Concerning the determination of trait variability

An analytic technique for the study of trait variability is presented. An expression for the average variance from test to test and an expression for the variance of these variances are derived in terms of the number of tests and the intercorrelations between them, and limiting cases are examined. The question of the true relationship between the nature of the distribution of test scores in a sample ofN persons and the nature of the distribution ofn traits in a single individual is discussed, and other problems are introduced.