Best test to show paired change per individual case over 4 separate survey instances (anova?, paired t-test?)

Having trouble figuring out the best statistical test to visualize outcomes of a mental health treatment intervention. Basically, I have a dataset with 4 separate survey instances (before treatment, at discharge from treatment, 6 months after discharge, and 1 year after discharge).

I would like to compare how each individual's scores changed across those 4 instances. I was thinking this was an anova test, but from looking into the formula, the anova does not match the samples, it just compares means of each group. I think I want something more like a paired samples t-test, but that can look for change over 4 instances (instead of 2).

Any help would be greatly appreciated! Also, any advice on how to best visualize the data output? Currently, I have created a graph with the mean score of each group at each instance connected by a line, but want to dive deeper to show the actual change per each case instead of just general change of means.

Thank you so much!!


TS Contributor
Have a look at repeated-measures analysis of variance. One can think of it as an expanded dependent samples t-test.
If your sample size is small, then maybe Friedman test (the "nonparametric" repeated-measures analysis).
Missing data will represent a major problem with this, by the way.

Regarding change per case, it probably depends on the number of cases and on the software you use.

With kindd regards

Thanks for the reply! Yes, the repeated-measures anova looks like the one for me. I will also look into the Friedman test. And, of course, eliminate cases with missing data across the samples.

As for visualizing change per case, I'm looking at about 1500 cases and using r. Maybe it comes down to just plotting the results of the anova?

Looks like I have more research to do on this one, but thanks for the helpful initial direction!!