Subsegments for One-Unit Acceptance-Rejection Sampling for Covering Missed Housing Units

• Activity Number: 314
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Section on Survey Research Methods
• Author(s): Avinash C. Singh and Kirk Wolter*+
• Companies: NORC and NORC

Abstract

To overcome the problem of missed HU in area sampling, the HOI (half open interval) method is often used although its proper implementation remains elusive. As an alternative, we propose to use rejection sampling for drawing a random sample from each cluster with unknown list size for HU. The basic idea is to create subsegments or zones within a selected segment, the number being the cluster sample size, and then one unit is selected at random per subsegment. Using a reasonably sharp upper bound for each subsegment size, FI is provided with an ordered list of a few alternate random selections to use if an earlier selection does not correspond to an eligible HU. Other applications of the proposed methodology include substitution for ineligible HU selected from an initial cluster list in address-based area sampling, and substitution for nonresponding HU in any nonresponse follow-up survey.