# Thread: Unequal sample sizes and statistical tests

1. ## Unequal sample sizes and statistical tests

Dear all,
I need you help urgently, and I want to thank you all for that.
My question is as follows,
I have two groups’ j and k and I have two conditions, condition A and condition B
Group j has 29 members and group k has 623 members,
What I want to test here is, if the members in group j (N=29) have a significant change in condition B compared with A, against the change of the K (N=623) group from condition A to condition B. Since the K group have a higher group size it has stable mean. However, the J group has a smaller group size and the mean can change easily.
Can you suggest me any means (statistical test) I can use.
I tried to take the 30 from group J and randomly selected 30 from K and did a test, but I am not sure if I am doing the right thing. The group size is worrying me a lot.
Thank you very much
With kind regards
Daniel

2. SPSS will compute TYPE III SS for you automatically I believe, and this takes into account unequal sample sizes.

3. ## thank you

Thank you, I would be very happy if you or anyone in this forum can add some facts on this point. More detailed and more alternative
Thank you again

4. It sounds like you might want to run an ANCOVA with condition B as the dependent variable and condition A as the covariate.

To evaluate whether the group size is a problem, you can run Levene's test for homogeneity of variance at the same time that you run the ANCOVA. Under ANCOVA in SPSS, select the options menu and then select homogeneity tests. In your output, if the Levene's test is significant, it means that the variances aren't homogenous and you should try to correct your data for homoskedasticity. I am currently trying to figure out how to do this--I know a weighted least squares regression will do the trick, but I can't figure out how to do one.

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts