Comparison of Several Populations

Abstract

Throughout this book we have stressed the basic statistical concept of variability. When some measurement, such as height or aptitude for a particular job, is made on several individuals, the values vary from person to person. The variability of a quantitative scale is measured by its variance. If the set of individuals is stratified into more homogeneous groups, the variance of the measurements within the more homogeneous groups will be less than that of the measurements in the entire group; that is what “more homogeneous” means. For example, the variance of the heights of pupils in an elementary school is usually greater than the variance of heights of pupils in just the first grade, the variance in the second grade, and the variance in each of the other grades. At the same time, the average height of pupils also varies from grade to grade.