The Perils of Polls

Survey questions themselves may affect behavior:

Simply asking college students who are inclined to take drugs about their illegal-drug use in a survey may increase the behavior, according to a study that’s making researchers understandably nervous.
“We ask people questions, and that does change behavior,” study co-author Gavan Fitzsimons, a marketing professor at Duke University’s Fuqua School of Business, said Thursday. The provocative effect, he added, can be “much greater than most of us would like to believe.”

Read the rest, it is quite interesting. My first thought – can frequent polling during the election year, using, of course, conservative frames, influence the outcome of the election?

4 responses to “The Perils of Polls

  1. Have you ever seen, at a sporting event after a player has done something good: hit a homerun, score a touchdown, make a spectacular dunk, someone in the crowd wearing a facsimile jersey with that player’s name on it, standing up pointing to his jersey as if to say “see, I am good because I have associated myself with a winner”
    I think that polls have the same effect. Many people who vote have no idea of the actual issues involved or the positions of the candidates on those issues. They will vote for whomever the latest poll indicates is leading because they want to be able to say “I voted for the winner”.
    I think that all political polling should be illegal just for that reason.

  2. “My first thought – can frequent polling during the election year, using, of course, conservative frames, influence the outcome of the election?”
    I believe that was one of the succsesful tactics used against John McCain, the question being about his having a black child using a push poll.

  3. Hamsterbaffle

    It’s like Heisenberg’s Social Uncertainty Principle. The act of measuring the subject’s opinion will affect the subject’s opinion.
    This is one of those findings that doesn’t surprise me in the least, because I’ve been thinking this for years. Another possibility: certain opinion polls may ask questions about which many people actually have *no opinion.* But the act of taking the poll causes them to “discover” their opinion on an issue they otherwise would never have an active opinion on.
    Then, the publication and promotion of the results of the poll have a ripple effect, bearing the mark of “statistic” which may influence even more people, who “discover” their own opinion after hearing about the results of the poll.
    But what can you do? I know of no way to survey people without their knowedge.

  4. This study doesn’t suprise me in the least. I’m all for evidence, but parents and teachers have understood this phenomenon for a long time.
    That said, there’s an important distinction to point out here. Some crazy person is going to say that this study proves that non-abstinence only sex education classes leads to more kids have more sex, so let’s deal with this now. Teenagers are inclined to want to have sex because – well they’re teenagers. But a survey asking likely users about their drug habits is completely different than a class, taught responsibly, that teaches students about their bodies, their choices and their responsibilities regarding sex. Seems obvious, and it is, but you just know somebody is going to misuse this study’s results.