|
|
 | | From: | K Fernandes | | Subject: | Proc Logistic | | Date: | 21 Jan 05 20:58:13 GMT |
|
|
 | Hello,
I am experimenting with Proc Logistic and I have the typed the following
code:
proc logistic desc data=merging; /*Logistic Regression*/
model y/m = x ;
ods output ParameterEstimates = BetaEsts;
output out=results p=predict;
by Rand Sim_No;
run;
proc logistic data = merging; /*Logistic Regression*/
model y/m = x ;
ods output ParameterEstimates = BetaEsts1;
output out=results2 p=predict;
by Rand Sim_No;
run;
Each proc logistic is producing the same output, which confuses me since the
first proc logistic has a desc and the other one doesn't. So I would think
that my parameter estimates for x would be different, but the output from
both statements is identical. Could someone shed light on why this might
be?
Thank you for your help,
K
|
|
| | From: | Robin High | | Subject: | Re: Proc Logistic | | Date: | 21 Jan 05 21:40:26 GMT |
|
|
| >
> I am experimenting with Proc Logistic and I have the typed the following
> code:
>
> proc logistic desc data=merging; /*Logistic Regression*/
> model y/m = x ;
> ods output ParameterEstimates = BetaEsts;
> output out=results p=predict;
> by Rand Sim_No;
> run;
>
> proc logistic data = merging; /*Logistic Regression*/
> model y/m = x ;
> ods output ParameterEstimates = BetaEsts1;
> output out=results2 p=predict;
> by Rand Sim_No;
> run;
>
> Each proc logistic is producing the same output, which confuses me since the
> first proc logistic has a desc and the other one doesn't. So I would think
> that my parameter estimates for x would be different, but the output from
> both statements is identical. Could someone shed light on why this might
> be?
The descending option only works with binary data (0 or 1) entered as the
response variable -- it doesn't have an effect with y/n = .....
Robin High
Univ. of Oregon
|
|
|
