hey guys, so I'm confused as to how to derive the geometric mean from a set of given values and also the odds ratio formula, can anyone help? Wait.... what is this thread about?
Incredibly well-stated. Major, you already know I don't like Dems either, but how anyone objective cannot see the national GOP is a bunch phony, empty-headed losers is beyond me. Sarah Palin, Christine O'Donnell, John Boehner, Mitch McConnell are going to solve our problems? Oh brother! We are in deep trouble.
I fully agree. I'm a fan of Obama as a leader, but the Congressional Dems are just a bunch of morons to not be taking advantage of their situation. The fact that they can't even decide whether to vote on tax cuts that the GOP is opposing is amazing to me.
Also the genertic congressional ballot seems to be trending towards Democrats <object width="450" height="323"><param name="movie" value="http://polltracker.talkingpointsmemo.com/images/TPMlinechart.swf?1284580280"></param><param name="allowFullScreen" value="false"></param><param name="allowScriptAccess" value="always"></param><param name="FlashVars" value="xml=http://polltracker.talkingpointsmemo.com/contests/us-cong-generic-ballot/Democrats-Republicans.xml&swf=http://polltracker.talkingpointsmemo.com/images/TPMlinechart.swf?1284580280&width=450&height=323&xMin=&xMax=&yMin=12&yMax=63&defaultHiddenCand=&defaultHiddenMode=Internet&defaultHiddenPollster=156:Internet,95:Internet,7:Internet,53:Internet&hideCandidate=&hideModes=&hidePollsters=&grid=false&plots=true&trend=true"></param><embed src="http://polltracker.talkingpointsmemo.com/images/TPMlinechart.swf?1284580280" FlashVars="xml=http://polltracker.talkingpointsmemo.com/contests/us-cong-generic-ballot/Democrats-Republicans.xml&swf=http://polltracker.talkingpointsmemo.com/images/TPMlinechart.swf?1284580280&width=450&height=323&xMin=&xMax=&yMin=12&yMax=63&defaultHiddenCand=&defaultHiddenMode=Internet&defaultHiddenPollster=156:Internet,95:Internet,7:Internet,53:Internet&hideCandidate=&hideModes=&hidePollsters=&grid=false&plots=true&trend=true" type="application/x-shockwave-flash" allowfullscreen="false" allowScriptAccess="always" width="450" height="323"></embed></object>
This goes back to my original question. From the Siena link: Two things to keep in mind: The Siena poll tests registered voters, not likely voters. It's the likely voter models that are putting all these races very close. But likely voter models are basically just best guesses of how turnout will shape up. If the likely voter models are assuming the entire GOP is excited and my theory is correct that it's just the Tea Party subset that's really excited, then these models all overestimate the GOP vote. Realistically, it's just a theory on my part and there's no way to know until the election happens - but that's the basis of why I'm thinking Dems have a good chance to overperform the polling.
Likely voter models are fine. I use them all the time at work. Its just a question of how refined are their models. For example, likely voters can mean a pool of registered voters that have voted in 3 of the last 4 elections. Well on the surface that would make sense but there are some problems with a model that basic. For example, an 85 year old man who voted in 3 of the last 4 elections would have a tough time getting to the polls because he's old. Even though he might have voted in 3 of the last 4 his age makes it more unlikely that he would vote this time despite past history. On the flip side, an 18 to 21 year old will have no election history whatsoever so using that model they would be considered an unlikely voter even though they might be an active college student who actually has a fairly decent chance to vote. There are other examples as well but that's just a couple of basic sampling errors that can come from poor polling practices. (and plenty of polls just flat out sample terribly) The only thing that was somewhat hard to measure was the Obama surge in 2008 because a lot of people with no election history came out to vote for the first time. But even that wasn't incredibly difficult as you had plenty of other sampling data (primary voting for example) that already hinted as to what the voting demographic was going to look like.
So I just read up on how they sampled their data and I have no clue what they were smoking. I cant even figure out what their intentions were. Its a random sample of literally random people (some are registered, some are likely voters, some arent registered and dont give a damn) I wish news organizations wouldnt give garbage like this the light of day.
Well Quinnipiac's poll that came out today was somewhat similar (Gillibrand by 6). Rasmussen released theirs last week and it had Gillibrand by 10. The Siena poll(Gillibrand by 26) seems further off then the SurveyUSA one.
You have primaries which can give you a good idea of voter enthusiasm. If Republicans are coming out in droves to vote in the primary (which they are) then you can assume they would come out a few months later for the big election.
Right - but they aren't coming out universally in the primaries. The tea partiers are coming out at a much higher pace than traditional GOPers. That's why you have so many unexpected upsets all over the place. The issue is whether pollsters are doing what you're doing - assuming that the GOP enthusiasm is universal to the entire GOP - or are accounting for what we're seeing in the GOP primaries. I'm guessing it's the former, but I don't know. Absolutely - I just believe the pollsters might be misreading this very unique election cycle and would be making incorrect assumptions as a result.
See all my previous posts about the tea party effect / enthusiasm not translating to general GOP enthusiasm. We've seen the results in all the surprise GOP results where the pollsters were completely wrong. I suspect they are making the same mistakes in the general election.
