Grouping non-mutually exclusive variables into categories in one variable

[aayomiposi]

Hi Everyone,

I have this survey data that includes checkboxes (select all that apply), and I am trying to create a new variable that groups all of these separate variables into a new variable. An individual can check more than one role or facility. I am new with SAS, and I know using "Else if" would not give the right frequencies (since its for mutually exclusive responses). Initially, I was going to use macro since I had a lot of variables to recategorize, but I couldn't get it to work so I decided to go with this but still not getting the right frequencies for each category. Can anyone help correct the code below, and let me know what I might be doing wrong thank you? SAS 9.4

*creating a new category called "respondent's role"*;
Data FSurvey1;
set FSurvey;
if respondent_role___1= 1 then role=1;
if respondent_role___2= 1 then role=2;
if respondent_role___3= 1 then role=3;
if respondent_role___4= 1 then role=4;
if respondent_role___5= 1 then role=5;
if respondent_role___6= 1 then role=6;
if respondent_role___7= 1 then role=7;
if respondent_role___8= 1 then role=8;

*creating a new category called "facilities description"*;
if facility_description___1= 1 then facility=1;
if facility_description___2= 1 then facility=2;
if facility_description___3= 1 then facility=3;
if facility_description___4= 1 then facility=4;
if facility_description___5= 1 then facility=5;
if facility_description___6= 1 then facility=6;
if facility_description___7= 1 then facility=7;
if facility_description___8= 1 then facility=8;
format facility facility. role role.;
run;
proc freq data= FSurvey1;
tables facility role;
run;

[PaigeMiller]

So, if for a given respondent, the value of respondent_role___1 is equal to 1 and the value of respondent_role___3 is equal to 1 and the value of respondent_role___5 is equal to 1, what result do you want?

Would this work?

proc freq data= FSurvey1;
tables facility_desc: respondent_role: ;
run;

[aayomiposi]

Thank you for your response. That worked and I got this

So, 0= unchecked, and 1= checked, meaning that for each of those variables for example respondent role__1, we have 1 and 0. All of the different variables should be in a variable as categories and not as an individual variable. So i need SAS to count everyone that checked role 1 as one category irrespective of if they checked other roles. 

[PaigeMiller]

---

@aayomiposi wrote:

So, 0= unchecked, and 1= checked, meaning that for each of those variables for example respondent role__1, we have 1 and 0. All of the different variables should be in a variable as categories and not as an individual variable. So i need SAS to count everyone that checked role 1 as one category irrespective of if they checked other roles. 

---

I don't understand these words. Categories that look like what? SHOW US what you want.

[aayomiposi]

So when I use this code
Data FoamSurvey1;
set FoamSurvey;
if respondent_role___1= 1 then role=1;
if respondent_role___2= 1 then role=2;
if respondent_role___3= 1 then role=3;
if respondent_role___4= 1 then role=4;
if respondent_role___5= 1 then role=5;
if respondent_role___6= 1 then role=6;
if respondent_role___7= 1 then role=7;
if respondent_role___8= 1 then role=8;
format role roles.
run;
proc freq data= FoamSurvey1;
tables role;
run;

I get this,

This is how I want it categorized. However, it is not correct. State fire training director should be 11 not 10, deputy fire chief should be 10 not 8. 

[PaigeMiller · Accepted solution]

Okay, I think I understand. I believe this is what you want:

Data FoamSurvey1;
set FoamSurvey;
if respondent_role___1= 1 then do; role=1; output; end;
if respondent_role___2= 1 then do; role=2; output; end;
if respondent_role___3= 1 then do; role=3; output; end;
if respondent_role___4= 1 then do; role=4; output; end;
if respondent_role___5= 1 then do; role=5; output; end;
if respondent_role___6= 1 then do; role=6; output; end;
if respondent_role___7= 1 then do; role=7; output; end;
if respondent_role___8= 1 then do; role=8; output; end;
format role roles.;
run;

[aayomiposi]

Thank you so much. It worked. I appreciate your help and patience.