As an engineer that has taken many data samples of various things that are hardware and/or software, I've found that it is difficult to resolve differences of 5% or so accurately. I think therefore, that expecting polls to resolve 5% differences in political polls is silly. What polls can show is trends. A few of them are good enough and unbiased enough to show trends. Most of the time trends can be related to events. I didn't vote in this poll, no point in it.
Depends on the poll. Loaded/fixed selections are the most obvious, then there's the selection of people who are contacted. I can't imagine it would be too damn difficult to primarily target a specific demographic who is most likely to vote the way one wants in order to get their numbers up. For all we know, there's probably a room somewhere in Virginia full of chimps trained to write random numbers on dry erase boards, and whatever they come up with is that weeks poll results.
Last edited by Surtr; 09-18-12 at 12:16 AM.
I love the NSA. It's like having a secret fan-base you will never see, but they're there, watching everything you write and it makes me feel all warm and fuzzy inside knowing that I may be some person's only form of unconstitutional entertainment one night.
In most cases I say no. In some cases I wonder if they are while trying the give the pollster the benefit of the doubt. Its really odd, for example when one pollster out of numerous ones consistently give one candidate higher numbers. Or when a website that gives the averages of all polls waits about a week to include one poll from a hugely significant media outlet seemingly waiting for other polls to drop a little so that their average doesn't bring up one side's numbers too much.
I also think if its happening, they're only hurting their own credibility when the election occurs and setting both themselves and the audience their trying to give false hope for disappointment.
Now that I'm thinking of it I recall there was some deliberate shenanigans going on on election day 2000 in one state. A dishonest activist pollster reported to the media that one side was doing better than expected in the exit polls in that state. Then when that candidate lost that state big time later that night, they had egg all over their faces and were very disappointed.
Having opinions all over the map is a good sign of a person capable of autonomous thinking. Felix -2011
I find it best to just ignore polls as they are always inaccurate.
We went from a historic shift in turnout and voting demographics in 2008 to a complete turnaround in 2010. The fact is, nobody knows what the demographics and turnout are going to look like this year. The various polling companies are making their best guesses in creating samples, and yes the liberal ones are probably estimating a little liberal, and conservative ones conservative. As always, the most accurate polls are probably the ones that aggregate those from several different companies.
But I don't think any of them are trying to push an agenda. A polling company that is inaccurate doesn't seem likely to last very long. I don't think polls effect momentum as much as they simply start a conversation going. But in presidential campaigns, the conversation is already going, and the stakes are high. I don't think presidential polling has much of an effect on election outcomes.
Last edited by Camer☑n; 09-22-12 at 10:49 AM.
(avatar by Thomas Nast)
This poll is most definitely skewed.
I think far more problematic is the electoral college and the media that reports the election over the minute the polls close. Takes the fun and legitimacy right out of the whole thing.