Analyzing Survey Nonresponse

• Activity Number: 201
• Type: Roundtables
• Date/Time: Monday, August 4, 2014 : 12:30 PM to 1:50 PM
• Sponsor: Survey Research Methods Section
• Author(s): Polly Phipps*+
• Companies: Bureau of Labor Statistics
• Keywords: nonresponse bias ; response rates ; propensity models ; tree regression ; logistic regression ; auxiliary data

Abstract

Survey nonresponse and associated risks of nonresponse bias are major concerns for government agencies and other organizations conducting surveys. At the time of data collection, respondents may choose to decline survey participation (unit nonresponse) or participate selectively by responding to certain data items and not others (item nonresponse). Nonresponse analysis often requires modeling the probability that a given unit will respond to the survey based on known unit characteristics. Often, the choice of method and the intended use of the model is different in this context than when adjusting estimates for nonresponse. During this roundtable, we will discuss methods for modeling and identifying unit characteristics associated with nonresponse and nonresponse bias. We will emphasize methods designed to help understanding the effects these unit characteristics have on the response rate throughout the data-collection process.