Identifying and Comparing Characteristics of Nonrespondents throughout the Data Collection Process

• Activity Number: 542
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Government Statistics Section
• Author(s): Morgan Earp*+ and Daniell Toth and Polly Phipps and Charlotte Oslund
• Companies: Bureau of Labor Statistics and Bureau of Labor Statistics and Bureau of Labor Statistics and Bureau of Labor Statistics
• Keywords: Nonresponse ; Regression Trees ; Establishment Surveys

Abstract

Establishment surveys usually go through a process of verifying address and contact information, as well as attempting to gain participation before data collection begins. At each phase in the data collection process, there is potential for nonresponse. Characteristics of establishments that are difficult to locate and contact may not be the same as those that refuse to participate or respond to the survey; therefore, it is important that we assess nonresponse at each phase of data collection. The Bureau of Labor Statistics Job Openings and Labor Turnover Survey (JOLTS) uses a three-fold data collection process: address refinement to verify address and contact information, enrollment to recruit participants into the survey, and data collection when the survey is administered. Using auxiliary data related to key JOLTS estimates, we identify and compare characteristics associated with nonresponse at each of the three phases of nonresponse, as well as both unit and item nonresponse during the data collection phase. The results of this study can be used to better allocate resources when attempting to reduce nonresponse.