This is a question from past paper about Bayesian statistics.

Suppose you have a non statistician friend Fred who wants to form a prior distribution for the average time, 8, he takes to drive to work. (a) Assume that Fred\" 3 opinion about 6 can be represented by a normal distribution N (a, b). Give question(s) that you might ask Fred in order to choose a and b, explaining to Fred what the questions are asking for according to the following: (i) Assessment of location (give 2 questions). [8] (ii) Assessment of spread (give 1 question). [4] (b) (i) Using Fred's answers to your questions, how can the assumption given in part (a) be checked using Fred's answers to your questions? [2] (ii) What answers would suggest a normal distribution is indeed a reasonable choice to represent Fred' s opinion? [3] (c) Suppose that a normal distribution is not a good representation of Fred' s opinion. Give 2 reasons why a gamma distribution might be a good representation instead. [2]