I am currently doing an undergraduate mathematical statistics course. One questions in Bayesian methods that i can’t seem to grasp is knowing when to use an improper uniform prior? After learning conjugate priors, my instincts are to directly go on to use conjugate priors whenever possible. However some times in the solution manual, it uses an improper uniform prior. Is there a trick to knowing when to use an improper uniform prior than a conjugate prior? Thank you.