Power analysis for longitudinal data (не гугельные)
- Mplus would allow you to conduct a Monte Carlo power analysis for the model you are describing (http://statmodel.com), (http://www.statmodel.com/bmuthen/ED231e/RelatedArticles/Article_097.pdf).
- If you want to use a multilevel analysis, you can do simulations using Mplus as mentioned by Meredith. Or you might try PASS 14, which has some capability for two-level models: http://www.ncss.com/software/pass/ PASS and Mplus are not free, but you can download a trial version of PASS.
- I suggest treating it as a multilevel analysis and using methods outlined by Gelman: http://www.cambridge.org/us/academic/subjects/statistics-probability/statistical-theory-and-methods/data-analysis-using-regression-and-multilevelhierarchical-models You can find R code to do so for free.
- Raudenbush, S. W., et al. (2011). Optimal Design Software for Multi-level and Longitudinal Research (Version 3.01) [Software]. Available from www.wtgrantfoundation.org. https://sites.google.com/site/optimaldesignsoftware/home
- I would recommend a simulation based approach if you envisage a large N. A simulation based approach gives you control over all possible scenarios. If you get your hands on the full text version of Rick Wicklin's book (https://www.sas.com/storefront/aux/en/spsimulationofdata/65378_excerpt.pdf) and with a little review of appropriate chapters, you can have a simulation based program that you can fine tune for most possible scenarios, save your codes for future projects and get your job done!
Statistical Computing Seminars/Introduction to Power Analysis