Tuesday, April 15, 2003

Poll Pall.... The polls Every day we hear about polls--depending on how much they agree with the outcome, people are much more likely to quote poll numbers as truth. There are many ways to criticize a particular poll--perhaps most relevant is the notion that questions lead to their own answers: how you phrase the question will lead to desired results. This is all the more true when the answers are multiple choice. Take the recent extreme example that Atrios found: What would you do if an anti-war protester stopped you on the way home in rush hour? Introduce them to the grill of my car! Ignore them. Join them. So, it's worth considering some of the other limiting factors in polling that Bob Morris of Politics in the Zeros hits upon. I reported on this a while back, and it's worth repeating. Polls are generally done by telephone. Polssters get lots of answering machines, and hangups. They never call cell phones, only land line phones, and they don't call unlisted numbers. So, you can see, wide swaths of people get filtered out from pollster phone calls. There's lots of people pollsters never speak to, such as people who 1) aren't home much. 2) use a cell phone as their main phone. 3) let the machine get it. 4) check the caller id before answering. 5) have an unlisted nunber. 6) don't immediately hang up when getting an unsolicited phone call.   In short, among other groups, they've virtually eliminated from their polls, the highly mobile urban population, who might be expected to be more liberal. No, it's not a conspiracy, just changing technological times, with the pollsters quite aware of the current shortcomings of their methods - and not sure what to do about it. [Politics in the Zeros] 12:48:12 AM    comment []

© Copyright 2003 nick ring. Last update: 5/7/03; 11:37:45 PM. This theme is based on the SoundWaves (blue) Manila theme.