st: Re: generating dataset with specific params

[Christopher F Baum <baum@bc.edu>]

wrote

After reviewing the faq's and the archive I've been unable to find 
guidance
or a module that would assist in generating data sets from scratch or at
least near it. For instance, is it possible to identify statistical
parameters, input them into stata, which would produce a data set as its
product, such as, generate a newvariable with 109 observations that 
have a
mean of 10.62 and a standev of 1.74. Maybe I've missed some guidance or
perhaps such a command is not possible (or desirable?), or this is a 
goofy
thought to start with. I reviewed -generate-, -exp-, -if-, etc but
everything appears to build on a primary data set. When reading 
articles or
text examples, one doesn't have access to the data set often, and 
sometimes
I'd like to test what's being put forward on a dataset, just simply run 
the
numbers. I have used -display- and some of the -i- commands, that's 
fine,
but one can't work from other angles what's being put forward in the
article/text, etc.


. set obs 109
obs was 0, now 109

. g double v1 = 1.74*invnorm(uniform())+10.62

. su

    Variable |       Obs        Mean    Std. Dev.       Min        Max
-------------+--------------------------------------------------------
          v1 |       109     10.6402    1.734998   6.296962   14.36916



This would work better with a larger sample:

. set obs 10900
obs was 0, now 10900

. g double v1 = 1.74*invnorm(uniform())+10.62

. su

    Variable |       Obs        Mean    Std. Dev.       Min        Max
-------------+--------------------------------------------------------
          v1 |     10900    10.61984    1.736441   3.801061   17.71306

Kit

*
*   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/