In previous chapters, we discussed descriptive statistics as ways of illustrating patterns embedded in data. We continue that discussion by examining correlation, the statistical process that measures the extent to which the values of two (or more) variables are related or linked. Correlation processes also allow the researcher to understand spuriousness more clearly.
Most everyone has an understanding of the basic principles of correlation; it is somewhat intuitive. Technically, correlation is the statistical process of measuring how changes in two variables are related to one another. Thus, for example, we may observe that students who get the highest reading achievement test scores are also the ones who read the most. Or, stated another way, as the amount of time spent reading increases, so do the achievement scores (and vice versa). Thus, in this example, the values of both variables increase together, or covary.
Of course, as we have already learned, not every relationship is what it seems! There may be additional variables not taken into account in the analysis that give the original two variables the appearance of covarying. The examples showed how this might happen on a theoretical level. By studying correlation statistically, we can document to the extent to which spuriousness is present among a set of study variables and help to debunk mistaken theoretical assumptions.
Researchers use correlation to explore ...