In the longitudinal estimation for the Survey of Income and Participation (SIPP), a weighting cell adjustment approach is currently employed to account for nonresponse. This paper presents the results of a study of the impact of selected nonresponse weighting schemes on SIPP longitudinal estimates. Emphasis is initially placed on the modeling of the nonresponse mechanism. A general framework for the weighting alternatives is presented, and empirical comparisons and evaluations of the procedures are provided, based on panel data for several of the survey's principal items.
Source: U.S. Census Bureau, Statistical Research Division
Last revised: 03-MAY-2005