Welcome to the LimeSurvey Community Forum

Ask the community, share ideas, and connect with other LimeSurvey users!

Two step randomization

  • knightnight21
  • knightnight21's Avatar Topic Author
  • Offline
  • New Member
  • New Member
More
3 years 10 months ago - 3 years 10 months ago #199922 by knightnight21
Two step randomization was created by knightnight21
I am trying to make an experiment where there are 2 main parts. I currently have the format as three questions group QG1/QG2/QG3

The first part is a survey that everyone takes QG1.

The second part is the experiment where I need to randomize twice separately. First I want to do an A/B randomize by a group of QG2 or QG3, I have tried following the online tutorials of making a hidden question in QG1 with the equation {rand(1,2)} with the code random. However, when I put random ==1 in the relevance section of QG2 it says undefined variable. I have tried this to A/B randomize within a question group and have been successful but it is not what I am intending to do.

Next is the part that I am unable to find any answers for, within QG3 I have six questions that are separated into two groups or bundles with three questions each (B1 and B2). I want to randomize the order in which they appear so B1 presents itself first half the time followed by B2 and then vice-versa. Since I want both to appear it is different than A/B and because there are questions that need to be followed up after each other I can't randomize all the questions.

I have attached my lss report.
Last edit: 3 years 10 months ago by knightnight21.
The topic has been locked.
  • Joffm
  • Joffm's Avatar
  • Away
  • LimeSurvey Community Team
  • LimeSurvey Community Team
More
3 years 10 months ago - 3 years 10 months ago #199926 by Joffm
Replied by Joffm on topic Two step randomization
Your first question is answered quickly.
There are two small mistakes:
a. You have a typo in the second group "rando==2" instead of "random==2"
b. You entered it into "ramdomization group name" instead of "relevance equation"

And the second question?
Two options:
a. Split the group into two:
Both have the same relevance equation "random==2"
And give both the same name in "randomization group name" like "R1"
Endo1
Endo2

b. If you really want to display all six questions on one page:
copy the first three to the end
create a second random number (1,2)
EndoYa1 : random2==1
EndoYb1 : random2==1
EndoYc1 : random2==1
EndoNa
EndoNb
EndoNc
EndoYa2 : random2==2
EndoYb2 : random2==2
EndoYc2 : random2==2

Of course you can't use the same random number - as you do in your example - that selects EXO or ENDO.
If EXO is selected, you do not see anyhing of ENDO.

Seeing your survey I'd prefer option a. It's a clearer separation of the two cases.
And in my opinion the second group is far too long and discouraging.
Here you should consider to display each question separately.


And which version of LS are you using?
Seems to be an older one, isn't it?

Joffm

Volunteers are not paid.
Not because they are worthless, but because they are priceless
Last edit: 3 years 10 months ago by Joffm.
The topic has been locked.
  • knightnight21
  • knightnight21's Avatar Topic Author
  • Offline
  • New Member
  • New Member
More
3 years 9 months ago #200252 by knightnight21
Replied by knightnight21 on topic Two step randomization
Thank you!

I am using Version 2.58.0+170104. It is what is provided by my university.

So I have made the changes that you have suggested. However, when I preview the survey I go through my first two sections that everyone needs to fill out, and then the survey is completed without going into the experiment of Exogenous vs Endogenous.

I have attached my new lss

thank you
The topic has been locked.

Lime-years ahead

Online-surveys for every purse and purpose