I'm runing MCMC using Metropolis-Hasting algorithm to fit an equation with 6 parameters on a dataset composed of 30 instances. How will the fact that my dataset is so small impact the posterio distribution of each parameter? And therefore its confidence interval?
As normally, the less data there is, the larger are the uncertainty intervals(posterior/confidence). The intervals will still be correct, just not as useful.
The 6:30 ratio is big, but I have seen examples where that would be more than enough.