用Enterprise Guide分析，得到了OR呢。不知对不对？

LIBNAME ECLIB000 "C:\";

/* -------------------------------------------------------------------
   由“SAS 任务”生成的代码

   生成时间: 2014年1月8日星期三 09:59:28
   任务: Logistic 回归

   输入数据: C:\www.sas7bdat
   服务器: Local
   ------------------------------------------------------------------- */
ODS GRAPHICS ON;

%_eg_conditional_dropds(WORK.SORTTempTableSorted);
/* -------------------------------------------------------------------
   对数据集 C:\www.sas7bdat 排序
   ------------------------------------------------------------------- */

PROC SQL;
        CREATE VIEW WORK.SORTTempTableSorted AS
                SELECT T.Pain, T.Sex, T.Treatment
        FROM ECLIB000.WWW as T
;
QUIT;
TITLE;
TITLE1 "Logistic 回归结果";
FOOTNOTE;
FOOTNOTE1 "由 SAS 系统 (&_SASSERVERNAME, &SYSSCPL) 于 %TRIM(%QSYSFUNC(DATE(), NLDATE20.))%TRIM(%SYSFUNC(TIME(), NLTIMAP16.)) 生成";
PROC LOGISTIC DATA=WORK.SORTTempTableSorted
                PLOTS(ONLY)=ALL
        ;
        CLASS Sex         (PARAM=EFFECT) Treatment         (PARAM=REF);
        MODEL Pain (Event = 'No')=Sex*Treatment                /
                SELECTION=NONE
                LINK=LOGIT
        ;
RUN;
QUIT;

/* -------------------------------------------------------------------
   任务代码的结尾。
   ------------------------------------------------------------------- */
RUN; QUIT;
%_eg_conditional_dropds(WORK.SORTTempTableSorted);
TITLE; FOOTNOTE;
ODS GRAPHICS OFF;

as cross effect was involved in the model, OR for one effect depends on other effects level, and versa. You can use ODDSRATIO statement to output OR for a specific effect. For example, you want OR for sex

proc ...; model ...; odsRatio sex; ods output OddsRatioWald =ORs; run;

However, you need to notice that under this situation, the OR needs to be interpreted cautiously.
JingJu

"contrast" will give what you want.

楼上两位说的很对，因为有了交互作用，计算OR值的时候就必须要考虑到交互变量是在什么水平，比如在算A vs P的OR值在sex = F 的水平上，SAS code如下：

请问最后那行的treatment 2  1  treatment*sex -2 -1是什么意思呢，如果我想做改动应该怎么改呢