Recent Solutions from SSRS:

The rapid growth of cell-phone only households in America poses a challenge to survey research. Randomly dialing landlines can no longer assure a representative sample of the population, since 1 in 5 households cannot be reached via landline. Over the last two years, SSRS has gained expertise in two effective methods of addressing the cell phone challenge: Address-Based Sampling (ABS) and Cell Phone Sampling.

Addressed-Based Sampling

Cell-phone only respondents can be reached efficiently and cost effectively through the use of an address-based sample (ABS). Address-based samples are created using the Delivery Sequence File (DSF), generated by the post office, as the universe of households. Since this list is based on people's addresses and not phone numbers, it includes cell-phone only households, landline households, households with both cell phones and landlines, and those with no telephones.

All sampled addresses are run against databases of listed phone numbers to maximize the ability to reach respondents by telephone. Those without a landline telephone number receive invitations to participate in the survey through one or more modes: (1) calling an 1-800 number and completing the survey with a telephone interviewer, (2) completing an on-line questionnaire, or (3) completing a mail questionnaire. The mode of interview is dependent on the particular project needs.

SSRS's experience using ABS in the 2008 Massachusetts Health Insurance Survey has shown it is possible to use ABS to reach cell-phone only households in proportions that closely match the National Institutes of Health estimates for percent of the population that has only wireless telephone service.

ABS can be used as a single sampling frame or in dual-frame designs that include an RDD sample component.

Cell Phone Sampling

SSRS has been dialing cell phone samples for the last two years. The majority of our most recent custom surveys employed a dual-frame design, coupling standard RDD surveys with cell phone frames. Through the use of these designs, we have interviewed significant numbers of cell-phone only respondents who would otherwise have been overlooked. In some of our surveys we screen specifically within the cell phone sample for those who report having no landline phones in their house; in others we interview a large proportion on the cell phone frame, which naturally leads to the inclusion of cell-phone only respondents. As always, the details of the design are decided in consultation with our clients, to meet their expectations and goals.

Equally important as reaching this population is ensuring the proper representation of respondents through weighting and data analysis. In dual-frame studies (including both cell phones and landlines), SSRS adjusts the data to represent the share of the population that is cell phone only, landline only or dual-frame users. The weighting design employed by SSRS reduces the bias caused by missing cell phone only households.