Implementing variable restrictions in full factorial designs

FranHartung · May 2017

Hi!

I just started using OpenSesame and so far I love how easy and robust everything works compared to alternatives. I have trouble to figure out stimulus selection restrictions though. I have a 2(context) x 2(target) x 20(item) design, adding up to 80 possible combinations. The same target however should not appear twice for the same participant though, which means that each item X target combination should only be presented in one of the two possible contexts. Across my entire testing population (N=30), the distribution of unique target X item combinations should be balanced (no exact match needed).

Any ideas how to implent this?

Franziska

sebastiaan · May 2017

Hi Franziska,

What you're describing is not a full-factorial design, because that (by definition) would have all combinations of factors in it. In your case, you have the following design:

target (2 levels) and item (20 levels) are fully crossed, leading to 2 × 20 = 40 rows in the loop table
Once you've created these 40 trials, you add a column context where the first 20 trials are A and the last 20 trials are B (or whatever your contexts are called).
To make the pairing between context and target × item fully random, we need to shuffle the context column separately. You can do this using an advanced loop operation: shuffle context, as described here.

Does that make sense? Once you let go of the idea that this is a full-factorial design, it becomes very easy.

Cheers!
Sebastiaan

FranHartung · May 2017

Hi Sebastiaan,

Thank you so much for your fast response! The documentation and help forum are amazing! Great job! My new fMRI experiment is ready to start!

Franziska

Howdy, Stranger!

Categories

Implementing variable restrictions in full factorial designs

Comments

Howdy, Stranger!

Quick Links

Categories

Implementing variable restrictions in full factorial designs

Comments