Dear all,
I am a Stata user (IC v. 13) and I'm doing the analyses for a longitudinal study, which aim was to test if a performance improved during time after a treatment (the treatment is only one - essentially, no improve in the performance is expected if no treatment is done).
I have 3 time of measurement (time) per each subject, where subjects are identified through an individual ID.
I wrote the command as it is a nested model (in Stata 13):
anova performance time / ID|time ID, repeated(time) bse(ID)
Is it correct to consider nesting in my case? If not, how should I run the model in Stata (given that it is the first time I'd have to do that kind of model, and I wasn't able to make any other command run)?
If you only have one group, you should be able to write the command as:
.anova performance ID time, repeated(time)
If you are more comfortable with a regression approach, you can use the -mixed- command:
.mixed performance i.time || ID: , var nolog
Then, test for the time effects:
.testparm i.time
Hope this helps,
Paul
Dear Paul,
regarding the first command line you wrote, if I remember correctly it gives an error because the software cannot identify the between-subject units... anyway I'Lloyd check on Monday!
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS