Dealing with under-coverage bias via Dual/Multiple Frame designs: a simulation study for telephone surveys

Abstract: Multiple-frame surveys are commonly used to deal with under-coverage bias. The use of more than one frame introduces the possibility that frames overlap leading to increased inclusion probabilities of units that appear in multiple lists. Following the guide example of a dual frame set-up (DF) in telephone surveys, this contribution presents an extensive simulation study where different types of screenings to deal with the overlap issue are compared with the proper DF approach. We empirically show that the efforts of screening do not guarantee estimators more efficient than the DF estimators that do not need screening. Moreover simulation results show that screening at sample level does not correct for the increased inclusion probability of units in both frames produced by the DF set-up nor improve efficiency. The different estimation options will be compared with regards to survey costs, amount of information required and statistical properties of the final estimates