[open] 4 group randomization

Comments

• EoinTravers

  August 2013 edited 4:35AM

  Hi Johannes,
  There may be more elegant way of solving your problem, but I've done something similar as follows.

  • First, you'll need to generate your four sequences, and place them within the experiment structure so that they would run one after the other. You'll probably want to place them after the instructions, if the same instructions apply to each sequence.

  • Next, you can place an inline_script section at the start of your experiment, containing something like this:

  from random import randint
  sequence_number = randint(1,4)
  exp.set('sequence_number', sequence_number)
  

  • Finally, go to the sequence item which contains your four possible sequences (sorry if this is starting to sound confusing!), which in this case will probably be the automatically generated one, called 'experiment'.
    Beside each of your sequences, there's a 'Run if' option, which is set to 'always' by default. Change this to `[sequence_number] == 1' for the first sequence (and 2, 3, and 4 for the others).

  What all this will do is essentially:
  - Generate a random number between 1 and 4.
  - Run sequences 1, 2, 3, or 4, depending on the number generated.

  Hope this helps!

  Eoin
• EoinTravers

  August 2013 edited 4:35AM

  Oh, another possibility is, instead of completely randomizing which sequence is presented, you counterbalance it according to the subject number.

  subject_nr = exp.get('subject_nr')
  sequence_number = (subject_nr % 4) + 1 
  # Gives the remainder of subject_nr divided by 4, plus 1.
  # Answer will be either 1, 2, 3, or 4.
  exp.set('sequence_number', sequence_number)
  
• johannes

  August 2013 edited 4:35AM

  hi,
  cool thank you very much! For your first option I would not need any loop right?
  your second option is very nice, to make sure I have equal sample sizes in every group. But if you just divide the subject number by 4 and ad 1 ist that getting bigger and bigger, since the subject number increases with every participant?

  1000 thanks!
  Johannes
• johannes

  August 2013 edited August 2013

  ok, never mind, I figured it out (about the subject number divided by 4 +1

  solved....