topic Re: nested case-control sampling macro analysis matching in SAS Programming

nested case-control sampling macro analysis matching

Mystik — Wed, 18 Mar 2020 23:53:10 GMT

Hello,

I am running a nested case-control sampling macro analysis matching on n=10 controls.

i.e..

proc surveyselect data=m4_case&i method=srs n=10 seed=2009 out=m6_case&i;
id IDnum case_id ctrl_id dx grp_cc;
run;

However, i get these messages when the matching number is less than 10;

ERROR: The sample size, 10, is greater than the number of sampling units, 9.
NOTE: The SAS System stopped processing this step because of errors.
WARNING: The data set M6_CASE2096 may be incomplete. When this step was stopped there were 0
observations and 5 variables.

I have relaxed my matching criteria to accommodate more case-control matches as possible.

For now i want to force the sas program to accept those n=9 etc to be matched using the same procedure stated above whilst still using the n=10 criteria. Pls any help will be welcome.

Re: nested case-control sampling macro analysis matching

Reeza — Thu, 19 Mar 2020 00:41:46 GMT

Try adding the SELECTALL option without your PROC SURVEYSELECT statement.

https://documentation.sas.com/?docsetId=statug&docsetTarget=statug_surveyselect_syntax01.htm&docsetVersion=15.1&locale=en#statug.surveyselect.selectselectall

Re: nested case-control sampling macro analysis matching

Mystik — Thu, 19 Mar 2020 01:32:19 GMT

I replaced "PROC SURVEYSELECT" statement with SELECTALL as you suggested (find below) but got the error message:

6 selectall data=m4_case&i method=srs n=10 seed=2009 out=m6_case&i;

---------

180

ERROR 180-322: Statement is not valid or it is used out of proper order.

selectall data=m4_case&i method=srs n=10 seed=2009 out=m6_case&i;
id IDnum case_id ctrl_id dx grp_cc;
run;

Re: nested case-control sampling macro analysis matching

FreelanceReinh — Thu, 19 Mar 2020 11:07:43 GMT

Just adhere to the syntax described in the documentation Reeza linked to: SELECTALL is an option of the PROC SURVEYSELECT statement, so it needs to be inserted anywhere into the list of the five options you're currently using (DATA=, METHOD=, N=, SEED=, OUT=).

Re: nested case-control sampling macro analysis matching

Mystik — Thu, 19 Mar 2020 17:37:49 GMT

Thank you. I did as Reeza says and it worked.

Re: nested case-control sampling macro analysis matching

Reeza — Thu, 19 Mar 2020 17:48:59 GMT

Whoops, that was autocorrect, the original messages should have said, within not without 🙂

Glad you got it working.

Re: nested case-control sampling macro analysis matching

Mystik — Thu, 19 Mar 2020 19:22:16 GMT

Thank you Reeza!
Actually, it was your solution that actually works. Dunno why i marked mine as the solution.
Thanks again.