Re: st: How to tackle the r1400 error message

[Maarten buis <maartenbuis@yahoo.co.uk>]

--- On Thu, 26/8/10, Nicola Man wrote:
> > I am running into problems with numerical overflow
> > according to the r(1400) error message, but really
> > wanted to fit the multi-level model I am fitting. 
> > It was working fine with the model fitting only
> > random effects and intercept.

--- On Thu, 26/8/10, Maarten buis wrote:
> I would try rescaling your variables. So that the 
> value 1 has a reasonable unit. Say your model includes 
> the gross domestic product of countries, then you do
> not want to add that in dollars but in billions
> of dollars. So in that case you create a new
> variable by typing: 
> gen gdp_b = gdp / 10e9
> 
> and use gdp_b instead of gdp.

In addition I would rescale all variables such that the
value 0 happens withing or near the range of the data.
Say you have a variable year in there ranging from 
1910-1990, than --- combining the above advise with this
advise --- I would probably look at changes per decade 
rather than year and I would make sure that the new 
variable is 0 in 1910 by typing:

gen year_c = (year - 1910)/10

Hope this helps,
Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------


      

*
*   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/