# How to apply paired t-test seperated in two groups at time (SPSS preferable)?

I am trying to compare pain estimation within two groups. I am following a random control trial research and have to groups for that i.e. experimental and control one. Each group has the same variable i.e. pain estimation measured before and after. I want to check the significance of the same variable measured before and after within each group but NOT between groups as independent sample t test does (is meant for). How do I achieve this on SPSS?

• Before getting an answer I had a little idea to define pain variables for each group e.g var_pain_experimental_before, var_pain_experimental_after as 1st pair and the same for control: var_pain_control_before, var_pain_control_after as a 2nd pair Apr 5, 2020 at 20:06
• Your comment seems to contradict your question. // Can you do 'half' of this project by using only the Control group: comparing Control Before to Control After? If there are 20 Control subjects, you'd be looking at 20 differences $\mathrm{Diff} = \mathrm{After}- \mathrm{Before}.$ // if that is the case then you have two choices: (a) A paired t test of $\mu_A = \mu_B$ vs $\ne.$ (b) A one-sample t test on the 20 Dif scores with null hypothesis $\mu_D = 0$ vs $\ne.$ // Before doing any t test make sure there are no far outliers in your data. Apr 5, 2020 at 21:19
• @BruceET it seems like you trying to put my on a good track and may seem contradictive in between question and comment but as I said I need a paired t test for same variable in each group but not between groups, unless there is a choice to use some variance test on SPSS that would compare the signifinace (if any) of the following one variable variances measured before and after. In conclusion I don't think I could use only Control subjects, I have only 20 subjects equally distributed in two groups, so each is n=10. Too few. Apr 6, 2020 at 8:54