In 2019, I helped to develop a series of workshops to prepare researchers to submit a development-type or efficacy-type grant. We called it "gear up for your grant proposal". One of the talks I gave as part of that workshop series was an Introduction to Power Analysis. I posted the slides to it on figshare HERE. The organizers also recorded it, and while it's not yet available, I do have hopes that the video might be one day!
At the crux of my argument about power analysis is that it's not as linear as you might think. What you probably think is that you develop a research question, run a power analysis, and that power analysis tells you how many people you need to test.
In practice, power analysis is much more cyclical. It involves your budget, your research questions, analysis plan, data structure, how big of an effect you can reasonably expect to find, and the power analysis simultaneously and in combination with one another. Something more like this:
When I'm planning a study, I go around and around this interconnected web for a few days before narrowing down what we will actually do for a project. It's complex, deep thinking work. But it's also really fun. Have a read of my power-point introduction to how I approach power analysis, and consider whether using it will be useful as you are planning your own.
If you want to cite the presentation, please use the following:
Logan,Jessica(2019). Introduction to Power Analysis. figshare. Presentation. https://doi.org/10.6084/m9.figshare.8236409.v1