Polling–maybe the worst thing ever to happen to our politics. Somewhere I came across a quote from Foucault–I think it was Foucault1–saying “truth is the product of the system of beliefs designed to produce it.” A Model T Ford was the truth of the Ford assembly line. If we think of the Ford assembly line as the physical embodiment of a set of beliefs–about efficiency, work, labor, economics, rationality, leisure–the Model T was the “truth” of this system. The assembly line was a set of ideas made real, the car was the truth, the visible evidence, this set of ideas was designed to produce.
In some religious systems the “truth” of disease is evil spirits: if you believe in evil spirits the disease proves their existence. “See? Evil spirits!” In science, a very different system of beliefs, disease proves an entirely different set of things. The “truth” of disease could also be described as a product of the system of beliefs designed to produce it.2
That same insight can apply to polls. The questions a pollster asks are just like the elaborate and carefully calibrated machines in Ford’s assembly line–they’re designed to work together to produce a truth which they have already imagined.
Elena Razlogova wrote an excellent dissertation, soon to appear as a book, on early radio and its relationship with audience. At first, radio programs solicited letters from listeners and actually read them , incorporating suggestions and corrections to into the plot. By the end of the 1930s, however, they had begun to switch to Gallup polls of listener opinion. They still solicited letters, but now they ignored them. In effect, they created the audience they wanted, and accustomed listeners to to a passive role.
I don’t mean simply deliberately misleading polls, “push polls” designed to produce disapproval of candidate A or B. Those are obvious.
But polls always work to reproduce a narrow, prexisiting set of assumptions–they are always already trying to find out what they think people always already believe. Here’s an example.
You could argue that the Model T didn’t exist without the assembly line, but “public opinion” does exist without polls. I have to disagree–“public opinion” is an invention, and idea, a product. It’s an 18th century phrase: it has to do with the invention of a distinct separation between public and private. People have opinions: aggregating them into public opinion involves narrowing the range of possible opinions, which you have already done before you start polling.
Take the wide wide field of human ideas, and reduce it to three possibilities, which your questions already show are the only possibilities you believe exist. You’ve pretty much locked the doors to the truth factory and guaranteed the Models T’s will keep appearing. And the Model T, in the case above, is “move to the right.”
Yes polling is effective–it’s extremely effective at reducing the field of possibility: it produces a feedback loop of received wisdom. We’d be better off with no polling, and politicians proposing actual solutions which we could then vote on, rather than polls designed around questions like “which of these three answers best describe your feelings about big government.” The polls can’t do anything other than reproduce what they started out believing in: machines for re-manufacturing the status quo.
Are polls effective? Of course they are. Just ask yourself if you feel like our politics works well; does it have room for innovation, or for a wide range of ideas? Or do you have three choices, none of them useful?
- lil’ help? ↩
- Yes, the objection here is that HIV for example, really is a virus. It’s effective to treat it as a virus, and persuasive; it brings good results, but there are still millions of people who believe it’s a virus and the virus is god’s curse, or karmic punishment for immorality. Two very different accounts of the “truth” of disease. And of course it’s not as if we know everything about viruses and HIV: it’s not been cured. That being said, if I get sick I go to a MD rather than a priest. ↩