DATA Step, Macro, Functions and more

controlling for variables

Accepted Solution Solved
Reply
Contributor
Posts: 69
Accepted Solution

controlling for variables

Hello

 

I am writing after trying to understand it but I am not sure with what I found so decided to ask here.

 

I want to use PROC MIXED as I have fixed effect and also random effects in my model. Which statement in SAS reprents the CONTROL VARIABLES in my model? is this the CLASS statement?

 

Thanks for any help :-)


Accepted Solutions
Solution
‎06-02-2016 11:15 AM
Trusted Advisor
Posts: 2,113

Re: controlling for variables

Yael,

 

Questions like this are better suited for the Statistical Procedures forum.

 

You should start by reading the entire chapter on PROC MIXED.  The CLASS statement is for identifying variables that are represented by levels (e.g. race) rather than on a continuium (e.g. age); this is the way that the CLASS statement is used in all of the SAS stat procedures.

 

"Control" variables are another type of independent variable.  The SAS procedures don't treat them any differently, but the analyst thinks about them differently.  I see the term most often in epidemiology.  They are usually independent variables that effect the outcome but aren't the topic of interest in the analysis (e.g. In a general regression framework, someone might want to "examine the effect of BP on heart rate controlling for age.").  In MIXED, most control variables are fixed effects.

View solution in original post


All Replies
Solution
‎06-02-2016 11:15 AM
Trusted Advisor
Posts: 2,113

Re: controlling for variables

Yael,

 

Questions like this are better suited for the Statistical Procedures forum.

 

You should start by reading the entire chapter on PROC MIXED.  The CLASS statement is for identifying variables that are represented by levels (e.g. race) rather than on a continuium (e.g. age); this is the way that the CLASS statement is used in all of the SAS stat procedures.

 

"Control" variables are another type of independent variable.  The SAS procedures don't treat them any differently, but the analyst thinks about them differently.  I see the term most often in epidemiology.  They are usually independent variables that effect the outcome but aren't the topic of interest in the analysis (e.g. In a general regression framework, someone might want to "examine the effect of BP on heart rate controlling for age.").  In MIXED, most control variables are fixed effects.

Contributor
Posts: 69

Re: controlling for variables

Thanks a lot for your answer, it really helps me.

☑ This topic is solved.

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

Discussion stats
  • 2 replies
  • 218 views
  • 0 likes
  • 2 in conversation