Challenges in the Treatment of Nonresponse for Selected Business Surveys

• Activity Number: 644
• Type: Topic Contributed
• Date/Time: Thursday, August 2, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Katherine Thompson*+ and Katrina Washington
• Companies: U.S. Census Bureau and U.S. Census Bureau
• Keywords: weight adjustment ; ratio imputation ; prediction model ; propensity model ; balance complex

Abstract

Probability sample selection procedures gift methodologists with quite a bit of control before data collection; the "optimal" design for a given frame ensures that the selected sample is representative. This situation can change after data collection. Not all sample units respond and those that do will not always provide data on every questioned characteristic, which can lead to biased estimates of totals. The degree of bias is determined by several factors, including the representativeness of the respondent set, the magnitude of the aggregated missing data values, and the effects of improper adjustment procedures. Business surveys are characterized by highly skewed populations, so a large proportion of the estimated totals originate from a small set of cases. When surveys collect totals and associated additive details, the effect of non-response bias can be particularly evident with the