Millisecond Forums

script view in Mac and even group assignment for completed exp only

https://forums.millisecond.com/Topic31409.aspx

By tecnika - 4/26/2021

Hello,

I would have some queries about the script linked below that I wrote in a Windows laptop with the latest update of Inquisit 5.
https://mili2nd.co/5qzb
1) A colleague tried to open the script file with the same version of Inquisit 5 in an iMac. The experiment could run fine, however he could see multiple repetitions for each line of the script (that are not part of the original version). 
Any idea on how to solve the issue? Not sure you can see the screenshot attached, in case I can send you an email with it as well as the registration number if you suggest to try re-registering the desktop version.

2) The experiment has a between subject design with 24 conditions and permutations assigned with groupnumber (random assignment without replacement). However the data collected in the pilot study (https://mili2nd.co/vkyb) don't have all the conditions evenly distributed. Some participants quit before completing or even loading the experiment...this may disrupt the number assignment maybe? Is there a way to control this so only the completed experiments count in the condition assigned?

Many thanks,

Elena
By Dave - 4/26/2021

 
tecnika - 4/26/2021
Hello,

I would have some queries about the script linked below that I wrote in a Windows laptop with the latest update of Inquisit 5.
https://mili2nd.co/5qzb
1) A colleague tried to open the script file with the same version of Inquisit 5 in an iMac. The experiment could run fine, however he could see multiple repetitions for each line of the script (that are not part of the original version). 
Any idea on how to solve the issue? Not sure you can see the screenshot attached, in case I can send you an email with it as well as the registration number if you suggest to try re-registering the desktop version.

2) The experiment has a between subject design with 24 conditions and permutations assigned with groupnumber (random assignment without replacement). However the data collected in the pilot study (https://mili2nd.co/vkyb) don't have all the conditions evenly distributed. Some participants quit before completing or even loading the experiment...this may disrupt the number assignment maybe? Is there a way to control this so only the completed experiments count in the condition assigned?

Many thanks,

Elena




> 1) A colleague tried to open the script file with the same version of Inquisit 5 in an iMac. The experiment could run fine, however he could see multiple repetitions for each line of the script (that are not part of the original version).

The Inquisit 5 Lab editor is not compatible with macOS Big Sur, and the repeated lines are a consequence of that. If need be, the script can be viewed using the trial version of Inquisit 6 Lab under Big Sur.

>
2) The experiment has a between subject design with 24 conditions and permutations assigned with groupnumber (random assignment without replacement). However the data collected in the pilot study (https://mili2nd.co/vkyb) don't have all the conditions evenly distributed.
> Some participants quit before completing or even loading the experiment...this may disrupt the number assignment maybe?

Correct. A groupnumber is generated when somebody visits the experiment's start page, regardless of whether that person ends up starting or completing the study. This can (and under real-world conditions typically will) lead to some imbalances across groups, unequal amounts of participants per group.

> Is there a way to control this so only the completed experiments count in the condition assigned?

No. The only thing you can do is to pre-assign conditions to your participants and pass the groupnumber in via URL parameter. This of course is only workable if you know your participant pool in advance.