Ononot — Consider a two-phase sampling design. We estimate the mean o

Guest Session: 1 Question Remaining. Create Account to save progress.

Data Collectionhard

0:00.0

Consider a two-phase sampling design. We estimate the mean of $Y$ using a regression estimator $\bar{y}_{lr} = \bar{y} + b(\bar{X}' - \bar{x})$ . What is the main condition under which this estimator is more efficient than a simple mean estimator $\bar{y}$ ?