Yes. I would proceed as follows:
gen group=yourvariable
recode group (11/13=1) (21/23=2) (31/33=3)
sort group <---not sure this is needed
set seed ******* <---pick a number, so you can reproduce results
by group: sample 25, count
makes it give you a sample of 25 units for each group, without count it
will sample 25% of each group.
I think this'll work to give you the samples you need.
Hope this helps.
Sam
On Thu, 12 Jul 2007 [email protected] wrote:
> Dear statalisters,
>
> I have a variable in which observations take one out of 9 values:
> 11,12,13,21,22,23,24,31,32,33
>
> So, observations assume the values 1`i', 1`i', 3`i', for i = 1,2 or 3.
>
> Assuming that I have 1000 observations , I would like to:
>
> (1) randomly sample, say, 25 observations satisfying the condition ==1`i'
>
> (2) randomly sample 50 observations satisfying the condition ==2`i'
>
> (3) randomly samply another 25 observations satisfying the condition ==3`i'
>
> Then, I would like to count the frequency of each `i' in the first, second
> and last samplings.
>
> For example, considering yet the example above, assume that among the
> first sampling (25 obs), 12 observations were 1`3', 10 were 1`1' and 3
> were 1`2' (12+10+3 = 25 sampled observations).
>
> I would like to obtain the following result:
>
> `1' = 10
> `2' = 03
> `3' = 12
>
>
> Is it possible to perfom such sampling scheme using Stata?
>
> Thank you very much.
>
> Tiago
>
>
> *
> * 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/
>
*
* 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/
