Linear regression uses a correlation matrix as its starting point, as does factor analysis. Before employing these statistical techniques, it is important to examine the underlying correlation matrix to understand the bivariate patterns that will serve as the foundation for multivariate data modeling. The Pearson correlation is embedded into linear regression and factor, so it is appropriate to discuss correlations before moving on to these topics. One useful property of correlation coefficients is that they are bounded by -1 and 1, so it is easy to compare their strength across a set of fields that may have very different means.
Under the Analyze menu in SPSS Statistics, the Correlate choice includes four options shown in the following screenshot:
This chapter will cover the first two, Bivariate... and Partial.... The Distances... and Canonical Correlation...
