The Declarer (Floyd McWilliams' Blog)

Sunday, October 17, 2004

Mickey Kaus links to this Mark Blumenthal post on the effects of cell phones and caller ID on polling accuracy. The post drew a fair number of comments, the most interesting being from a poller:

By the way, I work for a market research companies as a telephone interviewer, and I have to say the Caller ID worry is vastly overrated and betrays a basic misunderstanding of the basic reliability of polls. Sure, the personalities of people who obsess over Caller ID might differ and with personalities might differ politics. But that begs the whole questions: even if they do answer your call, how do the personalities differ among people who are likely to agree to participate, versus those who don't. I work in the polling trenches, and I know that these personalities are markedly different. I can guess with very good accuracy whether that willingness is there from the moment they say hello, from their whole demeanour, from their use (or misuse or unuse) of politeness, and from their tone. Some of these things are related to their daily situation, but a lot is due to personality and the difference between a YES person and a GET LOST person is much greater than any you can attribute to those who screen versus those who don't.

In other words, polling itself is a suspect activity, and the more you try to pick it apart to find response bias the more you will recognise that it's bias all-the-way-down and you will finally arrive at the proper entirely skeptical position with regards to polling accuracy, as only a telephone interviewer can truly appreciate.


Regarding the DO NOT CALL list. We don't have it in Canada (where I live and work), but the majority of my research calls are in fact to the United States, and I am constantly having to fend off (dozens of times a day!) accusations that I am breaking the law by calling someone who is on this list, and when I explain that it doesn't apply to polls or research, am I almost uniformly disbelieved. So being on the Do Not Call list does have a big effect. It automatically predisposes the caller to view you as a criminal. The exception for non-soliciting calls is NOT widely understood AT ALL.



Post a Comment