Error in estimate and contrast statements - need help interpreting error

Accepted Solution Solved
Reply
New Contributor
Posts: 2
Accepted Solution

Error in estimate and contrast statements - need help interpreting error

All,

I am a student and keep getting an error when setting up estimate and contrast statements. I'm trying to build a model for the association between mean glucose and exercise status, and control for possible confounding. Below is the log, and I keep getting this error message "Effects used in the ESTIMATE statement must have appeared previously in the MODEL statement." I don't understand why I keep getting this - the variables 'waist' and 'diabetes' are in the model statement. Does anyone know what I'm doing wrong/can help me out?

Thanks!

89   proc glm data=testdata1;

90   class diabetes;

91   model glucose = waist diabetes/solution;

92   estimate 'diabetes: waist = 80' diabetes 1 -1 waist*diabetes 80 -80;

ERROR: Effects used in the ESTIMATE statement must have appeared previously in the MODEL

       statement.

NOTE: The previous statement has been deleted.

93   estimate 'diabetes: waist = 110' diabetes 1 -1 waist*diabetes 110 -110;

ERROR: Effects used in the ESTIMATE statement must have appeared previously in the MODEL

       statement.

NOTE: The previous statement has been deleted.

94   estimate 'diabetes: waist = 140' diabetes 1 -1 waist*diabetes 140 -140;

ERROR: Effects used in the ESTIMATE statement must have appeared previously in the MODEL

       statement.

NOTE: The previous statement has been deleted.

95   contrast '80 vs. 90'  diabetes 1 -1 waist*diabetes 80 -80, diabetes 1 -1 waist*diabetes 90

95 ! -90;

ERROR: Effects used in the CONTRAST statement must have appeared previously in the MODEL

       statement.


Accepted Solutions
Solution
‎02-04-2013 02:17 PM
Respected Advisor
Posts: 3,777

Re: Error in estimate and contrast statements - need help interpreting error

Yes the variables are there but they only define the main effects effect for each variable.  You need to include the interaction effect waist*diabetes.

you can use vertical bar
glucose = waist | diabetes;

View solution in original post


All Replies
Solution
‎02-04-2013 02:17 PM
Respected Advisor
Posts: 3,777

Re: Error in estimate and contrast statements - need help interpreting error

Yes the variables are there but they only define the main effects effect for each variable.  You need to include the interaction effect waist*diabetes.

you can use vertical bar
glucose = waist | diabetes;

New Contributor
Posts: 2

Re: Error in estimate and contrast statements - need help interpreting error

Thank you. I updated that line to model glucose = waist exercise waist*exercise/solution;

It worked.

☑ This topic is SOLVED.

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

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