## Using Multiple Response to Combine Three Variables

Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

spssuser_70
Posts: 4
Joined: Tue Apr 10, 2012 4:06 pm

### Using Multiple Response to Combine Three Variables

Hi,

I am stumped!

I have three variables with the same 29 values and want to combine them into one. The three questions which created the three variables were "What is the most important reason for X?", "What, if any, is the second-most important reasons for X?" and "What, if any, is the third-most important reason for X?" Respondents could select one of the same 29 options for each of the three questions.

I want to combine these three variables (all numeric) into one, so that I am able to say what percentage of respondents noted the value 1 as their response to any of these three questions, the value 2 as their response, etc. and then do a crosstab of that variable (against a different variable). Since most respondents responded to all three questions (although some did not), computing a new variable through syntax does not work.

I am afraid that what I need to do is to create 29 dichotomous variables out of each of the three variables, for a total of 87 variables, and then define a multiple response set... But I was really hoping that there might be a somewhat less time-intensive way of doing so... It would be great if you could help!

Thank you!!
Penguin_Knight
Posts: 473
Joined: Thu Apr 05, 2012 5:58 pm

### Re: Using Multiple Response to Combine Three Variables

spssuser_70 wrote:Hi,

I am stumped!

I have three variables with the same 29 values and want to combine them into one. The three questions which created the three variables were "What is the most important reason for X?", "What, if any, is the second-most important reasons for X?" and "What, if any, is the third-most important reason for X?" Respondents could select one of the same 29 options for each of the three questions.

I want to combine these three variables (all numeric) into one, so that I am able to say what percentage of respondents noted the value 1 as their response to any of these three questions, the value 2 as their response, etc. and then do a crosstab of that variable (against a different variable). Since most respondents responded to all three questions (although some did not), computing a new variable through syntax does not work.

I am afraid that what I need to do is to create 29 dichotomous variables out of each of the three variables, for a total of 87 variables, and then define a multiple response set... But I was really hoping that there might be a somewhat less time-intensive way of doing so... It would be great if you could help!

Thank you!!
If it's for ANY of the three reason, you should only need 29 dichotomous variables, not 87 unless position matters.

Use COUNT. You may try the syntax below. Suppose your three response variables are called reason1, reason2, and reason3. If not, change the red parts below:

* Start of the syntax .
DEFINE !Option29 (option=!ENCLOSE("[","]")).
!DO !I !IN (!option).
COUNT !CONCAT("option", !I)=reason1 reason2 reason3(!I).
!DOEND.
!ENDDEFINE .
EXECUTE.

!Option29 option = [1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29] .
EXECUTE .
* End of the syntax .
spssuser_70
Posts: 4
Joined: Tue Apr 10, 2012 4:06 pm

### Re: Using Multiple Response to Combine Three Variables

Thank you! I've just tried that, and it seems to work... It's certainly closer than anything I have tried before.

Except... When I run frequencies for the three initial variables separately, and then for the new variables, I come up with slightly different counts for some of the options. (i.e. Option 1 is 55 if I add the counts from the three initial variables together manually, but 52 from Option 1, but for Option 2 I come up with 345 both times). Do you have any idea why?

Thank you!!
Penguin_Knight
Posts: 473
Joined: Thu Apr 05, 2012 5:58 pm

### Re: Using Multiple Response to Combine Three Variables

spssuser_70 wrote:Thank you! I've just tried that, and it seems to work... It's certainly closer than anything I have tried before.

Except... When I run frequencies for the three initial variables separately, and then for the new variables, I come up with slightly different counts for some of the options. (i.e. Option 1 is 55 if I add the counts from the three initial variables together manually, but 52 from Option 1, but for Option 2 I come up with 345 both times). Do you have any idea why?

Thank you!!
My first guess is some people might have picked #1 twice or more in their three answers. Try make a frequency table of option1 and see if there are only "1s" and "0s", or there are also "2s". If you see 2, it means across the three reasons, someone stated the same option twice.
spssuser_70
Posts: 4
Joined: Tue Apr 10, 2012 4:06 pm

### Re: Using Multiple Response to Combine Three Variables

Yes, that's it! Thank you so much.

### Who is online

Users browsing this forum: No registered users and 1 guest