Prev: Help needed regarding the urgent task.
Next: Histogram with gplot
From: Yaw on 25 May 2010 23:25
Dear all:

I have this logistic model that is producing very high odds ratio. I
have checked(& purged) all my predictors for missing values thinking
they maybe to blame; but not headway.


Has anybody encountered this problem before?

My working code:

proc logistic data=girls;
class race mstat(ref='2');
model abtg(event='1')= ageyrs ageyrs*ageyrs mstat race race*ageyrs/
expb;
*weight wtvar;
run;

Any help is appreciated.

Thanks,
-Yaw
From: Ryan on 26 May 2010 15:07
On May 25, 11:25 pm, Yaw <link...(a)gmail.com> wrote:
> Dear all:
>
> I have this logistic model that is producing very high odds ratio. I
> have checked(& purged) all my predictors for missing values thinking
> they maybe to blame; but not headway.
>
> Has anybody encountered this problem before?
>
> My working code:
>
> proc logistic data=girls;
>         class  race mstat(ref='2');
>         model abtg(event='1')= ageyrs ageyrs*ageyrs mstat race race*ageyrs/
> expb;
>         *weight wtvar;
>         run;
>
> Any help is appreciated.
>
> Thanks,
> -Yaw

Yaw,

I can't speak to your specific situation, but generally speaking, odds
ratios can get extremely high when the probability of event for one of
the groups is very low.

Ryan
 | 
Pages: 1
Prev: Help needed regarding the urgent task.
Next: Histogram with gplot