topic Re: Help separating categories in proc surveymeans in SAS Procedures

Help separating categories in proc surveymeans

CJxyz — Fri, 28 Apr 2023 16:41:05 GMT

Hello,

I'm trying to get into SAS programing and I'm having a lil trouble separating categories in my variable age (<15, 16-17, >18 years old). I need to used proc surveymeans but the output only shows all 3 categories together and 1 mean in 1 row. How you I get the output to show 3 means in the separate rows?

Thank you for helping

Code:

age_new =.;

if age le 15 then age_new =1;
if age = 16 or age = 17 then age_new =2;
if age ge 18 then age_new =3;

label
age_new = "1=15 or Under" "2=Between 16-17" "3=18 or Older"

run;

proc format;
value age_new 1='1=15 or Under' 2='2=16 or 17' 3='3=18 or Older';

run;

proc surveymeans data=syc mean clm sum std stderr clsum plots=none;
stratum stratum;
cluster psu;
weight finalwt;
var age_new;

run;

Re: Help separating categories in proc surveymeans

ballardw — Fri, 28 Apr 2023 18:03:33 GMT

Your code requested the mean and other statistics of a variable. It does nothing to use levels.

In Surveymeans variables that used for categories belong in a DOMAIN statement. At which point the mean for age_new will be the value of age_new. Are you sure that you aren't looking for SURVEYFREQ to estimate percentages of age_new?

Or is there another actual variable that you want the mean and other statisics for such as height, test score or such that would belong on the VAR variable?

If you want the actual mean Age within the Age_new group that would be:

proc surveymeans data=syc mean clm sum std stderr clsum plots=none;
stratum stratum;
cluster psu;
weight finalwt;
domain age_new;
var age;

run;