Descriptive Statistics by Groups (Breakdown) Introductory Overview - Breakdowns vs. Discriminant Function Analysis

Breakdowns can be considered as a first step toward another type of analysis that explores differences between groups: Discriminant function analysis. Similar to breakdowns, discriminant function analysis explores the differences between groups created by values (group codes) of an independent (grouping) variable. However, unlike breakdowns, discriminant function analysis simultaneously analyzes more than one dependent variable and it identifies "patterns" of values of those dependent variables. Technically, it determines a linear combination of the dependent variables that best predicts the group membership. For example, discriminant function analysis can be used to analyze differences between three groups of persons who have chosen different professions (e.g., lawyers, physicians, and engineers) in terms of various aspects of their scholastic performance in high school. One could claim that such analysis could "explain" the choice of a profession in terms of specific talents shown in high school; thus discriminant function analysis can be considered to be an "exploratory extension" of simple breakdowns. For more details concerning this procedure refer to the Discriminant Analysis module. See also, Exploratory data analysis and data mining techniques.