The implications of alternative allocation criteria in adaptive design for panel surveys

Abstract

We compare four methods for allocating sample members to one of two data collection protocols, in an adaptive survey design framework. The methods aim to optimise, respectively, response rate, R-indicator, coefficient of variation of participation propensities, and effective sample size. The data collection protocols are CAPI single-mode and web-CAPI sequential mixed-mode. We find that two of the targeted allocations achieve a better response rate than single-mode CAPI at a lower cost. All four targeted designs out-perform both single-protocol designs in terms of representativity and effective sample size. For all but the smallest sample sizes, the adaptive designs bring modest cost savings relative to CAPI-only.