THEORIZING AND STATISTICS.

This article presents the main outline of the first semester of the Statistics course at Yale University in New Haven, Connecticut in 1970. The developing reproachment between theorizing and methods in general has not directed enough attention to the connection between theorizing and statistics. Sociologists have paid most attention to this problem at relatively advanced levels, for example the use of statistics to test causal inference. In order to integrate theory and statistics, the course begins with symbolic logic, set theory and probability. On this foundation we enter the study of statistics, spending minimal time on measurement theory, estimates of central tendency, and dispersion of distributions. Symbolic logic is used to get students to state hypotheses clearly and to see what empirical results must occur if the hypotheses are to be considered true. The first semester closes with measures of association, taking correlation and regression as the basic paradigm, then introducing ordinal and nominal measures.