Assignment based on previous random assignment/Random Startpoint | XM Community
Skip to main content

Dear Community,
I have a question and hope someone might be able to help. I am conducting experimental research and am unsure how to realize something in qualtrics:
I have scenarios that I want participants to read that either have a male or a female protagonist (Afemale Amale; Bfemale Bmale; Cfemale Cmale (...) Lmale Lfemale).
I want two things to happen:

  1. Participants should be randomly assigned to the first scenario, i.e. some should read C first, some B, and some A.

  2. If they saw Amale first, I want them to next see Bfemale, and then Cmale; so that the gender always alternates.


I thought about it for a while and I am unsure how to do this other than to have a crazy long complicated survey flow. Does someone have an idea?

Edit: What would already help is if someone could tell me how it could be done, that participants get a random startpoint, i.e. if I have 12 scenarios, how can I achieve that some start with scenario 1 and then co to 12, and others start at scenario 5 and then see 6,7,8,9,10 etc.?

Thanks a lot!

DG1989 I think the attached QSF might work for you. (I did an example of 4 scenarios but you can expand it to 12).
In the Survey Flow the respondent get randomly assigned to a voice (male or female). This will be the starting voice for Scenario A.
image.pngThen a random number generator is used to pick an integer from 1-4 to assign the start point where A=1, B=2, C=3, D=4. You can edit the 4 to be a 12 and it will serve your needs.
image.pngThen there are are embedded data fields for the text in male and female voice you want to appear. I recommend you put the whole block of text where the voice appears and not try to use piped text within the editor. Just pipe these in as a big chunk. These need to be filled in by voice and again you will need to expand it up to 12.
image.pngNext there is a block only used for testing, so you can see where you were to start and in which voice Scenario A should be.
image.pngEach Scenario is added as it's own block to make the last part of the flow work. You will need 12 branches that define the start point and then cycle through the scenarios starting at the start point (see flow to make this easier to understand).
image.pngRandom_Gender.qsf



Hi bstrahin,
thank you so much for your detailed answer and especially for the time to make all these screenshots and the qsf file !!!

One more short question: Do you have any idea if it also would be possible to have A,B,C,D in a random order that is different for everyone, while still keeping the gender alternated?

Thank you so much!


Maybe to better illustrate what I mean: My first thought of making the flow was the one below:
image.pngHowever, the problem there would be that some people then, based on chance, may only see female options while some may only see male options and that is not at all what I want.


DG1989 I can't think of an easy and efficient flow that would keep the gender alternating when the scenarios are in a random order. You could have a simpler flow if you are open to having a random assignment of gender for and random order for scenarios. You could get the genders to balance 6 male and 6 female easily, but I am stuck on how to make sure you alternate genders between each scenario.


image.pngimage.pngHi bistrahin,
I was just playing around a bit more and now thought that another solution might be to have a flow like the one above. If I am not having a brain freeze then this would ensure that everyone is seeing 6 male scenarios and 6 female scenarios, right? The only downside here might just be that there is a possibility that someone will see first 6 female and then 6 male scenarios but that would then also only happen for very few participants, right?


DG1989 its possible they would get any permutation of 6 males and 6 females. So for example all kinds of chunking is possible, such as:

  • MMMMMMFFFFFF

  • MMMMMFMFFFFF

  • MMMMFMMFFFFF

If you are okay with this, then I suggest a more simple flow for you would be:
image.pngimage.pngimage.pngWithout testing your flow I worry about possibly selecting a group more than once. I just haven't set it up that way and it looks more complicated than it needs to be.


@bstrahin thanks a lot! I see how putting the text in the embedded data makes the flow way cleaner, however, to every male/female version there is also a corresponding picture and corresponding questions which is why I thought it might be easier to order it in the different blocks.

Again, thanks a lot for your help! I will leave the question open for now in case someone comes along in the next couple of days who has some out-of-the-box idea how to avoid the gender-chunking




DG1989 that makes sense to use the different blocks if you have different follow up questions. Images can also be stored in the embedded data field using HTML code, but again that's an annoying complication. So your approach of different blocks makes a lot of sense. In that case, You would gut out the branching that I have and you can make 2 branches with a randomize like above in the branch. The blocks you show would be the ones you want to be eligible. It should look just as clean...
image.pngimage.pngimage.png(please ignore the verification block that snuck into the second picture where Voice=male. At this point I think you get it without updating the picture.)


Leave a Reply