By jessrh - 1/25/2016
Hi all,
I want to randomize subjects into three different conditions, but want the number of subjects in each condition to be equal. I found here that one way to do this is to use separate .iqx files for each condition and host them in a batch file which will randomize, like this:
<batch>
/ subjects = (1 of 3)
/ groupassignment = groupnumber
/ file = "MOODROC_positive.iqx"
</batch>
<batch>
/ subjects = (2 of 3)
/ groupassignment = groupnumber
/ file = "MOODROC_negative.iqx"
</batch>
<batch>
/ subjects = (3 of 3)
/ groupassignment = groupnumber
/ file = "MOODROC_neutral.iqx"
</batch>
However, I've run the batch script several times and it has not randomized the conditions (it only presents the 3rd condition, "neutral"). Any input into how to resolve this issue would be much appreciated!
|
By Dave - 1/25/2016
I assume you are running this via Inquisit Web and that you've set your launch page up to generate random group numbers.. If so, you need to refresh the launch page between your test runs. The group number is generated when your browser first loads the page. If you don't refresh it, no new group number will be generated and you will get the same condition repeatedly.
|
By jessrh - 1/25/2016
Hi Dave,
Thank you for your quick reply. This fixed my randomization issue, but after several runs, the experiment began ending prematurely, presenting the "The session is complete, and your data were successfully uploaded. Thank you for your participation!" screen before the task even started. It seems like now the website might not be calling the scripts in my batch file, though it was working before online and continues to work on the desktop version of Inquisit. Any thoughts on this?
Thank you so much!
Jess
|
By Dave - 1/25/2016
This indicates an issue in your script(s). Specifically, you are passing in a group number or subject number for which no condition is assigned (and hence nothing will be run). You *will* see this happening in Inquisit Lab also, you just need to test out all the conditions (i.e., enter different group and/or subject numbers).
|
By jessrh - 1/26/2016
Is there a way to assign the group numbers 1, 2 and 3 on Inquisit Web and have it repeat this assignment (1, 2 and 3) for several hundred participants?
|
By Dave - 1/26/2016
You can opt for generating group ids sequentially. There is no need to only repeat the exact numbers 1,2 and 3, because assignment is based on the modulus of the provided group number. I.e.
1 mod 3 = 1 -> condition 1
2 mod 3 = 2 -> condition 2
3 mod 3 = 0 -> condition 3
4 mod 3 = 1 -> condition 1
5 mod 3 = 2 -> condition 2
6 mod 3 = 0 -> condition 3
and so forth.
For details see this post: https://www.millisecond.com/forums/Topic13856.aspx
|
By jessrh - 1/26/2016
Thank you for all of your help! I have one last issue that arose after implementing the sequential query. The experiment only runs on Google Chrome when a new Chrome session is established - i.e. I have to close Chrome and open a new session to begin the experiment, or else it goes straight to the "thank you" page as before. In both Safari and Firefox, the link worked when a I began a new web session, did not work when I refreshed the search bar in the same tab, but did worked when the link was copy-pasted into a new tab on the same session. Is there anything I can do to fix this issue?
|
|