Hi there, I try to use Group Randomization Test Survey ：

This survey will test the group randomization feature, it contains 3 groups with 3 questions in a group. The first question in each group is in rand group1, the second question in each group is in rand group2, and the last question in each group is in rand group3. The survey groups are Group A,B,C.

To Illustrate this:

---

Rand Group1 Q1 | Q4 | Q7

---

Rand Group2 Q2 | Q5 | Q8

---

Rand Group3 Q3 | Q6 | Q9

surevy link：
www.pushknock.com/projects/limesurvey/index.php/88881?lang=en

version 2.5


this survey type will show all the 3 question in this survey,
But I just want part of respondent only answer the Group1, and the other will only answer the Group2 and the others will answer
like the treatment group and the control group. (Experimental Design）

is anyone can help me to solve my problem.
many thanks!

Rita

Hi, Rita,

first: Is your usage of the word "group" the same as used in LS? Then do it by group relevance?

How do you select the respondensts who only answer group 1?
Are the predefined, Is it at random?

In case of the first I would recommend Token Attributes, in case of the second a random number.

In both cases you enter in "Edit groups" in group relevance the equation, e.g. "TOKEN:ATTRIBUTE_1==1" or "myrandomnumber==2"

If your usage of "group" is different, display the survey "question by question" and use question relevance to determine which questions are diplayed to the respondent.

Best regards
Joffm

I follow here www.limesurvey.org/manual/Expression_Man...Randomization_Groups
and import the sample survey ：Randomization_Group_Test.lss

How do I select the respondensts who only answer group 1?
It's at random.

and I want to use limesurvey to do the Experimental Design.
and different people will see 1 out of 2 group.
so is that could be implement in Limesurvey?


many thanks

I think you're confusing the purpose of that downloaded survey. It simply displays questions in a random ORDER, not a random visibility.

If you want half of your participants to see one set of questions or groups and the other half to see another set of questions or groups...

1) Add an equation type question (let's give it a code of "equation1) at the beginning of the survey.

2) Load that question with a random number between 1 an 2 with an equation like this:
3) Place relevance on following questions or groups such as: or: