Why Polls Don't Prove Anything

Polls reflect the views of the ones who come up with the questions and not much more. Polls are inherently unreliable, but that doesn't keep it from being a multi-million dollar business or from being a major policy driver.

First, a little whimsy:

[CP] First-ever polls comparing conservative and progressive activists are revealing to what degree these groups diverge when it comes to issue priorities, issue positions, and beliefs about scripture.

The surveys, the results of which were released Tuesday, also show how religious activists part ways on issue priorities.

While the vast majority of conservatives identify abortion (83 percent) and same-sex marriage (65 percent) as the most important priorities among eight issues listed in the surveys, less than 10 percent of progressive religious activists called abortion and same-sex marriage the “most important” issues.

Instead, progressive activists identify poverty (74 percent), health care (67 percent), environment (56 percent), jobs/economy (48 percent), and the Iraq war (45 percent) as the highest priorities.

This poll, touted as delineating differences between religious activists, doesn't reveal any new information. The reason the survey split the way it did has less to do with the fact of "religious activism" than it does with the fact that the respondents are conservatives and progressives (liberals); these are definitions.

Polls function as nearly self-fulfilling prophecies. Biases are baked in by the very act of formulating the question: how a question is asked (lead-in, background info), word choice, how large the sample is, who was asked, how they were asked, who did the asking, how respondents were chosen, the type of person who is willing to answer a survey, method of survey, time of day, geographic region, economic background, social background, recent events, choices presented, who paid for the survey -- all these things help to skew the results into near meaninglessness.

The best you can say is "Of the people who were willing to answer this question in this way, X number chose one of the predetermined answers -- which means nothing." Polls are nothing more than relative statistics, they cannot measure absolutes.

Me, I almost always refuse to answer polls or even poll phone calls, which means that my opinion is underrepresented. Of the few polls that I've participated in, I almost always found something to disagree with in the poll choices themselves or something that needed to be qualified, or my opinion was not one of the choices.

Everybody loves a poll, study, or survey because they promise a quick view into what's really going on, but in reality they say more about the questioner than about the answerer. Whoever gets to frame the debate gets to steer the answers.

That's what 4 out of 5 people say, anyway.