I'd modify your -forval- loop
and build up a list of names one
by one (add semi-colons as desired)
forvalues i = `start'/`end' {
sum income if age==(`i'+startage-1)
scalar age`i' = r(mean)
if `i' < `end' local ages "`ages'age`i',"
else local ages "`ages',age`i'"
}
Then
matrix data = startage,endage,male,female,`ages';
Nick
[email protected]
Incidentally, it looks as if your -summarize- should
be done ", meanonly".
Friedrich Huebler
>
> I work with household survey data and would like to create summary
> statistics disaggregated by gender, age and other characteristics. In
> my do-file the summary statistics are stored as scalars and then
> converted to a matrix. I encountered a problem with the disggregation
> by age because I use varying age ranges. Let's assume I want to
> summarize income by gender and age (for ages 20-24 in this case) with
> the following data.
>
> age male income
> 20 1 6
> 21 1 7
> 22 0 5
> 23 0 9
> 24 1 6
>
> * Set start and end age (NOTE: THE AGES VARY);
> scalar startage = 20;
> scalar endage = 24;
> * Summarize income by gender;
> sum income if male==1;
> scalar male = r(mean);
> sum income if male==0;
> scalar female = r(mean);
> * Summarize income by age;
> local start = 1;
> local end = (endage - startage + 1);
> forvalues i = `start'/`end' {;
> sum income if age==(`i'+startage-1);
> scalar age`i' = r(mean);
> };
>
> The problem is in the following step. Because I use varying age
> ranges I cannot list all age variables individually (age1, age2,
> age3, ...) but want to refer to them with a wildcard character.
>
> drop _all;
> matrix data = startage,endage,male,female,age*;
> svmat data;
>
> This leads to this error message:
>
> age* not found
> r(111);
>
> Can the matrix be created without listing all scalars individually?
*
* 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/
