Correlation tests on SPSS The last thing we’re going to look at is correlation testing using SPSS. Load up the data set alcohol-work-location.sav and carry out the following: We’re going to limit ourselves to what we’ve looked at so far – the Pearson and Spearman Rank correlation tests. Both are found, rather conveniently, in the same place – Under Correlate and Bivariate. Bivariate simply means “two variables”: The test is used to look for a relationship between two semi-dependent or paired data sets of equal size. When getting started, you’ll be confronted with this window. Note the “Correlation coefficients” box – Pearson is ticked by default – tick Spearman as well to run that test simultaneously. You can unclick Pearson if you solely want to run a Spearman Rank correlation test. Next, copy two related data sets across into the variables column. Here, the keypress data are used. The correlation here, if one exists will show that if one person does a lot of work on a Wednesday, they will also do a lot of work on a Friday and vice versa. You can opt for SPSS to mark out which correlations kick out significant values by ticking the “flag significant correlations” box in the bottom left corner. Click OK to run the test and you’ll get the following: In both cases, significance values are stated for each available combination – you will notice that when a variable is compared to itself, the correlation is always 1.000. The combinations we’re really interested in are comparing Wednesday to Friday and vice versa – in this case the correlation coefficients are the same. In both cases, the significance of the correlation is very high – those people who work harder when sober also work harder when drunk… or they moderate their drinking better on Friday lunchtime. Try running correlation tests for the party attendance figures – either en masse, or for each office in turn – are these results tightly correlated as well?