436 – Adaptive/Innovative Survey Design and Survey Cost
Implementing Adaptive Design on a Longitudinal Survey of At-Risk Youth: Empirical Evidence Based on a Deep-Dive Analysis
Hanzhi Zhou
Mathematica Policy Research
Jillian Stein
Mathematica Policy Research
A number of challenges are associated with conducting longitudinal survey research with at-risk young people who are often highly mobile and difficult to engage. This paper describes our use of statistical tools for dynamic monitoring of data quality and our assessment of innovative strategies for increasing sample retention and survey completion among a sample of at-risk young adults. Drawing on multiple data sources-including baseline information and paradata-we calculate quality indicators and construct prediction models to (1) assess the representativeness of our data, (2) identify any over- or under-represented groups, (3) investigate the efficacy of our engagement and retention strategies overall and specifically for those under-represented groups, and (4) adapt our data collection efforts to maximize the representativeness of our data. The findings from this research add to the knowledge base regarding the use of alternative measures of quality in survey practice and the efficacy of using texting and social media, as tools for retaining and engaging sample members in longitudinal research.