Reverse-Engineering the Prior — hard Bayesian Inference problem | Incognition

Reverse-Engineering the Prior

30 pts · 30 coins

Bayesian InferenceHard

Reverse-Engineering the Prior

A colleague hands you a finished analysis but has lost her working notes.

All you know is this:

She used a symmetric Beta prior $\text{Beta}(c, c)$ on a success probability $\theta$ , with $c > 1$ .
She then observed $8$ successes and $4$ failures.
Her reported MAP estimate — the mode of the resulting posterior — is exactly $0.60$ .

What value of $c$ did she use?

Hints

Select all that apply.