2

I have 10 participants who consumed an intervention for 3 weeks. I would like to look at any significance between pre-intervention and post-intervention readings. What SPSS test should I use?

I have tried doing a Paired samples t-test because it's the same sample at 2 different times but got confused on which data to place in variable boxes.

This is because I have 4 sets of data: mean systolic before, mean systolic after, mean diastolic before, mean diastolic after for each of the 10 participants.

If Paired T-test is correct, would it make sense to do this:

Pair 1: Variable 1 (systolic before) -- Variable 2 (systolic after)

Pair 2: Varible 1 (diastolic before) -- Variable 2 (diastolic after)

Isme Shinu
  • 23
  • 2
    Yes, if it makes conceptual sense to interpret SBP and DBP separately. You may also consider Mean Arterial Pressure (MAP). – Penguin_Knight Apr 17 '19 at 23:06
  • Judging the usefulness of an intervention based on only 10 participants is problematic. To what population would you refer when discussing results? Is this the same situation posted here? – BruceET Apr 18 '19 at 00:01
  • @Penguin_Knight Thank you, I will look into Mean Arterial Pressure – Isme Shinu Apr 18 '19 at 09:55
  • @BruceET Thank you for your response. Yes it is an issue but, unfortunately, I am a student and could only recruit a certain number especially due to university budget. I would refer to female university students. Yes it is the same situation; I decided to create a separate thread to not cause confusion :) – Isme Shinu Apr 18 '19 at 10:30
  • Maybe have a look at https://stats.stackexchange.com/questions/3466/best-practice-when-analysing-pre-post-treatment-control-designs – kjetil b halvorsen Apr 18 '19 at 21:38

1 Answers1

1

You were correct that the paired sample t-test was what you wanted. However, you can only test a single outcome at a time. You are running into problems because you are attempting to test two different outcomes (systolic and diastolic pressure).

Penguin_Knight suggested one solution in the comments: use a single measure which incorporates both, such as mean arterial pressure. I don't know anything about the health field, but the idea of moving to a single new measure which unifies the previous two makes sense.

Alternatively, you could test either systolic or diastolic pressure independently, or run two different t-tests.

Which should you pick? The one that fits your hypothesis! A statistical test is based on a hypothesis about what you think should happen. Is your hypothesis about diastolic or systolic pressure? If so, then you should use whichever your hypothesis includes.

indigochild
  • 253