Seeking Assistance on Randomized Treatment Assignment in Survey Design | XM Community
Question

Seeking Assistance on Randomized Treatment Assignment in Survey Design

  • 15 April 2024
  • 5 replies
  • 19 views
J
Badge +1

Hello Everyone, 

I am writing here because I am encountering some issues when creating my survey for my thesis. I am studying whether matching a chatbot personality to the user can yield to a greater acceptance. 

To do so, after asking their personality type, I would like to assign each participant randomly to different treatments. However, with the way I’m doing things, the participant goes through all these blocks one after the other. Here below is a screenshot of the Survey Flow. 

Is there a way to go in which the participant selects an answer from a questions and is allocated randomly to one of these following blocks once?

Your assistance upon this matter is highly appreciated.

Best regards, 

Jorge Ibanez

 

5 replies

A
Rank
Userlevel 7
Badge +21

Change this to display only 1:

 

J
Badge +1

Hello @ahmedA,

 

Thanks for pointing that out!


I changed that in my survey flow, and weirdly enough, in my previews it works with only one block appearing, but with the published survey link, I still get those 4 blocks one after the other. 

Do you know any idea why and how we can resolve that?

Looking forward to your reply. 

Best, 

 

A
Rank
Userlevel 7
Badge +21

You must publish the survey again after making any changes for them to reflect in the survey link.

This is so that you can test/preview changes before making them live.

 

Also, if you are using the same browswer, clear the cache and cookies before trying the new published link.

J
Badge +1

Hello Ahmed, 

Thank you for that advice! 

I published the survey again and it worked out how I wanted it to be. 

The thing is now, after collecting the data, the specific routing undertook by the participant is not reflected in the data-analysis. Do you know what I can do in the workflow or in the builder so that I can trace the specific treatment that has been assigned to the participant in the data analysis?

Thank you in advance for your help. I truly appreciate it!

Best regards, 

 

Jorge 
 

 

A
Rank
Userlevel 7
Badge +21

One option would be put each of your blocks in groups. Each group will contain an ED block, for example Treatment = X and the relevant treatment block. And put these 4 groups in the randomizer.

 

The other, post data collection, method is to look at your responses. If the question was shown to them and they responded, you’ll have data, otherwise it will be empty. If you download your data as CSV, you’ll get a question export tag, which you can use to identify which question they have answered.

Leave a Reply


Terms & Conditions