Computing “e” in Self-Administered Address-Based Sampling Studies

In recent years, address-based sampling (ABS) has become a widely used approach for obtaining samples for household surveys in the United States. In computing response rates for self-administered ABS studies, one challenge is in estimating the eligibility rate of sampled cases with unknown eligibility, denoted e in American Association for Public Opinion Research response rate calculations RR3 and RR4. The assumptions underlying standard approaches for estimating e are unlikely to hold in self-administered ABS studies. This note describes an approach for estimating e in such situations.