Best answer by PeeyushBansal
View original
How do I set up (start) a survey using a randomized incomplete block design?
None of my searches offered answers to this question.
Userlevel 7
+33
If you are looking to randomize blocks in survey, than in survey flow you can add Randomizer element and place all blocks under this that you want to randomize.
Dear bansalpeeyush29,
Thank you very much for your reply. I do want to randomize concepts within a block, but not randomize blocks within a single survey. I have attached a fuller description of the research project. I look forward to your ideas and recommendations.
Regards,
Bill
Thank you very much for your reply. I do want to randomize concepts within a block, but not randomize blocks within a single survey. I have attached a fuller description of the research project. I look forward to your ideas and recommendations.
Regards,
Bill
Userlevel 7
+33
Yes you can use loop and merge here and randomly select 5 of your concepts out of all.
Userlevel 7
+33
Read below for more detaild
https://www.qualtrics.com/support/survey-platform/survey-module/block-options/loop-and-merge/
https://www.qualtrics.com/support/survey-platform/survey-module/block-options/loop-and-merge/
So loop will create X number of unique blocks/surveys of five randomly selected concepts from the 15 and randomization will randomly order the concepts within a block/survey so that it is as if all respondents evaluated all concepts with a n=30 for a single concept - right?
Thanks for these pages - they were very helpful. If I am understanding these correctly in Qualtrics "blocks" are sets of questions that can be used for "looping and merging" in a survey. My reference to "blocks" are a subset of concepts (or any treatment stimuli) selected based on an experimental design. My survey questions are the same for all respondents and all concepts in a fixed order. So if Qualtrics does not select stimuli based on an experimental design to structure a set of surveys, the stimulus sets would have to be determined ahead and “piped in” from the design matrix for which the “block” of set questions would be “looped and merged.” As long as the design-determined number of individual surveys were run based on the “concept blocks,” would this work?
Leave a Reply
Enter your username or e-mail address. We'll send you an e-mail with instructions to reset your password.