I am computing some statistical analyses for my dissertation project and I am struggling a bit to run some of them in SPSS.

I have two groups of participants (Group 1 and Group 2), with two treatment times (PRE and POST), for two different output measures (wellbeing and mood).

My supervisor told me to run a Pearson's correlation between the wellbeing and the mood measures at PRE and POST and see if there are any changes for the groups combined and each group separately.

How would you lay out the data file in SPSS? Do I need two separate data files - one for groups combined and one for each group separately? And do I put both measures into one column or separate them into different scale variables in separate columns?


1 Answer 1


For paired t-test the variables are easier to use if they are side by side.

So, if it's the correlation between well-being and mood you'd like to get, paired by participant ID, then there should be a well-being variable, and a mood variable. If you to go Analyze > Compare means and pick paired t-test, you can provide these two variables and get the results.

Now, whether you should stack them vertically by pre/post and group or horizontally as new variables is less of an issue. But between the two I'd say vertical is better:

enter image description here

It's because you can always use "split file" or "select case" function to look into a finer subgroup. And if it turned out your supervisor was talking about all groups, one test for pre and one for post, you can easily accommodate with this long format.

Either way, it's not that hard to restructure data set in SPSS. Try to get familiar with that function (should be under Data > Restructure) and you should no longer be troubled by decisions like this.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.