# How to do repeated measures ANOVA for two groups in SPSS and 12 time points?

I am measuring one parameter (say weight) repeatedly at different time points (say monthly, for 1 year) for two groups (say experimental and control groups). My data looks something like data posted in this question: Analyzing repeated measures ANOVA with two groups. I want to compare the change in weight over time 'between' the two groups. I think I should use repeated measures ANOVA for this. In SPSS, when I try do that I'm able to get 'within' group comparison of the parameter over time but not 'between' group.

Can somebody help?

In order to get the between factors box, you must have a column for "groups" and the remaining columns of time up to a year. With this said, each row will be giving information for a subject.

This testing comes under mixed ANOVA where you have repeated measures as well as between groups.

### GLM Repeated Measures

If you have no missing data and everyone is measured at each of the 12 time points, then you could use Analyze - GLM - Repeated Measures. This allows you to include a between subjects factor and the 12 time points. Your data would need to be set out in wide format with one SPSS variable for group and 12 SPSS variables for the 12 time points.

However, there are several issues with this approach. First, you are probably interested in trends over time and not whether the means for the 12 points differ. You could still run polynomial contrasts and examine the linear, quadratic and possibly cubic effects. You could examine how these interact with group to look at differential changes over time.

Then there is the issue of the structure of your residuals. Data measured over time tends to intercorrelate even after you take out any mean trend component.

### Mixed models

Thus, you may want to look into linear mixed models. E.g., see here and here. This should give you control over how you want to model the effect of time and how you model the error structure in the data. You can also consider thinks like random effects.