Let X be a random variable with a distribution Pθ in {Pθ: θ ∈ Θ}, fθ be the p.d.f. of Pθ w.r.t. a measure ν, A be an event, and PA = {fθIA/Pθ(A) : θ ∈ Θ}.
(a) Show that if T (X) is sufficient for Pθ ∈ P, then it is sufficient for Pθ ∈ PA.
(b) Show that if T is sufficient and complete for Pθ ∈ P, then it is complete for Pθ ∈ PA.