# Two-way repeated measures--post-hoc analysis?

#### Sheesh19

##### New Member
Hello,
I have run a repeated measures two-way ANOVA using SAS. The repeated measure is time and it shows that it is significant in an interaction with one of my factors. How can I do a post-hoc analysis to determine which group is actually significant? I know the usual for regular ANOVAs, but I am not quite sure how to perform post-hocs with repeated measures.

Any help would be great.
Thanks.

#### Karabiner

##### TS Contributor
determine which group is actually significant?
What do you precisely mean by that? In an ordinary ANOVA one could investigate
whethera some group A has mean which is different from the mean of group B and
the mean of group C (or from the pooled mean of groups B and C, if that's what one
is interested in). You have to describe which COMPARISONS you want to make.

Regards

K.

#### Sheesh19

##### New Member
Hi, thank you for your reply. What you described is similar to what I would like to know. I have two factors and the repeated measure, time. In the within subject effects, time is significant as well as the time*1st factor. This factor has four groups: A, B, C, D. I'd like to know which groups are similar and what not. I'd like to know the differences similar to the results of a Tukey. Is this possible for a repeated measures ANOVA? Please let me know if this does not make sense because I've never done a repeated measures so perhaps the way I'm interpreting the output is completely wrong.

#### Karabiner

##### TS Contributor
I'd like to know which groups are similar and what not.
Similar in which respect? E.g. : scores pre-post in group A: 10 -- 20. pre-post in group B: 20 -- 30 .
The groups are similar and they are not.

Regards

K.

#### victorxstc

##### Pirate
I think we share a similar problem. The Tukey exists and works well in a repeated-measures ANOVA too. The problem is that virtually no statistical packages would allow you to examine the specific groups of a factor in repeated-measures or 2-way or repeated-measures 2-way ANOVAs. The experts here recommend R as the only option to compute such Tukeys (or other post-hocs needed in your situation) (check here).