Feasibility of a Single Frame Cell Phone RDD Sampling Design
Peytchev, A., & Neely, B. (2012, July). Feasibility of a Single Frame Cell Phone RDD Sampling Design. Presented at JSM 2012, San Diego, CA.
The swift adoption of cell phones in the last few years has now left some random digit dial (RDD) surveys with a substantial problem of undercoverage. A likely solution has been to adopt dual frame designs, augmenting landline samples with cell phone numbers. The telephone survey environment continues to be in a state of rapid change.
Current dual frame designs fail to sufficiently incorporate changes in the environment, data collection efficiency, and variances. There are also fundamental limitations in available sample allocation formulae. As a result, dual frame surveys tend to oversample the landline frame. We challenge the current sample optimization methods by incorporating missing components. We argue that we are approaching the point at which the optimal design for a national RDD telephone survey is a single frame cell phone design. Additional benefits include substantially reduced complexity in sampling, management, and weighting.
We extrapolate current trends in telephone use, and employ survey data to evaluate the use of a single frame cell phone RDD design. We compare estimates of bias, variance, and MSE of key survey estimates across different sampling scenarios.