he normal distribution is a good prior on the variance parameter, 2 , when the observations y1,...,yn are normally distributed, i.e., f(yi|,2) (consider the to be know). Question 5Answer a. Yes, and it yields a closed-form posterior on 2 (i.e., it is a conjugate prior in that model). b. Yes, but it leads to intractable posterior, so approximate methods should be used. c. No, because knowing turns the inference of 2 impossible regardless the prior. d. No. The support of such prior would not even be adequate for the required support of this parameter