Do Cell Phones Affect Telephone Surveys?

At MBE, our clients and colleagues frequently ask about the effect so-called "cell phone only" respondents have on traditional telephone survey research. In other words, with more people deciding to use only a cell phone and giving up their land line based telephone, does the data collected through traditional telephone survey research become less reliable? This certainly is a great question and one that is receiving significant attention in the opinion research industry. The American Association of Public Opinion Research (http://www.aapor.org/) which is the primary professional association for opinion research, has devoted countless hours, conference panels, columns and other writings to this topic during the previous few years. Without a doubt, the issue will require the industry to continually monitor and address.

Some government statistics indicate slightly more than one in 10 (13%) households in the U.S. can no longer be reached via traditional land line telephones. Just six years ago approximately 3% of US households were cell-only. In fact, there are some government researchers who believe the number of households using only a cell phone will more than double during the next year. Looking at just these statistics would indicate cell-only respondents currently pose a major problem in the opinion research industry. However, a closer analysis is required for appropriate planning and development of research methodologies.

Pew Research conducted an extensive study in 2007 (the entire survey can be found at http://people-press.org/report/276/) which provides insight into this challenge. First, the Pew survey found that cell-only respondents tend to be "younger, less affluent, less likely to be married or to own their home, and more liberal on many political questions." This then begs the question do traditional telephone surveys underrepresent these demographics and if so is the total data less reliable? The Pew Survey seems to indicate, and we at MBE agree, the answer is no. Pew conducted the same set of survey questions among a cell-only audience and among a standard land line sample. When the results of cell-only sample were included with the results of the standard land line sample "overall results of the poll (changed) by no more than one percentage point on any of nine key political questions included in the study." It is reasonable to assume that just because a respondent is cell-only does not mean he or she has wildly different attitudes or opinions on most subjects compared to respondents of similar demographics who happen to have a land line, or a land line and a cell phone. One must also consider the subject matter. As the Pew survey demonstrates, cell-only respondents are younger, tend to have less education, are single and less likley to own their own home. So if one is surveying the public on issues such as home mortgages or finance, Social Security benefits, the cost of higher education or marriage related issues, special consideration should be given to sampling cell-only customers. However, if the subject matter is partisan or political most studies indicate traditional land line sampling would be sufficient or appropriate. These issues should be easily managed through discussion and planning with your opinion research firm.

All of this of course applies to current demographic profiles. Moving forward, more and more people will likely become cell-only. As the current cell-only respondents mature, will they acquire a land line and not rely on a cell phone, will they continue to be cell-only, will they use a combination? These answers are hard to predict. Nonetheless, the cell-only population is going to continue to grow and opinion researchers must adapt accordingly.

Because national telemarketing regulations prevent automated dialing assisted calls made to cell phones (such as C.A.T.I.), cell phone respondents must be dialed "by hand" whereas a land line can be dialed, randomly, by a computer and then connected with a live interviewer. This greatly increases the cost per interview of cell-only customers in addition to significantly higher sample costs for cellular telephone numbers when compared to land lines. These costs can quickly become prohibitive.

In closing, in the current environment cell-only customers do not have a major negative affect on the reliability of telephone surveys (excluding a few narrow topics or subject matters) but the issue is still worthy of discussion and strategy. Many are suggesting the creation and development of national cell-only samples or databases and even Internet panels of cell-only respondents which could be used to augment the traditional land line telephone samples. The good news is that the industry is well aware of the short-term and long-term challenges posed by this issue and is moving in a manner that will ensure the integrity and reliability of our research products. For more information, feel free to visit http://www.markblankenship.com/.