In recent years cell phone coverage rates have increased rapidly. At the same time, more and more people have abandoned their landline telephone service. As a consequence, the proportion of the so called mobile-only population that has a cell phone only has increased. Thus, traditional RDD sampling procedures that rely on the landline telephone numbering frame only are questioned because of the potential coverage bias of samples drawn from this frame. In order to compensate for this coverage bias more and more telephone surveys make use of a dual frame sampling scheme consisting of randomly generated landline telephone numbers and mobile phone numbers.

Since a landline telephone number might reach a household with multiple members of the target population (e.g. adults) this number is subsequently weighted with the inverse of the number of persons in that particular household who belong to the target population multiplied with the number of landline telephone numbers that can be used to reach this household. By contrast, mobile phones are treated a personal devices, thus, the weight for these numbers is build on the number of mobile phone numbers of that particular person, only.

In this paper we will assess whether the underlying assumption is justi ed according to which a mobile phone is a personal device that is not shared among a group of potential respondents. Based on the rst wave of a small scale mobile phone panel we will report estimates for the proportion of mobile phone numbers that can be used to contact more than one respondent. In addition, we will provide socio-demographic background information n the sharing population in Germany.