A recent report from the American Association for Public Opinion Research (AAPOR) on the future of telephone interviewing in the US considers what is likely to happen to telephone sampling in the next decade and beyond. It concludes that telephone will decrease in usage as a proportion of all research, but that it will continue to be a valuable methodology, often used as one element in a multi-mode approach.
The task force who authored the report includes research leaders from academia, large research companies like Nielsen, and GfK, and organizations such as the Pew Research Center.
The group began its work in 2014 and published its findings in April this year. The full report includes results from two new surveys designed specifically for this study and six appendices covering topics including sampling, weighting, non-response, costs and legal issues.
The report’s main findings include:
- While many phone surveys today are dual-frame (that is, they use a combination of RDD landline and RDD cell phone numbers to make up a probability sample), in the future using just cell phone numbers will only be sufficient to effectively cover the US population
- Although non-probability online sample is the predominant data collection method today, an RDD cell phone frame will continue to give researchers broad coverage of the US population for many years to come
- Phone surveys are effective as part of mixed-mode methodologies
- Interviewers can cause variance in the data, especially when asking about sensitive topics, but today they are more successful than machines at getting complete data from respondents on the phone. It’s important for researchers to understand levels of variability in interviewer performance and results
- When conducting interviews by cell phone, environmental and audio quality issues can cause differences in the data; this is another area for researchers to continue to pay attention to and understand
- The cell phone frame provides good coverage of the population and will benefit from the future ability to append data to help with weighting, controlling bias and with analytics
- RDD cell phone samples will continue to have the advantage of being probability samples, thus allowing sampling error to be measured, unlike non-probability methods
- Telephone can be an effective recruitment and screening tool, preceding someone taking a survey on a mobile device
- Human interviewers can intervene during a survey, to encourage, clarify or probe for more information more effectively than machines can do today
The task force concludes that over time, fewer and fewer surveys will be done using only phone for sampling, recruiting and data collection. Increasingly surveys will use telephone alongside other methodologies. In the future “proportionally fewer surveys will use a telephone frame for coverage and sampling purposes,” the report concludes.