WE'VE MOVED! Democratic Convention Watch is now at http://www.DemocraticConventionWatch.com
Earlier this week, Gallup came out with two polls that seemingly contradicted one another. In the poll of "registered voters", Obama came out ahead, and in the poll of "likely voters", McCain came out ahead.
The difference between a registered voter and a likely voter is the poll respondent's answer to the question: "Are you planning on voting in November?" or some variation of that question. If someone says "YES! I wouldn't miss it. This is the most important thing I'm going to do this year", they're a likely voter. OK, that's over the top: the truth is, it's multiple choice: some variation of likely, unlikely, unsure. Sometimes the poll will then push for a lean to likely or unlikely.
Then, the pollster weights the responses in extrapolating the polls out. They give less weight to people in certain demographics who tend not to vote. Therefore, since young people vote at a lower percentage than other age groups, the pollsters often consider that they won't necessarily vote in November.
In one of the NY Times op-ed pieces yesterday, they talked about the potential affect of third party candidates on the November 2008 outcome.
The NY Times raised an interesting point yesterday. They talked in an op-ed piece about the potential effect of third party candidates on the November outcome. One of the factors that they looked at in terms of whether actual voters stick with their candidate relates to the enthusiasm of a candidate's supporters.
I believe that the single thing that will turn "registered voters" into "likely voters" is enthusiasm. I've never actually seen any research on this point, but intuitively, it makes logical sense.
The Times looked at the enthusiasm gap over the last three elections.
If you look at the table, you'll see that no matter what the poll number, the person with the higher enthusiasm "quotient" won the election over the last three cycles. So, two poll questions:
Logging you in...
del.icio.us
Yamaka · 872 weeks ago
If Obama goes with anyother VP, then the enthusiasm in the Hillary Camp will be very luke warm and, McCain will have a very strong chance of coming ahead.
It all depends on what Obama does in Aug '08. His judgment is on the line.
It 's Experience and Credibility, Stupid. Without Hillary on the ticket, Democrats will lose the WH, IMO.
Cheers.
Ed Espinoza · 872 weeks ago
In some cases, they also factor in voting history. In 2006 for example, if the voter hadn't participated in 04 or 02 according to the voterfile, they were not included in the survey, or if they were new registrants since 2004. The survey also asked if they had voted in the 2004 election, and if they said no, they were not included. But 2008 is going to be so high turnout, every indication is that they will not throw any people out of the survey. Otherwise, with lower turnout, the more questions. A congressional primary, for example, the screen gets much more complex.
They do not terminate by demographics, but they do weigh by demographics because we have some idea what % black or 65+ or whatever some place will be. And using past data, other surveys, they see how the data came in before applying weights.
Young voters are always weighted up in a survey, because they do not answer phones at a high rate, and in a presidential year, youth turnout is bigger than in other elections.
DocJess 71p · 872 weeks ago
Ed Espinoza · 872 weeks ago
That is to say, anything that is more than a month out from election day may not accurately show up in robo data (there was an infamous robo poll last year which showed Obama trailing 40 points behind Clinton in November of 2007, which wasn't accurate by any stretch - not even with a crazy margin of error). Survey USA is probably the best of the robos out there and you can see their site at www.surveyusa.com.
Nothing beats a live call, and robo polls more accurately reflect the live polls as we get closer to election day, which tells me that is more of a reflection of an electorate that becomes more engaging.
The nice thing about robos: they are significantly cheaper to field, because you don't need to pay live operators to conduct them. But in some cases that can have an affect on the resulting data.
Chad_Nielson 57p · 872 weeks ago
Peter Zenger · 872 weeks ago
It isn't really a poll, just a way to put lies about a candidate in your mind.
Joe Munoz · 872 weeks ago
Ubuwalker31 · 872 weeks ago
Tony Bittner · 872 weeks ago
DocJess 71p · 872 weeks ago
People like us will not vote against our values, and we shall not sit home, nor sit silent. But for some on the fence, a little moderation may be all the difference.
As a far, far left wing type, I wish the world were different, but sadly....
Mike in Maryland · 872 weeks ago
Look at this way -
1. Would you rather have at least 75% of the proposals you support at least put into play (introduced into Congress, and if passed, signed into law), as would be probable with Senator Obama? or
2. Get 100% of the proposals you abhor (stay in Iraq, increase government intrusion into the bedroom, legislate moral standards you don't agree with, off-shore oil drilling, etc.), and proposed legislation that is entirely for the uber-rich at the expense of the middle- and lower-income citizens, as would happen with McLame?
I'm sure almost all people who supported Senator Obama will agree with the first prospect, especially when they consider the consequences of the second.
As for FISA, would you rather have an opportunity to make sure those who ordered the invasion of citizen's privacy are held accountable, as could and probably will happen with a DoJ under an Obama Presidency, or let them go scot free, as would definitely happen under a McLame Presidency? In the long run, criminal penalties on the architects of the policy would be more effective than some fines on the telcos, who would only pass the cost onto the consumer. A ding on the quarterly earnings that they would then make up by permanent increases in prices, thus a long term, and probably permanent, gain to the profits of the telcos.
Mike
tmess2 · 872 weeks ago