<>
" Does anyone with Stata 11 get the same result?"
Yes, I do end up with a thousand zeroes after this code in Stata 11:
******
clear
set obs 1000
gen x=rnormal(100,0)
l
******
HTH
Martin
-----Ursprüngliche Nachricht-----
Von: [email protected]
[mailto:[email protected]] Im Auftrag von Feiveson, Alan
H. (JSC-SK311)
Gesendet: Freitag, 21. August 2009 15:51
An: [email protected]
Betreff: st: degenerate normal (mild annoyance)
Hi - In Stata 10, I just noticed that the function -rnormal(mu,sig)- returns
all missing values when sig = 0. I consider this a bug, since sometimes for
debugging purposes it is convenient to set a particular variance component
to zero. In other words, I would want -rnormal(mu,sig) to return a variable
with all values equal to the `mu' when `sig' = 0.
Of course, I don't have to use "rnormal" notation and could just as well
used statements such as
gen y = `sig'*rnormal(0,1) +`mu'
in my programs. But then why even bother with the "rnormal( , )" command?
(Buy the way, -drawnorm- does provide the correct output when the covariance
matrix is all zeros.)
Does anyone with Stata 11 get the same result?
Al Feiveson
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
