Survival analysis - Schoenfeld residual

[Jmbuote]

Hello,

Trying to test the proportional hypothesis. Can anyone tell me what is wrong?

PROC PHREG DATA=CBD.MERGE_CBD_CB1;
MODEL T_OS*OS (0) = MEAN_SC_CB1 SENT_cat cms_AGE cms_TABAC / TIES=EFRON;
WHERE 1 <= N_CORE_CB1 <= 4;
OUTPUT OUT = cbd.resschoe_CB1 wtressch=schMEAN_SC_CB1 schSENT_cat schcms_AGE schcms_TABAC;
RUN;

PROC SORT DATA=cbd.merge_cbd_cb1;
BY T_OS;
RUN;
PROC SORT DATA=cbd.resschoe_CB1;
BY T_OS;
RUN;
DATA cbd.resschoe_CB1_END;
MERGE cbd.merge_cbd_cb1 cbd.resschoe_cb1;
WHERE 1 <= N_CORE_CB1 <= 4;
BY T_OS;
LOGT=LOG(T_OS)
RUN;

PROC SORT DATA=cbd.resschoe_cb1_end;
BY T_OS;
RUN;

*****This section is not working:*******

DATA CBD.RESSCHOE_CB1_END;
SET CBD.RESSCHOE_CB1_END;
WHERE 1 <= N_CORE_CB1 <= 4;
LOGT=LOG(T_OS);

RUN;

***It says: NOTE: Invalid argument to function LOG(0) at ligne 76 colonne 6.***

Cheers,

J

[Reeza · Accepted solution]

That's not an error, it's a message. It should also tell you the value of T_OS at that line and if you check it, it's likely not valid for the LOG fuction - most likely missing T_OS or overall data.