12.7 In Example 12.7 try a cubic structural model yi = 1 + 2Fi1 + 3F2 i + 4F3 i + ui and assess its

12.7 In Example 12.7 try a cubic structural model yi = β1 + β2Fi1 + β3F2 i

+ β4F3 i

+ ui and assess its predictive performance against the quadratic model. Try other values of k apart from 1000 (e.g. k = 1, k = 10); this means formally obtaining the posterior means and variances of yi,new and xi j,new (see Gelfand and Ghosh, 1998, pp. 4–5).