Reducing nonresponse and attrition is a primary goal of the SIPP Methods Panel; a key strategy toward that goal is the design of a more efficient questionnaire. This paper describes an experiment which tests new questionnaire efficiencies for SIPP's asset ownership questions. The standard SIPP questionnaire includes a 12-question battery about different types of assets, some of which (for example, corporate bonds, royalties, mortgages held) are extremely rare - especially among those who own none of the more common asset types. Within the Methods Panel field experiment, one-half of the sample received the standard battery; the interview for the other half employed screening procedures to determine whether to ask the complete battery or not. In essence, those who reported ownership of any of the more common asset types in the initial half of the battery received the full set of questions; those who responded "no" to all of the initial questions received only a general "or any other type of financial investment" question in place of specific questions for the less common asset types. (A "yes" to this general question unleashed all of the specific questions.) We examine several outcome measures to evaluate the experiment: (1) efficiency gains; (2) interviewers' assessments; and (3) evidence regarding reporting completeness for the uncommon, not-always-asked-about-specifically asset types.
Source: U.S. Census Bureau, Statistical Research Division
Created: October 4, 2001