When I was 8 years old, my parents bought me a guitar for Christmas. Even then, I'd desperately wanted to play the guitar for years. I could not contain my excitement at getting this gift (had it been an electric guitar I think I would actually have exploded with excitement). The guitar came with a 'learn to play' book, and after a little while of trying to play what was on page 1 of this book, I readied myself to unleash a riff of universe-crushing power onto the world (well, 'Skip to my Lou' actually). But I couldn't do it. I burst into tears and ran upstairs to hide.¹ My dad sat with me and said 'Don't worry, Andy, everything is hard to begin with, but the more you practise the easier it gets.' In his comforting words, my dad was inadvertently teaching me about the relationship, or correlation, between two variables. These two variables could be related in three ways: (1) positively related, meaning that the more I practised my guitar, the better a guitar player I would become (i.e., my dad was telling me the truth); (2) not related at all, meaning that as I practised the guitar my playing ability would remain completely constant (i.e., my dad had fathered a cretin); or (3) negatively related, which would mean that the more I practised the guitar the worse a guitar player I would become (i.e., my dad had fathered an indescribably strange child). This chapter looks first at how we can express the relationships between variables statistically by looking at two measures: covariance and the correlation coefficient. We then discover how to carry out and interpret correlations in SPSS. The chapter ends by looking at more complex measures of relationships; in doing so it acts as a precursor to the chapter on multiple regression.

¹This is not a dissimilar reaction to the one I have when publishers ask me for new editions of statistics textbooks.