In the past few years, the growing number of Americans living in households without landline telephones has challenged survey researchers to develop a variety of approaches to deal with this non-coverage issue. One approach is to add interviews over the cell phone to traditional random digit dial surveys of landline telephones.¹

Adding cell phone interviews attempts to include people who cannot be interviewed via landlines, i.e. those without landline phones, particularly younger Americans. The major reason this is done is to seek as complete coverage of the population as possible. The secondary reason for these efforts (which can add substantially to the costs of surveys) is that those who are not covered may, in fact, differ from those reachable via landline on the key variables being studied. Findings from the National Health Interview Survey have provided a detailed look at the issue.² The contrary situation – where those who are interviewed via cell phone do not differ from those interviewed via landline on the key variables – would suggest that the added cost of cell phone interviews might not be necessary and the non-coverage issues could be addressed in other ways.

Both conducting these cell phone interviews and adding them to landline surveys pose a variety of operational, methodological and practical issues. These issues are important and unsolved problems remain.

Since most Americans have cell phones and most still have landline phones in their homes, there is major overlap between those one can be reached on a landline phone and on a cell phone. One approach to adding cell phone interviews has been to screen adult respondents reached via cell phone to those who have no landline phone at home, the so-called Cell Phone Only samples (CPO). Thus, in theory, the sampling frame for the CPO sample does not overlap with the sampling frame for the landline sample. Another approach is to interview all adults reached via cell phone, while still determining if they have a landline phone at home. Using this approach, some of the cell phone interviews are CPO and some are not.

The purpose of this paper is look at the implications of adding cell phone interviews to surveys in terms of the differences between the landline interviews and the cell phone interviews on a group of key parameters: Americans involvement with, use of and attitudes toward the internet. The major methodological questions raised by cell phone interviews are not addressed in this paper, for they are covered in other papers at the AAPOR conference in New Orleans in 2008. This paper’s findings do raise one possible methodological complication that may need further exploration.