You Will Love This Graphic: The poll result that explains the election

[trailmonkee]

Washington Post: Washington has been a bit perplexed by President Obama’s small but persistent lead in the polls. His administration would seem to fail the “Are you better off than you were four years ago” text. And presidents who fail that test lose, right?

But perhaps that’s the wrong question. We focus on the question “Are you better off than you were four years ago” because we assume voters aren’t sophisticated enough to vote based on the right question, which is “are you better off than you would have been if the other party’s candidate had won the presidency four years ago?”

The conventional wisdom: Voters don’t do counterfactuals. “It could have been worse” is a losing message. That’s been the Romney campaign’s theory of the case, certainly, and many in the media have bought it. But perhaps we’re not giving voters enough credit.

The new Allstate/National Journal/Heartland Monitor poll tested this directly. First, they asked the standard “are you better off now than you were four years ago?” A plurality said they were not. Then they asked, “are you better off because Obama won in 2008?? A plurality said they were. Here’s the graph:



more: http://www.washingtonpost.com/blogs/ezra-klein/wp/2012/09/21/the-poll-result-that-explains-the-election/

[Drale]

1. Thats the problem with polls

They can be easily manipulated with the way you ask the question or even your body language or how your voice sounds when asking the question.

[patrice]

3. +1,000,000 Polling IS a business. Some pollers are more valid than others.

Results have to look reasonable enough to sell, but they also have to be what sells. Money doesn't buy what it doesn't want.

It is widely recognized in research circles that you can poll a sample and get results that can be characterized in relatively "strong" terms and then take that same sample and interview them with similar MORE OPEN ENDED questions and capture that data and you will see the previous poll results almost completely reversed from what they "said" before. You know this from taking polls yourself or any kind of employment testing etc., how often you have to answer something from choices that don't authentically represent who you are. There's also the strong impulse to answer in ways that are socially approved, rather than the truth. The effect of social approval can be factored "out" of the results, but the tools for doing that are based, unavoidably, upon whole other sets of assumptions.

These two different approaches to research, polling compared to interviewing, fall under the general heading of QUANTITATIVE (e.g. polling) compared to QUALITATIVE (e.g. interviewing) research methodologies.

Lots and lots of polling, on authentically representative samples, systematically varied over time and content can give more authentic results, something approaching the qualitative dimensions of the information. I think that's the principle upon which google operates upon click trails.

Which brings me to my point in bringing this up in the first place, the importance of net neutrality, because what sells (e.g. FaceBook) shapes what is available.

[Indpndnt]

2. Excellent question.

"Are you better off than you were if McCain/Palin had won the election four years ago?"

Oh, hell, yeah!