st: Re: Generating correlated data with some unspecified correlations

[Andreas Baumann <andreas.baumann.privat@gmail.com>]

Hi all,

This might be dumb, but I'm at my wit's end. If anyone could help me, it
would be much appreciated.

I'm generating some data using drawnorm for use in a simulation model.
My model consists of six variables (m, f, m_i, f_i, r & r_i).
Some of these variables should be correlated: (m,f), (m,m_i), (f,f_i),
(f,r), (m,r) and (r,r_i).
The purpose is then to ascertain the variability of corr(m_i,r_i) and
corr(f_i,r_i).

However, if I use drawnorm to generate the data, I'm forced to specify the
correlation between these variables - which is exactly what I'm interested
in investigating. So, obviously, that is no good. I was wondering if there
was some way in which I could generate randomly distributed numbers with the
specified correlations without specifying the correlations mentioned above.

Thank you for taking the time to read this.

Best,
Andreas Baumann
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/