Community Groups are officially here!
We've released Groups, a new feature that enables us to connect community members of similar industries and interests in a shared, private space. You can check out all of the details here, including information about who can join, how to join, and what Groups are currently offered. Please leave your feedback through this Community Groups Feedback Survey.

Grouped randomization in a constant sum question?

Nickolas
Nickolas VancouverCommunity Member Qubie ✭

I want to randomize the choices within a constant sum question but keep 2 particular choices grouped together while still being in a random location as a group relative to all other choices. This seems to be possible using groups in a standard multiple choice question but groups do not exist for constant sum questions. Is there something I am missing within qualtrics or a bit of javascript that would perform this function?

Best Answer

  • TomG
    TomG Raleigh, NCCommunity Member, Qualtrics Brand Admin Wizard ✭✭✭✭✭
    Accepted Answer

    @Nickolas,

    I haven't tried this, but I think it will work. Create a MC question, put the subset in a group and randomize the question. Don't force response and hide it with JS. Then carry forward the displayed choices into your constant sum.

    If that doesn't work, using Kate's suggestion and putting the random subset in the middle of the list may decrease bias.

Answers

  • Kate
    Kate Community Member, Qualtrics Brand Admin Wizard ✭✭✭✭✭

    Kind of a weird work-around, but simple keep the two you want to stay together in the "random Subset" area together. So long as the random subset ones are kept together on the list to the left, they will always appear together in the randomized list. In this example, "C" and "D" always display together at the bottom of the randomized list.

  • Nickolas
    Nickolas VancouverCommunity Member Qubie ✭

    Hey Kate! Thanks for the answer reply. I had looked at this an option but am worried about there being an allocation bias of giving "whatever is leftover" to the bottom-most choice on the list. Ideally, I would be able to randomize the random subset's location relative to the rest of the group as well.

  • Kate
    Kate Community Member, Qualtrics Brand Admin Wizard ✭✭✭✭✭

    Great idea @TomG. I have definitely randomly carried forward the randomized lists before, that certianly feels like it will work if it's grouping correcty on MC.

  • Nickolas
    Nickolas VancouverCommunity Member Qubie ✭

    That worked perfectly @TomG! Thanks for the help @Kate and @TomG!