topic Re: Calculating %RSD of many variable at once in SAS Programming

Calculating %RSD of many variable at once

Job04 — Sun, 27 Feb 2022 20:52:52 GMT

My data has dimension of 80x200

patients in rows and variables in column.

I want to calculate the %RSD of each variable.

%RSD= (standard deviation of each variable / mean of variable)*100

The data looks like this:

ID,gender,AL,BK,MS,DF
2,F,13.4,7.3,-15.2,9.2
3,M,12.1,5.1,17.2,8.85
5,M,10.8,2.9,49.6,8.5
7,F,9.5,0.7,82,8.15
Thank you

PaigeMiller — Sun, 27 Feb 2022 21:12:51 GMT

This is essentially what SAS calls the CV (coefficient of variation) except it is not multiplied by 100.

proc summary data=have;
    var al bk ms df;
    output out=cv cv=/autoname;
run;

Job04 — Sun, 27 Feb 2022 21:17:42 GMT

My data has 200 variables. I need that on all of them.

PaigeMiller — Sun, 27 Feb 2022 21:23:07 GMT

Use a list of consecutive variables names, such as

al--bk

where al is the first variable name and bk is the last variable name; the double dash indicates that the variables in the list are consecutive.

Reeza — Sun, 27 Feb 2022 21:21:49 GMT

proc summary data=have;
   var _numeric_;
    output out=cv cv=/autoname;
run;

Using _numeric_ will run it on all numeric variables.

@Job04 wrote:

My data has 200 variables. I need that on all of them.