In social science research, many variables we are interested in are also difficult to measure, making measurement error a particular concern. Despite impressive advancements in measurement in recent years (see particularly Chapter 4 on Rasch measurement), simple reliability of measurement is still an issue in much research. Unreliable measurement causes relationships to be underestimated (or attenuated), increasing the risk of Type II errors. In the case of multiple regression or partial correlation, effect sizes of other variables can be overestimated if another variable in the equation is not reliably measured, as the full effect of that variable might not be removed.
This is a significant concern if the goal of research is to accurately model the “real” relationships evident in the population. Although most authors assume that reliability estimates (Cronbach alphas) of .70 and above are acceptable (e.g., Nunnally, 1978), and Osborne, Christensen, and Gunter (2001) reported that the average alpha reported in top educational psychology journals was .83, measurement of this quality still contains enough measurement error to make correction worthwhile (as illustrated below).
Correction for low reliability is simple and widely disseminated in most texts on regression but rarely seen in the literature. I argue that authors should correct for low reliability to obtain a more accurate picture of the “true” relationship in the population and, in the case of multiple regression or partial correlation, to avoid overestimating the effect of another variable.
Note that in this age of user-friendly structural equation modeling programs such as Amos and others, modeling and eliminating measurement error is a real and present possibility with almost every analysis. While this topic is discussed in detail elsewhere in this volume (Chapter 32), the discussion below pertains to this sort of analysis as well. Specifically, one can view use of structural equation modeling as another option for reducing the effect of measurement error. Furthermore, readers are encouraged to explore more modern measurement methodologies as another avenue to achieving the same goal—the more accurate modeling of that which we observe and study.