Randomly assign participants to differnet question groups

Plus d'informations
il y a 4 mois 2 semaines #179936 par Ellaisa
Hello everyone,

I tried to randomly assign participants to differnet question groups.
I have created 2 different szenarios within 2 different questiongroups. Some people should see szenario 1 and the other szenario 2.
Therefore I created a question group (that is always hidden) including in one question the randomization with: if(is_empty(randnumber.NAOK), rand(1, 2), randnumber.NAOK) - question type is equation

Now I have created another question group, that is my first szenario. In the relevance euqation I have added: randnumber == "1"

Afterwards my questions for this szenario follow.

I have done the same with szenario number 2, only changed: randnumber == "2"


I tought that this should lead to a 50/50 randomization. Now I have started a pretest and I mostly offers szenario 1 (85% of participants).

Can anyone tell me what I have done wrong or could do better?

Thanks a lot!

Connexion ou Créer un compte pour participer à la conversation.

LimeSurvey Partners
Plus d'informations
il y a 4 mois 2 semaines #179937 par tpartner
What is your sample size? The larger the sample size, the closer you will get to 50/50.

You might also try:

if(is_empty(randnumber.NAOK), rand(1, 100), randnumber.NAOK)

And then use relevance:

intval(randnumber) < 51
intval(randnumber) > 50

Cheers,
Tony Partner

Solutions, code and workarounds presented in these forums are given without any warranty, implied or otherwise.
Official LimeSurvey Partner - partnersurveys.com

Connexion ou Créer un compte pour participer à la conversation.

Plus d'informations
il y a 4 mois 2 semaines #179938 par Ellaisa
Thanks!
Until now I have 40 completed questionnaires.
I was just wondering, as the first 10 were only scenario 1, afterwards 2x scenario 2 appeared and then it went on with scenario one.

I will try your solution!
Can you see any mistake in the way I did it? Or should it work out with a bigger sample size?
I need at least 100 participants

Connexion ou Créer un compte pour participer à la conversation.

Plus d'informations
il y a 4 mois 2 semaines #179940 par Joffm
Hi, Ellaisa

everything you did was fine.
But as always:
If you throw a coin 10 times, you won't expect it to be a 5/5 distribution.
If you throw it 100 times, you won't expect a 50/50 distributiion, but will be nearer to a 50%/50%.

I just tested:
40 times with randnumber=rand(1,2): 19 times 1, 21 times 2
40 times with randnumber=rand(1,100): 18 times 1-50, 22 times 51-100


To be able to finetune a distribution survey, I prefer the following:
1. create your random number : r1=rand(1,100)
2. The relevance of your 2 scenarios is: r1<51, resp. r1>50.

If you see that the distribution gets unbalanced, just change the relevance equation (e.g. too much "2") like
scenario 1: r1<31
scenario 2: r1>30

or even (you only need "2")
scenario 1: r1==0
scenario 2: r1>0

You can do this, even if the survey is activated.

Best regards
Joffm


Volunteers are not paid.
Not because they are worthless, but because they are priceless

Connexion ou Créer un compte pour participer à la conversation.

Plus d'informations
il y a 2 semaines 1 jour #185207 par nannancy
Hey, you can try

{SAVEDID - floor(SAVEDID / 2) * 2 + 1}

in the equation field of the hidden question.

Best
Xinyuan

Connexion ou Créer un compte pour participer à la conversation.

Plus d'informations
il y a 2 semaines 1 jour #185210 par Joffm
Of course you can use the SAVEDID.

But it is not sufficient, because the SAVEDID is incremented each time a survey is started.
So "not completed" surveys have a SAVEDID as well.
So, if by chance each second respondent does not complete the survey only odd SAVEDIDs are completed and you have all respondents in one group.

Therefore I'd prefer the random number, but checking the distribution from time to time.

Joffm


Volunteers are not paid.
Not because they are worthless, but because they are priceless

Connexion ou Créer un compte pour participer à la conversation.

Plus d'informations
il y a 2 semaines 1 jour #185216 par DenisChenu
A good idea of puglin :woohoo:

Assistance on LimeSurvey forum and LimeSurvey core development are on my free time.
I'm not a LimeSurvey GmbH member, professional service on demand .
An error happen ? Before make a new topic : remind the Debug mode .

Connexion ou Créer un compte pour participer à la conversation.

Commencez dès maintenant !

Créez simplement un compte et commencez à utiliser LimeSurvey dès aujourd'hui.

Inscrivez-vous maintenant

Inscrivez-vous à notre Newsletter!