You would need to present evidence that rates of non-vaccination between immunocompromised and non-immunocompromised individuals are not equal for this to be true. Anecdotally, I know a handful of "immunocompromised" individuals (RA and its relatives run rampant in my family), and all of them are vaccinated and boosted.

20% of Florida hasn't received a single dose [1]. Even if you assume that 2.7% of Florida is immunocompromised (using a figure from a Google search [2], maybe not the most accurate), and that 0% of the immunocompromised population is vaccinated (bold assumption), that's still only ~14% of this pool that's immunocompromised.

You can't make an analytic deconstruction of bias and then just stop when you get to a level that makes you comfortable.

1: https://usafacts.org/visualizations/covid-vaccine-tracker-st...

2: https://www.google.com/search?q=percentage+of+population+tha...