Re: st: seed command with rndbin

[Roger Harbord <roger.harbord@bristol.ac.uk>]

--On 12 August 2003 15:07 +0100 "Allardice, Gwen" 
<Gwen.Allardice@scieh.csa.scot.nhs.uk> wrote:


> I have recently found a problem in  stata v7 . I have been using the
> rndbin program extensively in some simulations I am running.
>
> I would like to be able to replicate  my results and find that the seed
> command does not seem to work; I keep getting different results.
>
> e.g. a simple example should illustrate my problem. the following three
> lines of stata will generate and show me an appropriate random number but
> repeating the three commands then gives me a different number. I want it
> to give me the same number.
>
>
>
> set seed 2
> rndbin 1 0.33  6
> display   xb
>
> Can someone show me how to run the rndin command more than once and get
> the same result each time.
Thanks for pointing this out, I'd just assumed I was getting reproducible 
pseudorandom numbers when in fact I wasn't - i hadn't thought to check!

If you check the start of the code for -rndbin-, you'll find it's written 
in version 3.1.

If you check the help page for -version- and search for "seed" you'll find 
this:

If you set version to 3.1 or less, the following will be true
       1.  uniform() refers to the old random-number generator
       2.  set seed sets the old random-number seed


The solution is now simple:

version 3.1: set seed 2
rndbin 1 0.33  6
display   xb

- gives the same each time.  Unfortunately you can't replicate results 
you've already got..

Roger.
----------------------------------------------------
Roger Harbord     mailto:roger.harbord@bristol.ac.uk
Department of Social Medicine, University of Bristol

*
*   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/