Hello everyone,
I'm currently planning a survey with four time-points : 0 months, 6 months, 12 months, 24 months. The goal is to explore the consequences and causes of kinesiophobia, excessive fear of movement and physical activity. I can provide more details if needed.
What type of model is usually recommended for this type of analysis?
I was also wondering how you would calculate sample size for such a study. I have seen that it is possible on R with some packages, but are there any ressources out there that explain how to do it ?
Thanks!
you can use categorical models. its like A B C D vs Y. the problem is this can be difficult on sample size. the only clue i suppose is trying to check how data was taken to determine the sample size. the analysis is really easy. but the sample size could say use binomial distribution to describe the data. the sample size for pi defects on category i is (2 / d)² x p x (1 – p) d is 0.1 , 0.05 or 0.01 depending on how much money is on scheme.
A common approach for such questions is to fit autoregressive/cross-lagged panel models, which are special (longitudinal) models of path analysis. I have a Youtube video, in which I explain those models in more detail:
https://youtu.be/q0uGmvRHOCI
The lavaan package in R is capable of running autoregressive/cross-lagged panel models. Sample size estimations for path analysis are best done using simulations.