The nub of the problem is this. Consider these data:
month sales (sales of Stata 10)
1 1 million
2 1 billion
3 1 trillion
Then
gen sales1 = sales if month == 1
gen sales2 = sales if month == 2
gen sales3 = sales if month == 3
creates these variables
month sales1 sales2 sales3
1 1 million . .
2 . 1 billion .
3 . . 1 trillion
Now when you average, your averages will be all
missing, as they are all based on the sum of
one value and two missings, which in Stata is missing.
You are averaging across observations. You need
to average in columns.
Check out, for example, -tssmooth-.
(If the take home exam is yours, do I get to share
your grade?)
Nick
[email protected]
P chrisman
> I have a longitudinal data set with 12 records for each subject for a
> take home final exam. I need to create an average iron dose
> measurement
> for the first months of the 12 months study period. I have tried the
> following in long format:
>
> gen ironm1=iron if month==1
> (11132 missing values generated)
>
> . gen ironm2=iron if month==2
> (11134 missing values generated)
>
> . gen ironm3=iron if month==3
> (11135 missing values generated)
>
> . gen ironBL=( ironm1+ ironm2+ ironm3)/3
> (12144 missing values generated)
>
> The problem is that my ironBL variable is "missing" data for all
> subjects rather than the average of the 3 iron measurements.
>
> I also tried the following in wide format:
>
> gen ironBL=(iron1+iron2+iron3)/3
> (4 missing values generated)
>
> I honestly do not understand what went wrong here. I thought perhaps
> egen ironBL=mean(iron) if month<=3, but that does not work either.
>
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
