6.17. For a bivariate normal sample on Y1; Y2 with parameters y m1, m2, s11, s12, s22) and values of Y2 missing, state for

6.17. For a bivariate normal sample on ðY1; Y2Þ with parameters y ¼ ðm1, m2, s11, s12, s22) and values of Y2 missing, state for the following missing-data mechanisms whether (i) the data are MAR, and (ii) the missing-data mechanism is ignorable for likelihood-based inference.

(a) Prðy2 missing j y1, y2, y, cÞ ¼ expðc0 þ c1y1Þ=f1 þ expðc0 þ c1y1Þg, c ¼ ðc0, c1Þ distinct from y.

(b) Prðy2 missing j y1, y2, y, cÞ ¼ expðc0 þ c1y2Þ=f1 þ expðc0 þ c1y2Þg, c ¼ ðc0, c1Þ distinct from y.

(d) Prðy2 missing j y1, y2, y, cÞ ¼ 0:5 expðm1 þ cy1Þ=f1 þ expðm1 þ cy1Þg scalar c distinct from y.