9We will develop these ideas more fully in Chap. 10. For now, we simply note that Eqs. (6A.4) and (6A.5) are based on best linear unbiased estimators. Equation (6A.4) is closely related to the regression result. If we regress θ against z, i.e., θ_n = a + b · z_n + ∈_n, then E{θ_n |z_n} = a+ b · z_n. The regression coefficient b is Cov(θ,z) · Var⁻¹{z}. Assuming a = 0 leads to Eq. (6A.4).