Thursday, June 21, 2012

Why Polls Are Silly

They just released results from a Gallup poll questioning whether Americans would vote for a Mormon. Blah blah blah....and then this (as reported by CNN):
"Bias against a Mormon candidate is significantly higher among Democrats and independents than among Republicans, Gallup found.

Twenty-four percent of Democrats and 18% of independents said they would not vote for a well-qualified Mormon who was nominated by their party, while 10% of Republicans expressed such opposition."
Gallup apparently figured the phrase "...who was nominated by their party" would cut, like a scalpel, through Democrats' antipathy toward the current Republican candidate, who happens to be Mormon. Just insert those five magical words, and people will precisely clear their cognitive slates and offer crisp, meaningful data for precisely the case at hand. And we can, therefore, draw a clean conclusion that Democrats have this ever-so-puzzling prejudice against Mormons. Fascinatingly counter-intuitive! And the press, of course, passes this on without question.

Human cognitive slates are foggy at best. Our emotions and outlooks defy even our own conscious understanding. Even the language with which our emotions and outlooks are probed seethes with unconscious associations. We're not clean there, under the hood. So the conceit of rendering that primordial goo as precise data is, in the end, laughable.

Hmm. Am I being anti-science?

2 comments:

r. clayton said...

Dude, can you parse this out for me, because I'm not getting it.

I read the poll as saying 24% of Democrats would not support a qualified, nominated Democratic Mormon, while 10% of Republicans would not support a similarly credentialed Republican Mormon (I'll skip over what "the Independents and their party" might mean).

I agree with you if you're saying "bullshit polls are bullshit," and even if you're saying "ugh, another instance of trendy counter-intuitiveness" (on the other hand: data). What I'm not picking up on is the also trendy cognitive-affects reference. The poll seems to deal with that problem by following the high-school science injunction to change one thing at a time in an experiment.

You're correct - trivially, banally correct - to point out that cognitive affects are at play. But the interesting questions are Does the poll do anything to address these affects? and If so, how effective is the poll in doing so? As far as I can tell, you don't take up either of these questions.

Jim Leff said...

"The poll seems to deal with that problem by following the high-school science injunction to change one thing at a time in an experiment"

Yes, exactly. But my point is that human sentiment, unlike, say, copper wires or argon gas, does not lend itself to clean experimentation. People don't reset on command.

E.g.: Aside from the pale fellow aggressively plunging his fangs into your neck right this moment, how do you feel about vampires, generally?

Blog Archive