Welcome, Guest
Username: Password: Remember me
  • Page:
  • 1
  • 2

TOPIC: Attaching follow-up to random questions

Attaching follow-up to random questions 3 years 4 weeks ago #64813

  • barnaby1
  • barnaby1's Avatar
  • OFFLINE
  • Fresh Lemon
  • Posts: 12
  • Karma: 0
I have a question group of 12 Qs presented randomly. I now want to include a 'how sure are you' type follow-up question for each, but how can I make sure each follow-up is attached to the correct random main question? Obviously if I make the follow-up questions random, they all get mixed.

Any suggestions welcome.
The administrator has disabled public write access.

Re: Attaching follow-up to random questions 3 years 4 weeks ago #64815

  • tpartner
  • tpartner's Avatar
  • NOW ONLINE
  • LimeSurvey Team
  • Posts: 4185
  • Thank you received: 785
  • Karma: 359
You should be able to use conditions to only show a followup question if the corresponding random question is answered.
Cheers,
Tony Partner

Solutions, code and workarounds presented in these forums are given without any warranty, implied or otherwise.

LimeSurvey is open-source and run entirely by volunteers so please consider donating to support the project.
The administrator has disabled public write access.

Re: Attaching follow-up to random questions 3 years 4 weeks ago #64816

  • barnaby1
  • barnaby1's Avatar
  • OFFLINE
  • Fresh Lemon
  • Posts: 12
  • Karma: 0
Thanks Tony

I thought that, but the only problem is that the main question only has blocks of colour as answers (recall www.limesurvey.org/en/forum/can-i-do-thi...answer-options#64177) This means it's hard to attach <a href='docs.limesurvey.org/tiki-index.php?page=Setting%20conditions'>conditions</a>. Or is it possible to simply add the condition 'show if question x has been answered'?
Last Edit: 3 years 4 weeks ago by barnaby1. Reason: typo
The administrator has disabled public write access.

Re: Attaching follow-up to random questions 3 years 4 weeks ago #64817

  • barnaby1
  • barnaby1's Avatar
  • OFFLINE
  • Fresh Lemon
  • Posts: 12
  • Karma: 0
I've sorted the conditions thing. But it doesn't work. The follow-up question does appear when the appropriate random Q is answered, but it appears at the top, under main Q1 (when the relevant main Q might be any number).
The administrator has disabled public write access.

Re: Attaching follow-up to random questions 3 years 4 weeks ago #64822

  • atiut
  • atiut's Avatar
  • OFFLINE
  • Senior Lime
  • Posts: 76
  • Thank you received: 2
  • Karma: 1
Another solution is to put the *same* question to all but and then analyze the data considering what question was asked first. There are two ways imaginable in which you can do it:

1. If the random question is compulsory you can just look and see previous answers given by the respondent in the data analysis stage. Some re-coding in Spss or whatever program you use will do the trick.

2. Make questions pseudo-random. Make an attribute token field that takes values from one to thirteen, show "random question" depending on the number in the token field and then in the database you will see nice and clear what question was asked.
The administrator has disabled public write access.

Re: Attaching follow-up to random questions 3 years 3 weeks ago #64824

  • barnaby1
  • barnaby1's Avatar
  • OFFLINE
  • Fresh Lemon
  • Posts: 12
  • Karma: 0
Sorry, I don't quite understand either, although I think both sound plausible (if that makes any sense!)

Can you explain a bit more? Thanks
The administrator has disabled public write access.

Re: Attaching follow-up to random questions 3 years 3 weeks ago #64825

  • barnaby1
  • barnaby1's Avatar
  • OFFLINE
  • Fresh Lemon
  • Posts: 12
  • Karma: 0
But how do I know the order in which the random questions were asked? Aren't the results presented in the actual order of the questions (not the random presentation)?

Really need to get this sorted today if poss.

BTW, this is an open survey.
The administrator has disabled public write access.

Re: Attaching follow-up to random questions 3 years 3 weeks ago #64826

  • atiut
  • atiut's Avatar
  • OFFLINE
  • Senior Lime
  • Posts: 76
  • Thank you received: 2
  • Karma: 1
OK, so let us stick to option 1. Note that I assumed you will process the data offline and get the reports, rather than using the automated report tool.

After you export the data the table will look like below (this is the VV export, other exports will look slightly different.)

Q1_random Q2_random … Q13_random Follow-up quest
34463X136X65272 34463X136X65273 34463X136X6528 34463X136X6529
y {question_not_shown} {question_not_shown} 5
{question_not_shown} y {question_not_shown} 3
{question_not_shown} {question_not_shown} n 4
{question_not_shown} {question_not_shown} {question_not_shown} 5

You can clearly see, for example, that the first respondent chose Y at Q1_random, did not see Q2 thru Q13 and chose option 5 in the follow up question.

You can run frequencies and crosstabs on this structure of data and do all sorts of statistical processing.

Let me know if this helps and makes sense. If the table does not display properly use this image: prntscr.com/2plob
The administrator has disabled public write access.

Re: Attaching follow-up to random questions 3 years 3 weeks ago #64827

  • barnaby1
  • barnaby1's Avatar
  • OFFLINE
  • Fresh Lemon
  • Posts: 12
  • Karma: 0
It's making more sense, thanks, but still not sure it does what I need. I think the key is being able to tell the order in which the random questions were actually presented.

Sorry if I'm being dense here :)

I need all users to answer all 12 random Qs, I just need to know how sure they are about each answer. I tried recreating what you suggested, just couldn't see the answers I needed.
The administrator has disabled public write access.

Re: Attaching follow-up to random questions 3 years 3 weeks ago #64828

  • atiut
  • atiut's Avatar
  • OFFLINE
  • Senior Lime
  • Posts: 76
  • Thank you received: 2
  • Karma: 1
My bad, I was under the impression you would only show one question. If you show all questions I believe my solution would not work.

In theory, you should create 13 groups of 2 questions and randomize the groups. Sadly this is not supported by LS. I have seen several discussions on the forums, but I am unsure whether a workaround exist.

Do not despair though... The next best thing is to create two questions within a single one. If both your question are scales, then Array Dual Scale is what you need. Make sure you set your scales to be drop-down boxes in advanced settings.

PS:conditions should have also worked. I do not understand why they did not, maybe you can share a link
The administrator has disabled public write access.
  • Page:
  • 1
  • 2
Moderators: ITEd
Time to create page: 0.116 seconds
Donation Image