Strategies for Locating and Interviewing a Disadvantaged Population on the Moving to Opportunity Final Impact Evaluation
*Margaret L Hudson, University of Michigan, Survey Research Center
Keywords: respondent tracking, minority population, disadvantaged population, response rates
The Moving to Opportunity (MTO) for Fair Housing project was a multi-faceted study designed to help the U.S. Department of Housing and Urban Development (HUD) understand the long-term behavioral effects of families moving out of high poverty neighborhoods into less disadvantaged areas. Eligible families were recruited from high-poverty public housing projects. The 2008 Final Impact Evaluation of the MTO project included in-home interviews with adults and youth, math and English assessments with youth, physical measures and biomarker collection, neighborhood observations, and digital recording of interviews for both quality control and linguist analysis of data. Interviews were conducted with up to four persons per family.
MTO families were being contacted five or more years after their last contact with the MTO project. These families were largely non-white, were highly mobile, and many did not own a land-based telephone line. MTO 2008 was completed with an overall weighted response rate over 88%, with balanced representativeness across treatment groups and high cooperation rates for all measures included in the survey process.
This paper will discuss various strategies that were used by the Survey Research Center to locate and contact members of disadvantaged families in order to interview and conduct physical measurements and educational assessments. We will compare and contrast various methods of tracking (respondent-provided contact persons, National Change of Address, internet-based databases, and HUD databases) and discuss the overall cost investment in tracing systems versus gain in response rate. Other techniques such as family-centered sample management and tracing systems, developing cohesive teams of field interviewers supported by internet tracking specialists, and the use of adaptive design for emphasizing sample balance will be explored and discussed.