Creating New Variable

[GregBond]

Hi All,

I would like to create a new variable (health) using 3 variables which are already in the dataset (q1 q2 and q3). The three variables I am creating this new variable from all have the same ordinal response options. (How long in hours have you missed work, due two three separate conditions)

After creating this new variable(health) from the 3 variables,  I would like to get its mean, SD, CI. 

Thank you in advance for any advice you may have!

Regards

[PaigeMiller]

Number hours of missed work seems like it is numeric, not ordinal.

But, we cannot tell you how to combine the 3 variables, this is your problem, and so you have to decide how to combine them, and then we might be able to advise on how to code it in SAS.

[GregBond]

Thanks for your response Paige!

Combining the variables through addition, e.g. health = Q54+Q55+Q56.

Then, I would like to get the mean, SD, and CI of the 'health' variable sum.

[Reeza · Accepted solution]

proc means data=have n mean std ;

var health;

run;

[GregBond]

thanks for your response again, when running this code...

data work.CAREGIVER;

set work.caregiver;

length health $44;

keep health ;

health = Q54+Q55+Q56;

run;

proc means data=work.CAREGIVER n mean std;

var health ;

run;

I get the following error...

  proc means data=work.CAREGIVER_SPSS n mean std;var house;

ERROR: Variable house in list does not match type prescribed for this list.

  run;

[Reeza]

That's because you've created a character variable above, not a numeric variable.  Why do you have the length statement?

[PaigeMiller]

Yes, you shouldn't need the length statement, and I am still concerned that you consider the number of hours missed to be "ordinal".

Are the values of Q54 and Q55 and Q56 actual numbers, as in SAS variables that are numeric?