Help with selecting type of ANOVA

[Theodore Habarth]

Hello statalisters,

I am trying to perform an analysis on a dataset that contains information on patient values (r2-r4) from 4 different time points (denoted by the variable fu). These r variables each identify a different body part and the values are electrical signal. Patients are denoted by var recordid. I am trying to perform ANOVA to compare the means between two groups of patients (identified by variable innervation) of a body part over the period of time. I am aware that I should be using the repeated measure ANOVA but am not quite sure how to incorporate comparing the two groups of innervated vs non-innervated patients. Is this supposed to be nested? I have read the ANOVA help file but am not entirely sure based on the examples.

Any help is greatly appreciated.

* Example generated by -dataex-. For more info, type help dataex
clear
input int recordid byte innervation float(r2 r3 r4) str5 j float fu
29 1 26.6 37.7 26.5 "_12mo" 4
29 1 96.5  100 42.2 "_6mo"  3
29 1  100  100 74.3 "_3mo"  2
30 0 36.3  100  3.9 "_3mo"  2
31 1 64.2 93.3 17.4 "_6wk"  1
31 1 12.7 75.4 20.6 "_3mo"  2
32 0 95.8 98.2 61.2 "_6wk"  1
33 0 75.8 99.9 98.7 "_6wk"  1
33 0  100  100   84 "_3mo"  2
33 0  100  100  100 "_12mo" 4
end

[Joseph Coveney]

If that little snippet is representative of your dataset, then you cannot use conventional repeated-measures ANOVA, because you don't have an equal number of follow-up observations between patients. Try mixed instead.