Hi there,
I am new to the STATA program (and forum), so excuse me if my question sounds unclear/vague.
I formerly used SPSS and now switched to STATA but I am absolutely no statistical genius.
I would like to do an analysis on a relatively small dataset of 53 subjects.
For instance,
Column 1 = IDvariable
Column 2 = ‘7 intervals', labeled Phase
Column 3 = Point in time (PRE-BL-FU) (PRE comes first in time, than BL, than FU), labeled Rhythm
Column 4= is the parameter which I want to evaluate labeled AA
2. In the next step I would like to subdivide results even further basesd on column 5, labeled EROA --> this is a binary variable)
3. In the final step I would like to draw the longitudinal changes between the seven intervals with mean+SD with three separate curves (the three time points).
(Ideally I’d like to draw 8 points with point 8 again being identical to point 1 to complicate things)
I hope I am a little bit clear.
Thank you very much in advance,
I am new to the STATA program (and forum), so excuse me if my question sounds unclear/vague.
I formerly used SPSS and now switched to STATA but I am absolutely no statistical genius.
I would like to do an analysis on a relatively small dataset of 53 subjects.
- Each subjects had echocardiography measurements on 3 separate time points (PRE-BL-FU titled in my database)
- At each time point, 7 longitudinal measurements were made (meaning I measured the longitudinal change in an echocardiographic parameter on 7 fixed intervals and this was repeated at three different time points PRE-BL-FU)
- I would like to view these changes between the 7 intervals but also between the three time points for each interval.
- I already switched from a wide to a long dataset. I was planning to use repeated ANOVA for the changes between the 7 intervals with posthoc Bonferroni testing and do the same for each interval separately for the three time points, also with post hoc Bonferroni testing.
- Now I read that it is better to do mixed modelling for this purpose? Collegues told me to use mixed modelling with FDR post hoc testing?
For instance,
Column 1 = IDvariable
Column 2 = ‘7 intervals', labeled Phase
Column 3 = Point in time (PRE-BL-FU) (PRE comes first in time, than BL, than FU), labeled Rhythm
Column 4= is the parameter which I want to evaluate labeled AA
2. In the next step I would like to subdivide results even further basesd on column 5, labeled EROA --> this is a binary variable)
3. In the final step I would like to draw the longitudinal changes between the seven intervals with mean+SD with three separate curves (the three time points).
(Ideally I’d like to draw 8 points with point 8 again being identical to point 1 to complicate things)
I hope I am a little bit clear.
Thank you very much in advance,
Comment