Pseudo randomization

More
3 years 5 months ago #119223 by slowking
slowking created the topic: Pseudo randomization
Hi!
Let's suppose that I've created survey with two randomly assigned versions (let's say version 1 and version 2) (via rand(x, y) in hidden question and relevance equation). Of course it is possible that all taken surveys will be version 2. Is there a way to store somewhere somehow number of taken surveys and manipulate randomness? For example something like this

IF version1number - version2number > 10 THEN take survey 2
IF version2number - version1number > 10 THEN take survey 1
ELSE take random survey

(Sorry for poor english :) )

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

More
3 years 5 months ago #119229 by Ben_V
Ben_V replied the topic: Pseudo randomization
HI,
Maybe it's ok for you this way .
The only important think is you need a welcome page or a dummy 1st group. (won't work on the 1 first page)

Benoît

EM Variables => bit.ly/1TKQyNu | EM Roadmap => bit.ly/1UTrOB4
Last Releases => 2.6x.x goo.gl/ztWfIV | 2.06/2.6.x => bit.ly/1Qv44A1
Demo Surveys => goo.gl/HuR6Xe (already included in /docs/demosurveys)

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

More
3 years 5 months ago #119238 by tpartner
tpartner replied the topic: Pseudo randomization
Yeah, as Ben suggests, using the SAVEDID will work. But note that all responses get a new ID, regardless of whether the survey is completed so you still may not get a perfect distribution for completed surveys.

Cheers,
Tony Partner
Solutions, code and workarounds presented in these forums are given without any warranty, implied or otherwise.

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!