|The Declarer (Floyd McWilliams' Blog)|
Monday, January 12, 2004
this Pew Foundation study:
There's no very good reason why I should believe this. What is more likely: That one-fifth of young people get their campaign news from comedy shows, or that one-fifth of young people are smart asses? It's not hard to imagine someone getting a call from a Pew poller and thinking to himself: "These political pollers are just as annoying as telemarketers. I bet it would really bug them if I told them I heard about Howard Dean from SNL."
Of course we could extrapolate from this specific criticism to the more general: Most social science studies involving polls or interviews are probably bullshit because there's no reason to believe that their subjects tell the truth. Note that the typical study will ask us to believe almost anything: that most women will be rape victims at some point in their lives, that many young people are cretins who learn about politics from television comedies. But of course we are never asked to believe that a person who is asked intrusive questions by a stranger will lie about them.
How did I come by my skepticism about polling? When I was sixteen years old our class took PSATs (which are kind of a practice SAT test), which had multiple-choice questions that were answered by filling in ovals. The class clown made his answers in the shape of a Christmas tree.
(Via Jeff Jarvis. I also heard it on the radio news while driving home.)
Update: Here's some evidence that my skepticism is justified. The judge in the Scott Peterson murder trial said that Peterson could not get a fair trial in his home town of Modesto, California. The judge cited a California State University-Stanislaus professor's study claiming that 70% of locals surveyed thought Peterson was guilty.
Students who performed the survey are now claiming that they made up results.
Slave labor: It's worth what you pay for it!