The Benefits of an Increased Cell Phone Allocation for Dual-frame Surveys to Target Low Socioeconomic Persons
Jamie Ridenhour
RTI International
Marcus Berzofsky
RTI International
Bo Lu
Division of Biostatistics, Ohio State University
Timothy Sahr
Ohio Colleges of Medicine Government Resource Center
Research has shown that the rate of movement to majority cellphone use from partial or complete landline use is not proportional throughout subgroups within the U.S. population (see, e.g., Lu, et al., 2014). These studies have found that young persons with children, minorities, and those with income near the Federal Poverty Level have made the shift more quickly than other demographic groups. Therefore, dual-frame telephone studies such as the Ohio Medicaid Assessment Survey (OMAS), which are interested in examining low socioeconomic status (SES) subpopulations, need to increase the allocation of their sample that come from a cellphone frame. Moreover, even though data collection for cellphone samples generally costs more than landline samples, the increased rate of interviewed sample members from key subpopulations mostly negates the cost-per-complete difference. In 2012, 25% of the OMAS sample was randomly chosen from the cellphone frame. For 2015, the sample allocation was increased to 50%. This paper presents the impact of the increased OMAS sample allocation to the cellphone frame on respondent yield and data collection costs and serves as a guide for similar dual-frame survey