Welcome to the LimeSurvey Community Forum

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

Set Rule for Group Randomization

  • limette123
  • limette123's Avatar Topic Author
  • Offline
  • New Member
  • New Member
More
8 years 5 months ago #127264 by limette123
Set Rule for Group Randomization was created by limette123
Hi,
i`ve created five groups and set "Group Randomization” for these five groups.
With these randomised loops, it is now important for my analysis of data that each group needs to be evaluated on first position with the same frequency (n=60).

Thus, with a total sample size of n=300,
Group 1 shall be shown 60 times on first position
Group 2 shall be shown 60 times on first position
Group 3 shall be shown 60 times on first position
Group 4 shall be shown 60 times on first position
Group 5 shall be shown 60 times on first position

Can someone help me with this?
Many thanks in advance!
The topic has been locked.
  • limette123
  • limette123's Avatar Topic Author
  • Offline
  • New Member
  • New Member
More
8 years 5 months ago #127352 by limette123
Replied by limette123 on topic Set Rule for Group Randomization
has noone got an idea on how to solve this problem? :(
The topic has been locked.
  • holch
  • holch's Avatar
  • Offline
  • LimeSurvey Community Team
  • LimeSurvey Community Team
More
8 years 5 months ago #127359 by holch
Replied by holch on topic Set Rule for Group Randomization
Randomization does not guarantee quotas, as it is a real randomization (100% chance).

But depending on your setup, there might be a workaround? How do people access your survey? Is it open to anyone? Or do you use tokens? With tokens, you might be able to get to your result.

I answer at the LimeSurvey forum in my spare time, I'm not a LimeSurvey GmbH employee.
No support via private message.

The topic has been locked.
  • limette123
  • limette123's Avatar Topic Author
  • Offline
  • New Member
  • New Member
More
8 years 5 months ago #127387 by limette123
Replied by limette123 on topic Set Rule for Group Randomization
i actually did not plan to use any tokens since i want to adress everyone with my survey (only condition is the right nationality).

But how is there a way to manage this problem with the use of tokens?

But without using any tokens, would I at least see in the end when obtaining the data how often Group 1 was shown on first position, Group 2 on first position, Group 3 on first position, etc.? This is extremely important for my data analysis since I can only interpret and use the first question of each question group IF this whole question group was shown BEFORE the other four question groups.
Thanks a lot!
The topic has been locked.
  • limette123
  • limette123's Avatar Topic Author
  • Offline
  • New Member
  • New Member
More
8 years 5 months ago #127433 by limette123
Replied by limette123 on topic Set Rule for Group Randomization
Is there maybe a workaround to my problem?

I`ve got 5 question groups all structured in the same way (7 questions per question groups). These question groups are shown in randomised order.
Is there the option to control that Q1 of each question group is ONLY shown if the corresponding question group is shown on FIRST position?
The topic has been locked.
More
8 years 5 months ago #127437 by first
Replied by first on topic Set Rule for Group Randomization
Yes I guess this can be done.

Requirement: The question should not be a text display question and it should be a mandatory question.

Logic:
Place below relevance condition in first question of each group.

(First Question Of Group One Is Not answered) && (First Question Of Group two Is Not answered) ... && (First Question Of Group Five Is Not answered) ..

This way the question will appear only in first group.

Survey Designer and Programmer
The topic has been locked.
More
8 years 5 months ago #127443 by Ben_V
Replied by Ben_V on topic Set Rule for Group Randomization
You may have a look at this previous post

In your case a solution -even if not probably the best- can be:

1) Set a general quota n=300 for your survey
2) Create your 5 groups (G1, G2, G3, G4, G5) and a related duplicate set (G1b, G2b, G3b, G4b, G5b)
3) Use the following relevance equations for each group:

G1: ((substr(SAVEDID, -1) == "1") or (substr(SAVEDID, -1) == "2"))
G2: ((substr(SAVEDID, -1) == "3") or (substr(SAVEDID, -1) == "4"))
G3: ((substr(SAVEDID, -1) == "5") or (substr(SAVEDID, -1) == "6"))
G4: ((substr(SAVEDID, -1) == "7") or (substr(SAVEDID, -1) == "8"))
G5: ((substr(SAVEDID, -1) == "9") or (substr(SAVEDID, -1) == "0"))

G1b: ((substr(SAVEDID, -1) != "1") or (substr(SAVEDID, -1) != "2"))
G2b: ((substr(SAVEDID, -1) != "3") or (substr(SAVEDID, -1) != "4"))
G3b: ((substr(SAVEDID, -1) != "5") or (substr(SAVEDID, -1) != "6"))
G4b: ((substr(SAVEDID, -1) != "7") or (substr(SAVEDID, -1) != "8"))
G5b: ((substr(SAVEDID, -1) != "9") or (substr(SAVEDID, -1) != "0"))

4) Merge results for data analysis (G1+G1b … G5+G5b)

There is maybe a similar way using javascript, just moving (client side) G2, G3, G4 on G5 if necessary with no need to duplicate the groups… but I don't know if it's possible to retrieve {SAVEDID} value with js…

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)
The topic has been locked.

Lime-years ahead

Online-surveys for every purse and purpose