On Adjusting Sample Tabulations to Census Counts