DATA Step, Macro, Functions and more

Should we submit raw data or residuals for PROC(EDURE)s in SAS?

Accepted Solution Solved
Reply
Contributor
Posts: 57
Accepted Solution

Should we submit raw data or residuals for PROC(EDURE)s in SAS?

For example, we can test homogeneity of variance with PROC GLM.

The procedure calculates residuals as we can see it in the formula for Levene’s HOV test.

 

Is this implicit residual calculation common situation for SAS procedures or not ?

Can we save the residuals computed by a PROC into a table ?

 

Thank you.


Accepted Solutions
Solution
‎09-12-2016 06:50 AM
SAS Super FREQ
Posts: 3,475

Re: Should we submit raw data or residuals for PROC(EDURE)s in SAS?

SAS regression procedures enable you to output residuals (and standardized residuals, and predicted values, and confidence limits, and...) by using the OUTPUT statement.  Different procedures support different keywords, as applicable.  

 

To get the residuals in PROC GLM, add this line to your procedure call:

output out=MyOutput R=Residuals;

That will create an output data set calls "MyOutpuit" which contains the input data and a new column called "Residuals" that contains he residual value for each observation.

 

View solution in original post


All Replies
Solution
‎09-12-2016 06:50 AM
SAS Super FREQ
Posts: 3,475

Re: Should we submit raw data or residuals for PROC(EDURE)s in SAS?

SAS regression procedures enable you to output residuals (and standardized residuals, and predicted values, and confidence limits, and...) by using the OUTPUT statement.  Different procedures support different keywords, as applicable.  

 

To get the residuals in PROC GLM, add this line to your procedure call:

output out=MyOutput R=Residuals;

That will create an output data set calls "MyOutpuit" which contains the input data and a new column called "Residuals" that contains he residual value for each observation.

 

Contributor
Posts: 57

Re: Should we submit raw data or residuals for PROC(EDURE)s in SAS?

Thank you Smiley Happy.

 

The kindly advised code line outputs the residuals r_ij = y_ij - mean(y_ j), where j is a group indicator.

And if I'm not mistaken, they are the same as the ones calculated for and used by Levene's HOV test.

But the OUTPUT statement doesn't allow us to see another types of residuals used by other HOV test, do we ?

 

And by the way there are RSTUDENT, STUDENT keywords in the OUTPUT statement.

What are they used for?

☑ This topic is SOLVED.

Need further help from the community? Please ask a new question.

Discussion stats
  • 2 replies
  • 185 views
  • 1 like
  • 2 in conversation