Hi there,

I have a randomization question I couldn't find answered earlier in the forums. I have three tasks that I have participants performing in a random order. I have implemented this as follows: 

<expt ALL>
/ blocks = [1-3=noreplace(dilute, latency, choice)]
</expt>

This would be well and good, but the issue is that have three versions of each block that need to be coded separately. 

Thus, I have three versions of the "dilute" block, three versions of the "latency" block, and three versions of the "choice" block. They are coded separately because they are all substantively different. Because we don't need these fully crossed, our experimental design allows us to reduce this to just "three paths" through the study. 

1. dilute_1,  latency_2 , choice_3

2. dilute_3, latency_1, choice_2 

3. dilute_2, latency_3, choice_1 

To reiterate, that is 9 blocks with three "paths."  

I therefore wrote three versions of the experiment command. Is there a way to have inquisit3 randomly each participant into one of these three paths at the start of the study? 

/ blocks = [1-3=noreplace(dilute1, latency2, choice3)]
/ blocks = [1-3=noreplace(dilute3, latency1, choice2)]
/ blocks = [1-3=noreplace(dilute2, latency3, choice1)]

How would I code that? I recall hearing that in inquisit 4 there are ways to do it with lists I believe. How would I code this in inquisit 3? Help is greatly appreciated. 

You either need several (here: 3) <expt> elements or set up the groups using the <variables> element. Details are available in the documentation. For groupassignment, see the documentation for the /subjects and /groupassignment attributes.

You'll also find those basics covered in the "Simplified IAT" or the "Standard Picture IAT" tutorials included in the documentation.

This is rather elementary stuff, so you do not need <list> elements and Inquisit 3 will work just fine.

Regards,

~Dave

Thanks for the speedy reply. 

I can see how the variables command will do the trick. My other question now is, does that logic / solution still work if each task is composed of a series of blocks instead of individual blocks?

Tom

Why wouldn't it?