Hi guys,
I am interested in conducting a longitudinal study investigating the development of within-connectivity of the DMN in autistic and non-autistic children. Scans will take place at 8, 13 and 18.
My dependent variable will be DMN within connectivity calculated by averaging ROI-to-ROI connectivity and pairwise correlation between time series of regions within two hemispheres and between the hemispheres will be averaged. These averages will then be averaged to get a within-connectivity.
My fixed variables will be time and autism as well as their interaction. My random effects will be subject-specific intercepts and slopes and my covariates will be gender and education (I am expecting to add more covariates).
I am curious as to how I calculate a sample size a priori for this. I know I need to define an effect size which I have a rough idea of and of course power and alpha. Can this be done on G*Power (to my knowledge I dont think so, but I may be missing something).
Any help would be useful.
Thanks in advance
GLIMMPSE (General Linear Mixed Model Power and Sample Size) likely would be able to accommodate your multilevel longitudinal design. There is a free online version:
https://samplesizeshop.org/glimmpse-power-software/