1. Suppose Uncle Jack shows you a coin, and the probability that flipping the coin gives the heads side is p Î[0,1]. Jack is tricky, so either p=0, p=1/2, or p=1, with probability 1/3 each. Also suppose that Uncle Jack is going to flip the coin 2 times, and let X = the number of heads. So conditional on p, X ~ binom(2,p) . a. What is E(p)? ½ What is E(p2)? 5/12 What is var(p)? 1/6 b. What is E(2p)? 1 What is var(2p)? 4/6 c. In general, a joint distribution is the product of a __marginal______ distribution and a __conditional_______ distribution. Using this, what is the joint distribution of p and X? ì (0,0) 1/ 3 ï 0 ï (0,1) ï (0,2) 0 ï ï (1/ 2,0) 1/ 3 ´ 1/ 4 ï (p,x) = í (1/ 2,1) with probability 1/ 3 ´ 1/ 2 ï (1/ 2,2) 1/ 3 ´ 1/ 4 ï ï (1,0) 0 ï 0 ï (1,1) ï (1,2) 1/ 3 î d. What are the formulas for the marginal mean and variance of X? (Marginalizing over p, of course.) e. Apply those formulas to get the marginal mean and variance of X in this case. E(X) = nE(p)=2*1/2 =1 var(X) =E(var(X|p))+var(E(X|p))=E(np(1-p)) + n2 var(p) = 2(1/2) – 2(5/12) + 4(1/6) = 1 - 5/6 + 2/3 = 5/6 E(p) = 1/2 E(X) = nE(p)=2*1/2 =1 E(p2) = 1/3*0 + 1/3*1/4 + 1/3*1 = 5/12 Var(p) = E(p2) – (E(p))2 = 5/12 – ¼ = 2/12 = 1/6 var(X) =E(var(X|p))+var(E(X|p))=E(np(1-p)) + n2 var(p) = 2(1/2) – 2(5/12) + 4(1/6) = 1 - 5/6 + 2/3 = 5/6 -- I forgot that var(np)=n2var(p). Verify: Pr(X=0) = 1/3 + 1/12 +0= 5/12, Pr(X=1)=0 + 1/6 + 0 = 1/6, Pr(X=2)=0+1/12+1/3=5/12. E(X) =0+1*1/6 + 2*5/12 = 12/12 = 1 OK. E(X2)=0+1*1/6 + 4*5/12 = 22/12=11/6 Var(X)=11/6 – (1)2 = 5/6. Correct by simulation. X=2, pr(P=1 | X=2) = Pr(1,2)/Pr(X=2) = (1/3) / (1/3 + 1/12) = 4/5.