ABSTRACT
When observations are taken simultaneously on two or more variables, there are several ways of examining the relationship, if any, between the variables. For example, principal component analysis (or even factor analysis?) may be appropriate if there are several variables which arise ‘on an equal footing’. In this section we consider two rather simpler approaches. A regression relationship may be appropriate when there is a response variable and one or more explanatory variables. A correlation coefficient provides a measure of the linear association between two variables. These techniques are straightforward in principle. However, the examples in this section demonstrate the difficulties which may arise in interpreting measures of correlation and in fitting regression relationships to non-experimental data.
