st: RE: multiple repeated measures

[Nick Cox <n.j.cox@durham.ac.uk>]

Without commenting on whether this is a good way to do it, an answer to your Q1 is 

gen twoweek = cond(mod(week,2) == 1, week, week-1) 
egen max_myvar = max(myvar), by(id twoweek) 

The first statement is likely to be seem more cryptic. It keeps odd values as they are are and bumps down even values. So (1,2) -> 1; (3, 4) -> 3; etc. 

(mod(week, 2) == 1)  is equivalent to mod(week, 2). 

Depending on how many subjects you have, just plotting the data sounds a good idea. Missing values won't show, which is right. 

Nick 
n.j.cox@durham.ac.uk 

laura hendrix

I have a 14-item survey that was administered weekly to a group of
subjects for 8 weeks with responses to each question ranging from 1-5
with some subjects having missing weeks and/or missing responses:

              Week1                    Week2
------------------------------------------------------------------
            Question                   Question
------------------------------------------------------------------
 ID         1  2  3  4  5  6  7 8       1  2  3  4  5  6  7  8
----------------------------------------------------------------------------
1           1  3  .  2  2  1  5  .        2  4  2  2 .  3  4  5
2           4  4  4  3  3  .  1  1       2  3  2  2  4  5  1  1

and so on.

I would like to determine whether there has been a significant change
in score for each question between weeks 1 and 8.  Without going into
a complex longitudinal analysis which I have no idea how to do, how
can I

1.  Group questions into 1-2  3-4  5-6  7-8 and generate a new
variable containing the highest value in each group (counting
missing=0)
2.  determine whether there has been a significant change in score for
each item between weeks 1 and 8
??


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/