Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Supported by

[open] Conduct a power analysis for a multilvel analysis in Jasp?

edited March 2016 in JASP & BayesFactor

Hello,

I wanted to ask whether it is possible a) conduct a power analysis for a multilevel analysis. A reviewer asks me to repeat the study we already ran and conduct a power analysis in advance. So there is no data yet of the new study.

I was wondering whether you have some tips and advice for me regarding a power analysis for a multilevel design. The nested structure of the study is as follows:

country, id, face 1, emotion 1
country, id, face 1, emotion 2
country, id, face 1, emotion 3
country, id, face 2, emotion 1
country, id, face 2, emotion 2
country, id, face 2, emotion 3
country, id, face 1, emotion 1
country, id, face 1, emotion 2
country, id, face 1, emotion 3
country, id, face 2, emotion 1
country, id, face 2, emotion 2
country, id, face 2, emotion 3

Emotion recognition scores of participants (id) are collected in 2 different countries and 'face' and 'emotion' refer to stimulus characteristics. In addition, there is a between-subject variable that refers to stimulus clothing with which the presented face is covered (no coverage, niqab, a black bar). Unfortunately, the number of data points per individual is quite low.

Thank you in advance for your comment,

Best regards,

Eliska

Comments

  • edited 4:14PM

    Hi Eliska,

    Could you clarify a few things first?

    • What exactly is the design? (The table is not clear to me.) What is the dependent variable, which factors are varied within subjects, and which are varied between? If you find it useful, you could usethis notation to make your design explicit:

    • What kind of analysis are you going to do? A (non-Bayesian) repeated measures with between-subject factors?

    • For which effect do you want to do a power analysis? If you have multiple factors, the power is different for each factor, and all the interactions. For example, if you have a large effect of country, but a small effect of clothing, then you would have more power to detect the effect of country than that of of clothing.

    Cheers!
    Sebastiaan

    There's much bigger issues in the world, I know. But I first have to take care of the world I know.
    cogsci.nl/smathot

Sign In or Register to comment.