No government agency has been more influenced by the work of K.R.W. Brewer than the US National Agricultural Statistics Service (NASS). Kott and Bailey (ICES-II, 2000) described the then new sampling design (maximal Brewer selection) and estimator (based on Brewer’s cosmetic calibration) being adapted for the agency’s quarterly crops/stocks program. This paper will discuss more recent research at NASS on these and on another
Brewer suggestion, using instrumental variables in regression estimation.
