ABSTRACT
Regression analysis is an extremely versatile tool. However, like all statistical techniques, it is subject to misuse and abuse. This chapter examines simple linear regression. Although this is a somewhat limited tool, a clear discussion of it will facilitate an understanding of related techniques. Ordinarily, a researcher estimates a regression line in order to estimate a causal relationship that is believed to exist on a priori grounds. The anticipated cause and effect relationship is based on the experience and judgment of the researcher; regression analysis merely provides an acceptable arithmetic procedure for quantifying the hypothesized relationship. The interpretation of the estimated slope must be made within the context of the problem studied. What statistics offers is supporting evidence of the reliability of these statements within a probability framework.
