Abstract
To the Editor: Godfrey's article (Dec. 26 issue)1 on simple regression overlooked one point. Too frequently researchers assume that the statistical significance of a correlation coefficient is synonymous with biologic importance. Too often the estimated slope of the regression line is ignored. If a causal relation is assumed between a dependent and an independent variable, a statistically significant estimated slope can indicate the magnitude of an effect of a change in the independent variable on the dependent variable.Imagine that two separate regressions are fitted to a set of data on black and white men in which the dependent variable . . .

This publication has 3 references indexed in Scilit: