Implications of Moving Public Opinion Surveys to a Single-Frame Cell-Phone Random-Digit-Dial Design

Courtney Kennedy, Kyley McGeeney, Scott Keeter, Eileen Patten, Andrew J. Perrin, Amanda Lee, Jonathan Best

Public Opinion Quarterly · 2018 · 12 citations · 10 references

Concepts

Abstract

With the share of US adults living in households with cell phones climbing over 90 percent (Blumberg and Luke 2016), survey researchers are considering whether it is necessary to continue sampling landlines in addition to cell phones in RDD surveys. This transition raises a number of questions, including whether it will systematically change public opinion estimates, subgroup estimates, long-standing trends, as well as precision. To address these questions, we analyzed data from Pew Research Center national dual-frame RDD surveys conducted from 2012 to 2015. We compared the final survey estimates to those computed with an experimental weight that excluded the landline sample, thus simulating a single-frame cell-phone design. Analysis of more than 250 survey questions shows that when landlines are excluded, estimates change by less than one percentage point, on average. In addition, responding samples of adults reached via cell phone are much more demographically representative of the United States than responding samples of adults reached via landline, thus requiring less aggressive weighting and yielding lower design effects and smaller margins of error, relative to dual-frame RDD. While most estimates commonly reported from public opinion surveys seem unaffected by this design change, there are some exceptions. While considerations of cost and concerns about maximizing the population coverage rate may continue to favor dual-frame RDD for some time, these results demonstrate that by and large public opinion researchers have little to fear and potentially much to gain when landlines become a thing of the past.

References

10