I am trying to identify outliers by plotting conditionally externally studentized residual against predicted values obtained from fitting mixed model in SAS using PROC MIXED as follows:
proc mixed data = breed_trans noitprint noinfo noclprint covtest update;
by trial_type year loc trial trait;
class rep gen ;
model y = /ddfm = kr outp= dm_studenres influence;
random rep gen;
ods listing exclude solutionR fitstatistics ;
ODS output covparms= dm_para(drop = StdErr ZValue ProbZ);
where trait in("dm");
run;
quit;
The analysis was done for individual trial across different breeding stage over over year and location. My studentized seemed to be very high, varied from 10 to - 20. I would like to know if the influence option gives me the correct studentized residual. which is better to use between internally and externally studentized residual for outlier detection.
- Badges
- Science topic
- Similar topics
- Agricultural Science
- Agronomy
- Crop Science
- Crop