Desktop productivity for business analysts and programmers

Highlighting outliers

Posts: 58

Highlighting outliers



I notice in my dataset that there appears to be a few outliers which in turn can throw out headline figures. They appear to be related to keying errors. 


What I would like to do is include in my datastep an if statment that will mark an outlier with a Y if it 3 times higher than the average level of turnover in that sector. 


It would be something simple like this;


data i;
set m;
if NET_OPERATING_INCOME_1*(3)>mean_operating_income then outlier='Y';run; 

However, what this does is actually just multiply the variable net_operating_income *3 whereas I just want SAS to mark the observation with a Y anytime a value in net_operating_income_1 is 3 times larger then mean_operating_income.


Any input would be most welcome.


KInd regards,



Super User
Super User
Posts: 9,840

Re: Highlighting outliers

Posted in reply to Sean_OConnor
data i;
  set m;
  outlier=ifc(net_operating_income_1 > (mean_operating_income * 3),'Y','N');

You were nearly there - note I use the ifc function rather than if construct for brevity. 

Ask a Question
Discussion stats
  • 1 reply
  • 2 in conversation