How to recode continuous variables

Denali — Tue, 04 Dec 2018 20:20:02 GMT

I have 15 continuous variables and the data entry person defined "888" and "999" as missing. I don;t want these two numbers to be calculated in the proc means procedure as they appeared to be the maximum values. Is there a way to recode 888 and 999 as . ?

Could you please coach me how to do the data cleaning step in the data step? Thank you!

Attached is the data and below are the variables I was using.

%macro one(x) (where=(wrat_raw not in (888 999)));
proc means n mean std median min max maxdec=1;
var &x;
class site_id Phase visit_id;
run;
%mend one;
%one (wrat_raw);
%one (wrat_ss);
%one (wrat_perc);

%one (bacs_symbol_raw);
%one (bacs_standardized);
%one (bacs_percentile);

%one (hvlt_trial1);
%one (hvlt_trial2);
%one (hvlt_trial3);
%one (hvlttot);
%one (hvlt_standardized);
%one (hvlt_percentile);

%one (msceit_raw);
%one (msceit);
%one (msceitpercentile);

Re: How to recode continuous variables

Astounding — Tue, 04 Dec 2018 20:51:27 GMT

You can recode them in a DATA step easily enough. But I would suggest using special missing values, rather than . so here is an example using 3 variables:

data want;

set have;

array nums {*} wrat_raw wrat_ss wrat_perc;

do k=1 to dim(nums);

if nums{k} = 888 then nums{k} = .A;

else if nums{k}= 999 then nums{k} = .B;

end;

drop k;

run;

The special missing values will automatically be excluded from PROC MEANS calculations. But you still have a record of what the value used to be, if that ever becomes important to know.

topic How to recode continuous variables in SAS Programming

How to recode continuous variables

Re: How to recode continuous variables