## Open-ends: need to count multiple instances of "Other"

0     4     Edited Aug 16, 2016 by Anonymous

We’re having a problem in Report Builder where respondents could say multiple open-ended responses to a question and it doesn’t look like the entries we coded as “other” are all being counted. At first we had each “other” coded as 26 so we then tried creating a unique value for each additional mention but that still doesn’t seem to solve the problem. We haven’t had an issue calculating the total with any of the open-ended responses we coded as particular organizations. Do you have any suggestions on ways to combine these counts? Our new syntax is below to reference.

``````pctin(or(Q10mention1==26,Q10mention2==27,Q10mention3==28),true)
``````

#### 0

##### robsmithAug 16, 2016

Assuming 3 mentions... When “other” was coded as 26, I would have written it this way:

``````pctin(Q10mention1==26 or Q10mention2==26 or Q10mention3==26,true)
``````

Then, for coded verbs, I usually find that I need to re-base to all. So I would have changed it to this:

``````pctin(ifnull(Q10mention1==26 or Q10mention2==26 or Q10mention3==26,false),true)
``````

But if ‘Other’ is multiple values, it will complicate the formula to something like this:

``````pctin(
in(Q10mention1,26,27,28) or
in(Q10mention2,26,27,28) or
in(Q10mention3,26,27,28)
,true)
``````

And if you need to re-base to all, it would be this:

``````pctin(
ifnull(
in(Q10mention1,26,27,28) or
in(Q10mention2,26,27,28) or
in(Q10mention3,26,27,28)
,false)
,true)
``````

#### 0

##### AnonymousAug 16, 2016

So the issue is that we need to double count (or triple count) some of these respondents because they could have said multiple “other” companies. Your syntax is giving us the % that said “other” in any of the 3 possible mentions (I just verified this by computing a variable that does this in SPSS and my % matches what Report Builder says) and we need the total that said “other” across the 3 mentions. I thought we could fix this by coding the “other” mentions different in each variable (i.e., 26 if they said it in 1st mention, 27 if they said it in 2nd mention), but that didn’t seem to do the trick either.

#### 0

##### robsmithAug 16, 2016

Oh! Compute the pctins individually, then add them at the end, like this:

``````pctin(Q10mention1==26,true) + pctin(Q10mention2==26,true) + pctin(Q10mention3==26,true)
``````

#### 0

##### AnonymousAug 16, 2016

Lifesaver!!! That worked perfectly.

Thanks so much!