What is an uninformative prior?
hardAnswer
- Prior meant to encode minimal prior belief.
- Options: (1) Flat / uniform (Beta(1,1)) — feels natural but is not invariant to reparameterization.
- (2) Jeffreys prior — proportional to √; invariant to reparameterization (Beta(0.5, 0.5) for Bernoulli).
- (3) Reference prior (Bernardo).
- For Bayesian A/B testing with weak prior belief, Jeffreys is a defensible default; results converge to MLE as n grows.
Check yourself — multiple choice
- Always uniform
- Flat (Beta(1,1)) — not reparam-invariant; Jeffreys (∝ √I(θ)) — reparam-invariant; reference priors; results → MLE at large n
- Random
- Not needed
Uninformative priors: flat / Jeffreys / reference; Jeffreys reparam-invariant.
#bayesian#estimation
Practise Statistics Fundamentals
215 interview questions in this topic.