate variables using data collected in "A Study of Schooling." Appended are a list of 42 references and 18 tables presenting information about total, between, and within analyses. (FG)