Uniform prior for all parameters


Suppose I want to estimate a model and all I know about the parameters is that they lie within certain ranges. I don’t really know which values are more plausible than others within the ranges, and so am willing to assume uniform priors for all parameters. What could be potential problems with this?

with uniform priors, it is possible that the optimiser will hit the boundaries.
it is better to have loose, but slightly curved priors.
good luck. you need all the luck in the world, with these models.:slight_smile: