I'm working on an EEG experiment using the oddball paradigm. My experiment already has 2 within-subject factors. I want to introduce a third factor but I don't want to make the experiment too long.
Thus, If I make the third factor within-subject, I will have to cut the number of my target trials by half; if I keep the number of target trials, I will have to make the factor between-subject.
I know from stats classes that a between-subject design gives far less power than a within-subject design and the disadvantage probably can't be offset by doubling the trials. But I assume such a difference is based on simulated data or behavioral data with less noise than the brain data.
I'm wondering for the EEG data, which gives more power, a within-subject factor or a between-subject factor with trials doubled?
Thank you!
Dear Chao,
You can rigorously test the power with the superpower package, as described here: https://aaroncaldwell.us/SuperpowerBook/the-no-way-interactions.html
As a very popular post by Prof. Simonsohn indicates: "With between subject designs, two-way interactions are ambitious. Three-ways are more like no-way." This was an eye-opening article for me (http://datacolada.org/17).
Halving the trials would be like multiplying your erp noise by √2 (Article Assessing the internal consistency of the event‐related pote...
). I think you would have potentially much more to lose if you would test two groups. Just a question: would asking participants to come twice be an option, with session as a random effect?Best
Suddha
Thanks for the information, Suddha Sourav
This article seems to suggest having the same subject come twice is no better than a between-subject design: https://today.duke.edu/2020/06/studies-brain-activity-aren%E2%80%99t-useful-scientists-thought
But I haven't given the option a careful thought.
Thank you for your suggestion.
Chao
Hi Chao Han, very interesting article, thank you! I will read it in detail.
For EEG, variance acros sessions seems to be much lower than across subjects, maybe this paper would be of relevance?
Article Systems, Subjects, Sessions: To What Extent Do These Factors...
Best
Suddha
- Badges
- Science topic