How assign participants to different role using JATOS when conducting group study?

Comments

• kri

  August 2021

  Hi Bo,

  there aren't any other group study examples as the ones in https://www.jatos.org/Example-Studies.html. Writing and conducting group studies can be a bit tricky due to participants acting in parallel (but nothing impossible). Can you describe more in detail what you are trying to achieve with 'assign participants to different roles' and 'present them with different html content'?

  Best,

  Kristian
• elisa

  August 2021

  Hi Bo,

  On which basis would you like to assign participants in the group different roles? If you wanted to assign them by order of appearance, it's easy:

  You could have a first component in your study (or use the waiting room in our examples) where you have an if statement:

  if (jatos.groupMembers.length == 1){
      //assign role A
  } else if (jatos.groupMembers.length == 2){
      //assign role B
  }
  
  

  Something like this.

  But Kristian is right: Writing group studies is always a bit complicated. Not difficult, just a lot to keep in mind. So this might not answer your question entirely, just ask again if you need any more help.

  Best

  Elisa
• yuanbopsy

  August 2021

  Hi Kristian and Elisa,

  Thank you so much for your response! Basically I want to randomly assign participants to different role (e.g., employee or employer in the inspect game) and present them with different matrix payoff and button choice. I found a pair of participants' GroupMemberId are two numbers adjacent to each other in the group study, so I have tried a solution that assign participants to different role according whether their GroupMemberId is even. It did worked, however, I am not sure whether my solution is robust because I want randomly match a pair of participants every round (see below). Or Elisa's solution is better?

  A related issues is that I want let a group of participants (e.g, Max total members is 8) repeatedly play the game with a certain number of round, but a pair of participants (e.g, Max active members is 2) were randomly matched before very round. Is it possible to do that by JATOS, how to achieve this?

  Thank you very much!

  Best,

  Bo
• elisa

  August 2021

  Hi Bo,

  Your solution to randomily assign roles is as valid as mine, as far as I can tell. As long as you don't have any systematic differences in sign-up order (e.g. women start always first, younger people start always first, etc) then it should work.

  I really didn't understand your second question tough. Could you explain it again? Do you want the same two people to always be paired, and also restrict the maximum number of workers? If this is the case, you'll probably need to use the batch session, but to tell you how exactly to do this it would be good to know how you're recriuting participants (i.e. I need to know if they each have a unique, permanent ID)

  Best

  Elisa
• kri

  August 2021 edited August 2021

  Hi Bo!

  If I understand you correctly you want to use jatos.groupMemberId to assign a participant to a role based on it being odd or even. That's not a good idea - it will only work on your local JATOS but not on a JATOS server where there are potentially many group studies running in parallel. The jatos.groupMemberId is just increased by one for each new participant but this participant is not necessarily one of your study. So I think Elisa's approach is better.

  And I also couldn't understand your second question :/

  Best,

  Kristian
• elisa

  August 2021

  Ah, good point, I missed the odd/even idea.
• yuanbopsy

  August 2021

  Hi Kristian and Elisa, about the first question, I am also care about there is possible that a pair of participants' GroupMemberId ( jatos.groupMemberId) are both even (or odd) when I collect data on my JATOS server. So, maybe Elisa's approach is better.

  With regard to the second question, image that I want to conduct the Prisoner’s Dilemma study where eight individuals (Max total members is 8) are recruited as one group to play a game. Each individual play the game 20 rounds, however, each individual is randomly matched with another one from the group in each round. That is, in each round, each individual's partner was not fixed, but was randomly matched from seven other group member. So, may in round 1, A and C were matched and play the game, but in round 2, A and E were matched. Of course, A and C is likely to match again in the subsequent rounds. How to achieve this by JATOS? I plan to recruit participant though distribute the General Single Worker link to participants or let them to the lab together and play the game.

  I don't know if you understand what I mean, if not you can let me explain it again. Thank you very much!

  Best,

  Bo

  
  
• kri

  August 2021

  Hi Bo,

  What you describe is possible. Like you suggested I'd start with the Prisoners Dilemma example and change it to your needs. The rough study flow could look like:

  • You need a waiting room component for 8 people
  • When all 8 participants joined assign them randomly into subgroups of 2. Store the subgroups in the group channel.
  • Start the game (second component) and let them play
  • In the end of the round send the result data of this round to JATOS
  • Jump back to the waiting room until all rounds are done

  How far are you already with your study? I can help you with the group study part.

  Best,

  Kristian
• yuanbopsy

  August 2021

  Thank you so much, Kristian. I have already finished the main part of the study. Could I send it to you via message of the forum? Or sent it to your email?

  Best,

  Bo
• kri

  August 2021

  Either is fine