XIX. Effect Size
Author(s)
Prerequisites
Researchers often seek to learn more than whether the variable under investigation has an effect and/or determine the direction of the effect. This is particularly true for research that has practical applications. For example, an investigation of the efficacy of a pain-relief drug would seek to determine the extent of the relief and not merely whether there was any relief. Similarly, a study of a test-preparation course's efficacy would seek to determine how much the course raises students' test scores. Finally, a study of the relationship between exercise and blood pressure would seek to determine how much blood pressure decreases for a given amount of exercise. In all of these examples, a significance test would not be sufficient since it would only provide the researcher information about the existence and direction of the effect. It would not provide any information about the size of the effect.
Before we proceed with a discussion of how to measure effect size, it is important to consider that for some research it is the presence/absence of an effect rather than its size that is important. For example, Eddington in 1919 conducted a critical test of one of the key tenets of Einstein's general theory of relativity: gravity bends light. Since the alternative theory (Newton's) held that there should be no effect of gravity on light, it was the presence/absence of the effect rather than its possible size that was critical. Measurements taken during a total eclipse of the sun on May 29, 1919 supported Einstein's theory that gravity bends light and contradicted the long-held views of Newton. Thus, it was evidence of the existence of an effect rather than the estimation of its size that was important in this classic experiment.
Information about the Eddington experiment