Repeated measures data manipulation

PBalte — Tue, 28 Nov 2017 21:35:06 GMT

Hello all,

I am working with a very large (~65,000 observations) repeated measures data set. Currently, the data is setup in long format. How do I create an indicator variable based on change in variable at different timepoints?

This is what I have:

ID Count Smk_status

1 1 Never

1 2 Never

2 1 Former

2 2 Former

2 3 Former

3 1 Never

3 2 Current

3 3 Former

This is what I want:

ID Count Smk_status Smk_trajectory

1 1 Never Never

1 2 Never Never

2 1 Former Former

2 2 Former Former

2 3 Former Former

3 1 Never Inconsistent

3 2 Current Inconsistent

3 3 Former Inconsistent

I can do this in wide format of data but I am just wondering if anyone has any suggestions to do this in long format of data.

Thanks in advance!

Re: Repeated measures data manipulation

Reeza — Tue, 28 Nov 2017 23:12:11 GMT

SQL is pretty good at this one.

proc sql;
create table want as
select *, case when max(smk_status)=min(smk_status) then min(smk_status)
              else 'Inconsistent' end as smk_trajectory
from have
group by ID,
order by ID, count;
quit;

topic Repeated measures data manipulation in SAS Procedures

Repeated measures data manipulation

Re: Repeated measures data manipulation