More Wave Goodbye to Landlines: How SP&R’s Methods Continue to Evolve with Changing Trends
According to Census Bureau statistics, more than a quarter of U.S. households have ditched their landline phones in recent years. This phenomenon is fueled mainly by younger adults relying exclusively on cell phones, to the point where now a record low of 71% of households reported having landlines in 2011, the most recent year figures were available. Meanwhile, cell phone ownership has reached more than 90%, up 300% more than 10 years ago.
Any researcher in the industry who specializes in telephone-based research needs to worry about this trend, and take the necessary steps to ensure that cell phone-only households are being included in probability-based sampling techniques.
At SP&R our methods are continually evolving to ensure we do not overlook or under sample harder-to-reach households, whether they involve cell phone-only households, or non-English speaking citizens. Coverage bias, or the failure to reach certain types of households that cannot otherwise be reached by traditional telephone methods, is a problem that affects what’s known as “external” validity of a research project, and ultimately calls into question the accuracy of its results.
We have taken several steps to account for coverage bias. First, all SP&R’s scientific telephone-based polling now includes a blended sample frame of both landline and cellular-only households. This is hugely important for many reasons. For instance, in a typical survey, the percent of interviews conducted with cell phone-only households will vary from as little as five to as high as twenty percent of all interviews, depending on factors like the nature of the study, the amount of cell phone records available and whether the study is a probability or non-probability-based sample. Moreover, it’s important that a higher concentration of cell phone respondents come from younger adults mainly in the 18-34 age range, given that “millennials”, or those born in the early 1980s, are the ones least likely to own a landline.
When it comes to our automated/IVR (Interactive Voice Response) software, commonly known as “robo” or “push button” type surveys, we now augment surveys conducted via our automated software with interviews conducted using “live” telephone agents from the same survey script. This is critical considering it is illegal under federal law to use automated software to contact people via cell phones. Using live agents to conduct these surveys via cell phones is therefore crucial to the accuracy of results. Due to the quick and relatively seamless nature of obtaining automated responses, there are many automated polling vendors springing up in the Harrisburg market. Some of these firms claim to be experts in polling and are tied to otherwise reputable political consulting firms, or trade associations that offer other services. However, few of them have methods sanctioned by the polling industry’s most commonly-used guidelines or best practices. At SP&R, we participate in annual gatherings of polling experts through associations like AAPOR (American Association of Public Opinion Research) and incorporate the latest polling methods into our research. Placing a great emphasis on the means by which we obtain results, we also adopt guiding principles from real task force experts and account for hard-to-reach members of a sample in all of our research.