Hi,
In our survey, we have 3 branches based on the frequency of the usage of a product (Never, a few times, and frequent).
In the "a few times" and "never" branch, there is a text set question "do you wish to receive more info on how to use the product?". So the same question phrase occurs in both branches, but they are two different "questions" technically (so QID5 and QID10 let's say).
(yes, we should have put that question into a block, and repeated the same block in both branches, but the survey is already done)
Likewise in the "a few times" and "frequent" branches, there is a multi-answer question "what type of data are you looking for in the product"? that was asked in two different QID's.
How can we combine such questions for use in the dashboard? A Measure Group doesn't work as they are not numeric.
We want to have a widget that shows a breakout of the "what type of data are you looking for in the product" by the usage Frequency.
(so to see that people looking for Data_A are mostly frequent users, wile people looking for Data_B are mostly "a few times" users).
Likewise also for the "do you want more info", we want to see how many people want more info per usage frequency.
How to combine multiple Text Set or Multi-Answer questions for dashboards
+27Best answer by chasemcl
JoycaV yes that is the unfortunate part - you have to import the responses and then delete out all the old responses. I have done this several times as it often is the most efficient way of doing it. But it is simple enough to just delete all responses with a recorded date before the import. Just be sure to use start or end date rather than recorded date if you need it for reporting purposes
You could create an embedded data field and then at the end of the survey flow merge the responses to the 2 question into that field. It would look like this:
This would record the data in one field going forward. Moving the previously collected data into this field may have to be manual. I am not sure what the most efficient way would be without knowing the data set.
Hope this helps!
+27Thanks @chasemcl, good idea for a next time. But our survey is already done. What would be the fastest way to update all our existing responses? I think in the Data&Analysis, an export, and then an import but with the extra column of that embedded data field filled in?
I just noticed in the Data&Analysis that there's a new tab under "Create New Field" called combine:
Combining Fields - Qualtrics SupportThat's new I think? Seems exactly what I need though! 😁
Edit: only for text set fields... the multi-answerfields are not supported by the Combine function :(
+27Oh no, my idea for the multi-text fields won't work either...
So export all data, update the new combined field, import all data.
Apparently when you do that, you get duplicate replies (even if you map the Response ID field) :(
JoycaV yes that is the unfortunate part - you have to import the responses and then delete out all the old responses. I have done this several times as it often is the most efficient way of doing it. But it is simple enough to just delete all responses with a recorded date before the import. Just be sure to use start or end date rather than recorded date if you need it for reporting purposes
JoycaV Hi, just read your reply about getting duplicate replies. I think the solution is you have to export your data as you mentioned but delete all the responses out of the survey before re importing them back in. This is very nerve racking when doing this and also it will create new Recorded Date values. There is no way round this other than using API to manipulate the responses in situ. But the problem then with that is that if you have manually edited some entries already, API won't work with those specific responses until you've a) found them in the data & analysis section [which is like finding a needle in a haystack!] and then b) manually clicked on each entry to return it to the original figure [again there is no filter or function to show you which entries have been edited]. In the past the only way I was able to find all the responses was by exporting the data normally and in legacy format to identify the differences. Nightmare!
So whilst I welcome this new Combine fields feature, it's only half a solution as most entries use the text set field type I guess.
Sorry I don't have an answer for you on this but at least you are not alone!! ;)
Anyone with any other thoughts on this subject, please do respond - perhaps there is a simpler way that is not well known about just yet?
Thanks
Rod
I realize this is an older thread, but an easier solution might be to create an embedded data field (as described, or just blank if the survey is completed) and use a combination of filtering a response editing.
Response Editing | You would have to just do it once per multiple choice. Filter the data to show only where QID5 = 'yes' OR QID10 = 'yes', and edit your new embedded data field to equal 'yes'. Do the same for 'no' and/or whatever other response options you have.
@bmcgrath, Hiya, you are correct, providing the options and combinations are low then this could be done. It's frustrating that we can't do a blanket change but maybe one day this option will become available.
Thanks
Rod
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.