t-test question SPSS

Hi - when running a t test to compare scores at two time points (pre to post intervention) do you look at the paired samples test or paired samples correlations output in SPSS?

Also unsure whether to look at the one sides or two sides p for the significance level.



TS Contributor
It depends on whether the same subjects were measured pre and post, or whether
the pre-intervention sample consists of different subjects than the post-intervention

Regarding the p-value, use the two-sided result.

With kind regards



TS Contributor
Oh, sorry, I mis-read your question. So you were asking about t-test (comparison of means between pre and post)
versus correlation (how closely the relative positions of subjects pre are related to their relative positions post)?
It depends on what you want to find out. In interventional studies, usually one is interested in the pre-post difference
in the first place.