Millisecond Forums

counterbalancing with web delivery

https://forums.millisecond.com/Topic23962.aspx

By Paul Lattimore - 3/5/2018

Hi,
I'm using Single target IAT based on script available in millisecondlibrary. I will run it on web so I need to figure out how to counterbalance allocation to compatible and incompatible categories. The current PC based script does this by assigning group number (1,2) on start-up. That's fine if you are sitting beside the participant. Participants will be promted via wb delivery to enter 8-digit alphanumeric ID.
 Can anyone tell me how I would write the counterbalancing into the script or use some other means to ensure even allocation.
Thanks,
Paul
The script is currently set for PC platform as follows:

Groupassignment is done by groupnumber

<expt>
/subjects = (1 of 2)
/groupassignment = random
/ blocks = [1=practice_compatibletest; 2=compatibletestinstructions; 3=compatibletest;
 4=practice_incompatibletest; 5=incompatibletestinstructions; 6=incompatibletest; 7=summary]
/onexptend = [values.completed = 1]
</expt>

<expt>
/subjects = (2 of 2)
/groupassignment = random
/ blocks = [1=practice_incompatibletest; 2=incompatibletestinstructions; 3=incompatibletest;
4=practice_compatibletest; 5=compatibletestinstructions; 6=compatibletest; 7=summary]
/onexptend = [values.completed = 1]
</expt>


By Dave - 3/6/2018

Paul Lattimore - Tuesday, March 6, 2018
Hi,
I'm using Single target IAT based on script available in millisecondlibrary. I will run it on web so I need to figure out how to counterbalance allocation to compatible and incompatible categories. The current PC based script does this by assigning group number (1,2) on start-up. That's fine if you are sitting beside the participant. Participants will be promted via wb delivery to enter 8-digit alphanumeric ID.
 Can anyone tell me how I would write the counterbalancing into the script or use some other means to ensure even allocation.
Thanks,
Paul
The script is currently set for PC platform as follows:

Groupassignment is done by groupnumber

<expt>
/subjects = (1 of 2)
/groupassignment = random
/ blocks = [1=practice_compatibletest; 2=compatibletestinstructions; 3=compatibletest;
 4=practice_incompatibletest; 5=incompatibletestinstructions; 6=incompatibletest; 7=summary]
/onexptend = [values.completed = 1]
</expt>

<expt>
/subjects = (2 of 2)
/groupassignment = random
/ blocks = [1=practice_incompatibletest; 2=incompatibletestinstructions; 3=incompatibletest;
4=practice_compatibletest; 5=compatibletestinstructions; 6=compatibletest; 7=summary]
/onexptend = [values.completed = 1]
</expt>



<expt>
/subjects = (1 of 2)
/groupassignment = random
/ blocks = [1=practice_compatibletest; 2=compatibletestinstructions; 3=compatibletest;
 4=practice_incompatibletest; 5=incompatibletestinstructions; 6=incompatibletest; 7=summary]
/onexptend = [values.completed = 1]
</expt>

<expt>
/subjects = (2 of 2)
/groupassignment = random
/ blocks = [1=practice_incompatibletest; 2=incompatibletestinstructions; 3=incompatibletest;
4=practice_compatibletest; 5=compatibletestinstructions; 6=compatibletest; 7=summary]
/onexptend = [values.completed = 1]
</expt>

The condition assignment in the code you posted is _not_ done by group number. It is set to random, i.e. the condition assignment is basically like a coin toss, and the condition administered is unrelated to either the participant's subject id or her/his group number. If you want condition assignment based on group number, you should do

<expt>
/subjects = (1 of 2)
/groupassignment = groupnumber
/ blocks = [1=practice_compatibletest; 2=compatibletestinstructions; 3=compatibletest;
 4=practice_incompatibletest; 5=incompatibletestinstructions; 6=incompatibletest; 7=summary]
/onexptend = [values.completed = 1]
</expt>

<expt>
/subjects = (2 of 2)
/groupassignment = groupnumber
/ blocks = [1=practice_incompatibletest; 2=incompatibletestinstructions; 3=incompatibletest;
4=practice_compatibletest; 5=compatibletestinstructions; 6=compatibletest; 7=summary]
/onexptend = [values.completed = 1]
</expt>

To ensure approximately even assignment to conditions on the web, then, select e.g. sequentially generated group numbers in the web experiment's settings when you upload and configure the script.