Reweighting of the Primary Sampling Units in the National Automotive SamplingSystem

摘要

The National Automotive Sampling System (NASS) is based on a multi-stageprobability sample which requires periodic maintenance to preseve the probability basis and minimize any bias which can creep in over time. Since the NASS was first implemented in 1979, the first stage Primary Sampling Units (PSUs) have been reselected once. In an ideal world, the PSUs should be selected periodically to allow for any shifts in the number and type of crashes occurring across the country. This research note describes the process recently used to account for shifts in the number of crashes without a reselection. There are three stages to the NASS sample design: The first stage--the selection of the PSUs; The second stage--the selection of the police jurisdictions within a PSU; and, The third stage--the selection of crashes for investigation. The subject of this research note involves only the first stage.