Random question/group(uniform probability distribution)

More
5 months 3 weeks ago - 5 months 3 weeks ago #166231 by RitaShen
RitaShen created the topic: Random question/group(uniform probability distribution)
hey there, I have a question about the Random Design.

If I sent the 25,000 mail invitation, but I only want the 1,000 respondents to answer my question.
and I have 4 questions need to randomize (1 of 4 questions randomly appear)
how could I let the 4 question uniform probability distribution.

or if Question 2 is already 250 respondents, and the other respondents will automatically change to another questions

question frequency
A250
B250
C250
D250

every question could randomly and uniform probability distribution

because now it's seems cannot uniform probability distribution if use any equation.

Rita
Last Edit: 5 months 3 weeks ago by RitaShen.

Please Log in or Create an account to join the conversation.

More
5 months 3 weeks ago #166233 by Joffm
Joffm replied the topic: Random question/group(uniform probability distribution)
Hi, Rita,
unfortunately LS doesn't give access to the number of filled quotas, like other - commercial - tools.
In LS you don't have something like:
quota_male_Target (how many males are needed)
quota_male_ToDo (how many males are missing)
quota_male_Done (how many males answered already)

So the best is to use a random number (1-4)
Did you check the distribution of LimeSurvey's random number (1-4)
If you create a survey with only one question - the random number - and run it, let's say 200 time, how many 1, 2, 3, 4 do you get?

And if you see that one group is filled, like
A: 241
B: 252
C: 246
D: 240

change the relevance equation that Group B is never displayed, but Group D is displayed if random number in [2,4].


Personally I prefer to use a random number (1-100).
First:
A (1-25)
B (26-50)
...

Later I can fine tune
A (1-28)
B (29-50)


Joffm

P.S.
As Jelo explained (in the German section):
If you send out 10000 invitations you very probably will get fake interviews, speeder, happy clicker and so on.
So in reality you will need some more interviews than 250 to get 250 real usable interviews after cleaning.


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

Please Log in or Create an account to join the conversation.

More
5 months 3 weeks ago #166428 by Joffm
Joffm replied the topic: Random question/group(uniform probability distribution)
Hi, some additional information:

https://www.limesurvey.org/forum/design-issues/110017-balance-quotas-fill-least-full-quota

Joffm


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

Please Log in or Create an account to join the conversation.

More
5 months 3 weeks ago #166439 by holch
holch replied the topic: Random question/group(uniform probability distribution)
And I would not send out 25.000 at once. Go in steps and see how it goes. Go for 1000, wait for a response rate, than go for another 1000, etc. Depending on the quality of your list, the topic of your research, relevance to the contacts on the list, etc. 25.000 might actually be not enough to even reach n=1000. But this is a different story.

I'm not a LimeSurvey GmbH member. I answer at the LimeSurvey forum in my spare time. No support via private message.
Some helpful links: Manual (EN) | Question Types | Workarounds

Please Log in or Create an account to join the conversation.

Start now!

Just create your account and start using Limesurvey today.

Register now
Join our Newsletter!