Using Decks to create Treatment and Control Group

Sampling & Quotas
1

Hi,
I want my survey to split the respondents into two equal parts after the first 2-3 pages.
Group 1 and Group 2 should receive different experiments. I think this is done with either decks or segments but couldn’t figure it out by myself how to actually do it.

Thanks in advance,
Arne

2

Hi Arne,

you should indeed use a deck to achieve this. Create a deck of size 2 and call it e.g. “group”. For each respondent the value of the variable “$group” will be either 1 or 2 in a uniform distribution.

You can then add a display condition “$group == 1” e.g. to a question and the question will only be shown to those participants who got value 1 from the deck.

If you want to hide an entire section from a group of participants you can use a branch element with such a display condition. Those participants that fulfill the display condition of the branch will directly jump to the page selected in the branch element.

1 Like
3

Hello,
I have to conduct in my survey two DCEs concerning two different products A and B. So I will divide my sample into two and randomly ask respondents to answer one the DCEs (either A or B).
Still not sure how to do this though…
I create an experiment with all of the DCE questions (for product A and B), then at the end of the page add a deck of size 2… what will be the display condition in that case ?
Thank you in advance :slight_smile:

4

Fred here again.
Depending on what your research question is there are many ways to do this.

If your experiments are fundamentally and structurally different - and you want different strata to be directed to either one or the other, the least painful is simply to user two different DCE’s then use conditional expressions to display one or the other.

So in your case.

  1. create a deck of size 2 call it ‘product’ - this will ensure a random without replacement even allocation to either ‘1’ or ‘2’ - and will control for dropouts. This should be BEFORE the experiment section.

  2. then create 2 DCEs - one for each product. productA and productB

  3. then simply put 2 scenarios on each page for each DCE using the DCE selector.

  4. put the conditional display on the scenarios
    on all the productA scenarios put the condition ($product ==1)
    on all the productB scenarios put the condition ($product ==B)

Now you will have even stratification and two independent data sets.
Sweet!