DEMONSTRATING THE INFLUENCE OF SAMPLE-SIZE AND RELIABILITY ON STUDY OUTCOME

This article describes the use of a BASIC program for demonstrating the influence of sample size and reliability on study outcome. The program generates a random sample of size N from a population defined by the user. In one option, a correlation between two variables is computed based on a single sample. In the second option, the program generates sample correlations between the two variables for reliability values ranging from 0 to 1. The program is useful for demonstrating variability in parameter estimates imposed by small samples and for demonstrating the limits that unreliable measures place on estimates of population correlations. © 1991, Society for the Teaching of Psychology. All rights reserved.