Pre-Sampling Model-Based Inference III

• Activity Number: 413
• Type: Contributed
• Date/Time: Tuesday, August 2, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Stephen Woodruff*+
• Companies: Specified Designs
• Keywords: Model Based Inference ; Design Based Inference ; Pre-sampling Inference

Abstract

In survey sampling, a sample unit`s study variables are expanded to population totals by probability design based (DB) or model based (MB) expansions that implicitly treat a unit`s study variables as totals over entities called atoms contained in a unit. Expansions would make little sense if applied to unit statistics other than atom totals, for example, industry surveys where units are businesses, atoms are employees, and unit study variables are establishment totals for hours, wages, and number of workers. Other examples are household, mail, and ecological surveys. Woodruff (2010, 2009), derived Pre-sampling Model Based Inference from this atom structure, a structure that depends on probability sampling of population units and of atoms that comprise each unit. It provides estimates that retain the best properties of both MB and DB inference and that eliminate the main shortcomings of e