topic Re: CREATING A NEW VARIABLE FROM A "CHECK ALL THAT APPLY" VARIABLE in SAS Programming

CREATING A NEW VARIABLE FROM A "CHECK ALL THAT APPLY" VARIABLE

_maldini_ — Tue, 07 Aug 2018 20:12:23 GMT

I am analyzing data from a survey with questions whose answers are in the form of checkboxes (i.e. "Check all that apply"). In the raw data, each checkbox (i.e. Response) is its own variable (See below: addiction___1 is the first checkbox, addiction___2 is the second checkbox, etc.). I would like to create a new variable for each question and have each checkbox / response be its own category / level in that variable.

The code below works fine when I am coding one question at a time. Is there a way to automate / condense the syntax so I don't have to write this syntax for each question over and over again?!

	IF addiction___1 THEN DO;
	addiction = 1;
	
	END;
	
	IF addiction___2 THEN DO;
	addiction = 2;
	
	END;
	
	IF addiction___3 THEN DO;
	addiction = 3;
	
	END;
	
	IF addiction___4 THEN DO;
	addiction = 4;
	
	END;
	
	IF addiction___5 THEN DO;
	addiction = 5;
	
	END;

Re: CREATING A NEW VARIABLE FROM A "CHECK ALL THAT APPLY" VARIABLE

Astounding — Tue, 07 Aug 2018 20:27:50 GMT

I'm not sure that this code is really working fine. When both 1 and 2 are checked, your new variable ADDICTION will be 2 and won't capture the fact that 1 was also checked.

Re: CREATING A NEW VARIABLE FROM A "CHECK ALL THAT APPLY" VARIABLE

VDD — Tue, 07 Aug 2018 20:42:52 GMT

You should use the /debug on your data statement.

See what is happening yourself.

data fun / debug;

/* lots of stuff you want to check */

run:

Re: CREATING A NEW VARIABLE FROM A "CHECK ALL THAT APPLY" VARIABLE

hashman — Tue, 07 Aug 2018 21:47:53 GMT

@_maldini_:

You're saying "I would like to create a new variable for each question", yet it's not what your code is doing. Rather, it overwrites the value of the ADDICTION variable with a new value X every time the next addiction__X evaluates true. If this is what you really want, you can shorten your code by using:

array add addiction__1-addiction__5 ;

do over add ;
  if add then addiction = _i_ ;
end ;

This way, if none of the boxes is checked, the variable ADDICTION will remain missing.

However, if instead you want, as you say, to create a new variable for each box and populate variable #1 with 1 is addiction__1 is true, variable #2 - with 2 if addiction__2 is true, and so forth, you'll have to do it differently; for example:

array add addiction__1-addiction__5 ;
array new newvar_1-newvar_5 ;

do over add ;
  if add then new = _i_ ;
end ;

This way, if you had the addiction__1-addiction__5 variables valued, for example, as:

(unchecked, checked, unchecked, unchecked, checked)

you'd get newvar_1-newvar_5 populated as:

(. 2 . . 5)

But the approach depends on what you really want to get. Unfortunately, you're a bit vague on that head, and discerning your intent from your sample code is uncertain.

Paul D.

Re: CREATING A NEW VARIABLE FROM A "CHECK ALL THAT APPLY" VARIABLE

ballardw — Wed, 08 Aug 2018 15:08:52 GMT

You may want to visit this thread: https://communities.sas.com/t5/Base-SAS-Programming/consolidating-columns/m-p/484770#M125874

Which has a better description of starting data and I believe at heart is similar to your problem with possibly yours not having an issue with identical responses appearing in multiple variables.

Or perhaps you can provide a more explicit example of existing data set and what you want.