Use of Propensity Weights and Categories in a Two-Stage Survey

• Activity Number: 685
• Type: Contributed
• Date/Time: Thursday, August 5, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Pedro Saavedra+ and Martin Wulfe*+ and Daniel Geller and Lee Harding and Sheku Kamara and Ted Macaluso
• Companies: ICF Macro and ICF Macro and ICF Macro and ICF Macro and U.S. Department of Agriculture and U.S. Department of Agriculture
• Keywords: Improper payments ; cluster sampling ; nonresponse adjustments ; weighting

Abstract

In a survey of WIC participants, a telephone survey was conducted and fifty percent of the respondents were asked to participate in an in-home survey. A number of respondents to the telephone survey declined to participate in the in-home survey. The presence of a number of variables in the telephone survey lends itself to weight adjustments using either propensity weights or propensity categories. Variables associated with the in-home survey variables were used to predict responding to the survey. Two methods of adjustment were compared with results not adjusting for nonresponse. In addition, a variable from the telephone survey was used to simulate an in-home response so as to compare estimates using the two procedures with actual estimates from the telephone survey.