Table 2.

Numerical summaries of densities associated with random slopes and intercept estimates. The median (p₅₀) and middle 95% quantiles (p_2.5, p_97.5) based on the generated draws from the corresponding distribution are presented. All results correspond to prior {g, h} = {g1, h1} as described in the text. The b_0i and b_1i columns correspond to the posterior predictive distributions for a new random slope and intercept, respectively. The Mean, Var and Cov columns are derived from the (marginalized over cluster configurations, for M3 and M4) posterior density estimates of the hyperprior parameters corresponding to the random effects.

	p50(p2.5, p97.5)
Fixed Effect Parameters	β1 Gender	β2 hMLH1	β3 hMSH2	σ Error Scale
M1	1.5 (−0.3,3.2)	−6.8 (−9.7,−3.9)	−6.4 (−9.2,−3.7)	9.8 (9.1,10.5)
M2	1.5 (−0.3,3.2)			9.8 (9.1,10.5)
M3	1.5 (−0.3,3.3)	−6.6 (−9.6,−3.7)	−6.3 (−9.1,−3.5)	9.8 (9.1,10.5)
M4	1.5 (−0.3,3.3)	−6.3 (−9.8,-2.6)	−6.1 (−9.4,-2.6)	9.6 (8.9,10.4)

Random Effect Parameters	b0i	b1i	Mean[b0i]	Mean[b1i]
M1	57.3 (50.9,63.9)	−2.3 (−5.7,1.0)	57.4 (54.0,60.7)	−2.3 (−3.5,−1.1)
M2 hMLH1	51.5 (44.9,58.3)	−2.8 (−6.2,0.7)	51.5 (48.4,54.7)	−2.8 (−4.3,−1.2)
M2 hMSH2	51.3 (44.9,57.8)	−2.4 (−6.1,1.1)	51.4 (48.3,54.2)	−2.5 (−3.8,−1.0)
M2 hMSH6	54.0 (46.5,62.2)	−1.1 (−5.3,3.0)	53.9 (49.9,59.1)	−1.0 (−3.3, 1.1)
M3	57.2 (50.4,64.2)	−2.3 (−6.0,1.2)	57.2 (53.8,60.6)	−2.3 (−3.5,−1.1)
M4	57.2 (44.8,69.7)	−2.5 (−10.2,5.2)	57.3 (51.5,62.9)	−2.5 (−5.6, 0.6)

Random Effect Variance Parameters	Var[b0i]	Cov[b0i, b1i]	Var[b1i]
M1	5.6 (0.8, 26.1)	−2.2 (−12.0, 0.4)	1.9 (0.3, 7.2)
M2 hMLH1	6.0 (0.7, 31.0)	−2.2 (−12.0, 0.5)	1.8 (0.3, 7.0)
M2 hMSH2	6.0 (0.7, 27.4)	−2.5 (−13.4, 0.3)	2.1 (0.3, 8.5)
M2 hMSH6	6.4 (0.7, 39.6)	−2.5 (−18.6, 0.5)	2.1 (0.3,11.7)
M3	6.0 (0.7, 25.7)	−2.4 (−11.9, 0.3)	2.0 (0.3, 7.2)
M4	19.0 (2.5,100.9)	−6.7 (−41.4,11.9)	7.0 (1.0,72.6)