Earlier today, TPM's Paul Kiel asked Frank Luntz about a participant in his most recent post-debate focus group for Fox News that had also appeared in prior televised focus group:
[W]hen I spoke to Luntz today, he said that he uses repeat participants by design. In a segment to air on Fox News tonight, he said, there should be a "bunch of people" who had been in prior focus groups, some of them participating as early as May of last year. "It allows me to see how people's opinion have changed over time," he explained. "I'm trying to isolate that moment that made the difference."
Before Kiel spoke to Luntz, he called me for a reaction. My own view, quoted in Kiel's article, is that the presence of repeat participants is a troubling indicator about of recruitment quality. But then, I had never heard of this particular application of what Luntz goes on to describe as "anthropology." Draw your own conclusions, I suppose, but the survey researchers I know try to avoid repeat focus group participants.
I wrote a brief introduction to political focus groups, their strengths and weaknesses, back in 2006. Then as now, I also highly recommend this story that aired on NPR's Marketplace in March 2002 on what can go wrong in focus group recruitment (the focus group story begins at about 4:10).
Update - Reader Mike G. leaves this helpful comment:
As someone who worked in market research for four years, I can tell you that it is viable to use the same group. This is what's known as a longitudinal method. However, Frank Luntz should have been up front about this at the beginning of the segment. The entire segment was presented like these voters had not been interviewed previously. This is intellectually dishonest and, in polling, such intellectual dishonesty has no place.