252 – Retention, Attrition, and Respondent Burden
Measuring Attrition in Long-Term Longitudinal Surveys
Aref Dajani
U.S. Census Bureau
Toni Warner
U.S. Census Bureau
For longitudinal surveys, attrition is defined as a reduction or decrease in size, number, or strength of the sample, whether permanently or intermittently. Attrition in a longitudinal survey reduces the statistical reliability of both cross-sectional and longitudinal estimates. Attrition can come in many forms, and it is important to define attrition types for a survey before evaluating a survey's performance. Generally, care should be taken to separate reasons for attrition into those that are not related to the survey design and those that are or may be related to the survey design. In planning for the 2015 Redesign of the American Housing Survey (AHS), several metrics were calculated to measure attrition over the past twenty-six years.
