Exploratory Factor Analysis with SAS

Would Error-Free Measurement Make a Real Difference?

A Demonstration with Effect Sizes

To give a concrete example of how important good measurement is, we can use an example from the first author’s survey of the educational psychology literature (Osborne, 2008). This survey consisted of recording all effects from all quantitative studies published in the Journal of Educational Psychology (usually considered one of our top empirical journals) during the years 1998 to 1999.

The mean effect size and the reliability coefficient (Cronbach’s α), if reported, were used to identify nine potential measurement scenarios. First, a small, average, and large observed effect size were determined as the effect size one standard deviation below the mean (Cohen’s ...