Comment

The Summary Report of the American Association for Public Opinion Research (AAPOR) task force on nonprobability sampling is impressive. It covers a great deal of ground and reflects the evolving spirit of an organization that recognizes that it “needs to promote flexibility, not dogmatism, in determining which methods to use to achieve its objective of measuring, analyzing, and summarizing public opinion” during times when “nothing stands still, everything changes” (Newport 2011, p. 602). The observations and suggestions presented here are intended to clarify, amplify and build upon the many merits of the Summary Report'. They might also help to reduce some of the report's limitations and vulnerabilities. The breadth of experience of task force members and report reviewers stands out. It is not a roster of the usual suspects, with narrow views and limited practical experience outside of the survey research realm. Instead, members and reviewers represent many disciplines and domains, all of which value and depend upon representative samples. However, the roster includes only one person who resides outside of the United States. Given the global nature of the task force's mission “to examine the conditions under which various survey designs that do not use probability samples have scientific merit…” (p. 1), it is safe to say that the Summary Report would have benefitted from the involvement of distinguished colleagues from Great Britain, in particular, where concerns about nonprobability sampling have often been less pronounced than in the United States. As Robert Worcester observed many years ago, “All major polling organisations in Great Britain have now switched over to quota samples, whereas several had been known for random sampling in the past … the empirical results over the past decade have convinced all major pollsters that the speed of turnaround of quota samples, more or less tightly controlled, is preferred to the random methodology which requires callbacks” (Worcester 1984, p. 832).