Tracking and Interviewing Family Options Study Participants

IntroductionThis article examines the participant-tracking methods used to conduct the Family Options Study, launched by the U.S. Department of Housing and Urban Development (HUD) in 2008 to learn more about the effects of housing and sendees inten'ention for homeless families. The Family Options Study analyzes the effects of three housing and sendees interventions for a sample of 2,282 homeless families staying in emergency shelter in 12 locations across the country. The effects of the three interventions-(1) a short-term rental subsidy, (2) a long-term rental study, and (3) sendeeintensive transitional housing-were measured and compared with the effects of usual care. The study defined usual care as any housing or services that a family accesses in the absence of any other intervention. Because the study was a randomized, controlled trial that compared the study interventions with usual care (and with each other), it was very important to achieve high response rates to the followup surveys that were the main source used to measure the effects of the study interventions. The study achieved remarkably high response rates to the followup surveys-more than 80 percent for the survey conducted about 18 months after families enrolled in the study and 78 percent for the survey conducted 3 years after enrollment.Because of these high response rates, the study has been able to measure statistically significant effects not only on housing stability but in other areas as well, including family composition, adult well-being, child well-being, and self-sufficiency.Tracking study participants in a longitudinal study is difficult, because attrition is inevitable. Attrition rates of 20 to 30 percent, and even as high as 70 percent, are not uncommon (Gustavson et al., 2012; Launes et al., 2014). Some participants decide they no longer want to be part of a longitudinal study. The study more frequently cannot locate the study participants for followup surveys because they move or change their telephone numbers. The housing instability that is in the very nature of homelessness made tracking and surveying families over time particularly challenging for the Family Options Study.Low-income households, taken as a whole, tend to be more mobile than middle-income households. They move more frequently, change telephone numbers more often, or may have telephone numbers temporarily disconnected. Homeless families are even more mobile. During the followup period for the study, families in the Family Options Study were likely to relocate from the emergency shelters from which they were recruited into the study. They also were likely to move back and forth among their own housing units, temporary stays with family and friends, or returns to shelters that could be different from the shelters in which they were staying originally Further, many homeless families have experienced violence or trauma that may make them vulnerable and wary of engaging in research or efforts to contact them over time.Gustavson et al. noted that study participants who have low educational levels, are unemployed, or are not married are likely to have high attrition rates (Gustavson et al., 2012). These characteristics are all common among Family Options Study participants. At the time of enrollment, 83 percent of the study participants were unemployed; 30 percent were unmarried; and 73 percent had a high school diploma, GED (that is, general educational development), or less (Gubits et al., 2013). The Family Options Study implemented a rigorous participant-tracking strategy aimed at overcoming these challenges.Local interviewers conducted the study enrollment in person, which helped the interviewers build rapport with the families early on. After random assignment, the study team contacted the programs to which families were referred to collect information about whether and when the family enrolled in the assigned intervention. The local interviewers also maintained periodic contact directly with participating families to ensure that the contact information was as accurate as possible leading up to the followup interviews. …