[OT]BTW, how are you, LOUNE? Everything OK after the accident? [/OT]
Yeah, I’m good. I had whiplash for a few days, but I’m mostly better now. No permanent damage (I think). I was back in the office, working like a dog the next day.
Thanks for asking.
I actually don’t think that it is going to be so massive. What seems most notable about this election cycle so far is how stable it is. Very little has gone outside of statistical (and overanalyzed) noise. Look at that RCP trend line. Obama has moved in a very narrow range for more than two months. McCain has moved in almost a narrow of a line with the biggest swings being mostly due to the occasional reporting of results like the USAToday/Gallup LV poll that gave him a +5 result while everything else was reporting him as -3. Obama’s lead is not huge but looking at the whole polling picture over time its is amazingly nonvolatile.
There will be a bump, and a smaller counterbump during the GOP show, but massive? I will be shocked if you are right. (But I’ll be happily wrong.)
As to micro-analyzing swings in the electoral projections based on occasional pollings with small sample sizes - wake me up when the situation is other than that McCain still needs to win every toss-up, not lose any of his leans, and grab an Obama lean out to win. It may get there but it isn’t yet.
Hey, it serves Team Obama well for the perception to be that this is an exciting race. But the reality is that it so far is not. Least Orig you may have whiplash, but these polls won’t give the rest of us any! 
(Glad you’re okay btw.)
Well, I’m glad you’re not still a newt. 
Here’s when Obama was at 320.
The biggest changes between those is Ohio going back to McCain +2, VA going back to a tie, MO going to McCain +7 and Colorado going McCain +2.
However, you mentioned the “past couple of weeks”. Two weeks ago, we had this. From that, the major switches (aside from today’s Colorado flip) was almost entirely made up of ties going for McCain. I don’t think it’s “pretzel logic” to say that 51 EV’s worth of McCain leaning tied states going into the McCain column a week later doesn’t represent a big “loss of support”.
The only real “lost support” I see in there (switch-wise) is in Colorado where Obama had been enjoying a fairly decent lead. I’ll be interested to see where future polls place him. Non-switching states, I guess the big one for loss of support is Minnesota although I doubt we’ll see that flip (and I also doubt Indiana will go blue).
I’m not saying that it’ll be easy for Obama or even that Obama is guaranteed victory. I said in another thread (the “Why isn’t Obama leading by 15?” one) that I don’t think the current partisan divide and two independent appealing candidates will allow for anything but a pretty close race. But I still think this race is Obama’s to lose and suspect (only my opinion) that we’ll see an upswing once he gets back onto the campaign trail.
Uhhh…on July 18, Obama=325, McCain=199, Ties=14. How exactly is today’s Obama=275, McCain=250, Ties=13 (not to mention a large number of states going from dark blue to light blue) not an indication of a loss of support for Obama?
So, by “couple weeks” you meant a month? 
makes whatever noise a newt makes
Arizona the point is that those putative swings are really more statistical noise than anything else. While the source is biased, DemConWatch has a great overview of the different electoral trackers. The consensus of the trakers is Obama 260 and McCain 189 with 89 toss-up. Since June the only changes that haven’t flipped back are
Go to the most McCain leaning electoral tracker (I do not mean biased, just giving him the best numbers right now - Rasmussen - 538 is tied but is a modeller, not a tracker). Their last one is dated in July, but hey, it’s still the more McCain leaning than any other, so we’ll use it. McCain safe and likely = 165. Obama safe and likely = 210. McCain leaning = 62. Obama leaning = 63. Toss-up = 38.
Give McCain all of his safe and likely, add in all of his leanings and you get up to 227. Give him all the toss-ups, every one. You still have him short at 265. So in addition to not losing a single one of his leans, and winning all the toss-ups, he needs to pull one out of the Obama lean column, using the most McCain favorable tracker’s numbers. It can happen sure, but the essence of that being the electoral landscape has not changed. Yeah, some trackers are putting more with narrow Obama leads as Obama numbers and so when there is a small swing have to pull them out, but the Electoral-Vote.com that you cite as having lost so much still has that same dynamic - Obama by their numbers has more than enough to win even conceding the toss-up electoral votes to McCain. (Would I expect that site’s numbers to swing a lot though? Well yeah. They have Ohio in McCain’s tally and Indiana in Obama’s! That one will swing a lot with statistical noise.)
Again, the reality is that so far this race is boringly stable. Overanalyze statistical noise up and down all you want.
So I’ve noticed. Not sure my link will work but compare the 2004 summer polling to 2008’s. In 2004, Bush/Kerry preferences were up and down all over the place. This year, we quickly settled into a stable Obama lead in polling averages and haven’t really moved since.
Arizona, let’s have some more fun with overanalysis. Go to that July 18 you link to and add up the strong plus weak Dem (not the “barely”) and you get 246, same GOP and you get 178.
Go to the most recent and you get 255 Dem vs 165. An Obama improvement over that time period of 22! In the same tracker that you claim shows his support weakening so much. Which I also do not take too seriously.
Slice and dice. Barelies and tieds were 114 then and 123 now. Then Obama needed to win 24 of 114 in that group; now 15 of 123 in that group. Honestly, I’d say the current numbers are better than they were. If I’d say anything other than boringly stable.
Reverse Bradley Effect! Look out!
Seriously- my mom is in the tank for Obama, but she would not tell my dad that. I wonder if she would tell a pollster? Will cases like this balance these alleged legions of racists who will tell a pollster that they will for Obama, even though they won’t, despite a bunch of more PC reasons why they might want to vote for McCain? “Experience”, for one?
So interesting. I’ve learned a lot about the ins and outs of polling and media coverage during this campaign.
BTW jophiel, your link worked fine and is a way cool side by side. Can you do the same with other past elections?
Thanks!
The Kerry & Obama stuff was off of RCP’s site. I was actually curious enough myself to try to make one myself for Gore since RCP didn’t have one for me. I used numbers archived on PollingReport.com to create a week to week average.
A word of caution: My graph is worth exactly what you paid for it. I tried to get as much data as possible but I went week to week instead of RCP’s day to day for the sake of my own sanity. So there’s less data points actually marked. Also, RCP uses daily trackers from Gallup & Rasmussen whereas I was using polls that covered several days at a time. Finally, some weeks had considerably less data than others (1 poll at its worse for July 1st-7th, six polls at the best).
I resized the Kerry/Bush graph so that 50% is approximately 50% across the board. I suppose if I had more gumption I could have just combined all the graphs into one but that’s not likely to happen 
Graph of the last three lections and how they polled from mid-April through mid-August.
Gore really took a beating over the summer. I don’t think he really closed up until after the convention. Also, the stability of Obama’s polling is pretty unique among any of the last six candidates. If it means anything, I don’t know but it is interesting.
Thanks!
In an odd coincidence, many conservative bloggers are making hay of the (apparent; I don’t know if it’s true or not) fact that Obama’s lead over McCain is lower than Kerry’s was over Bush at this time of the year in '04. How much is this an “Obama is in big trouble” indicator?
Hey, that’s nothing compared to the Reverse Bel Air!
At this time last election, the Democratic convention was three weeks ago, and the Republican convention had not yet taken place.
I’m not going to worry about any numbers until November. Really, I’m not. I’m extremely worried about the latent racist vote, and I believe that polls are going to be extremely unreliable this year.
Which is not to say that I believe Obama will necessarily lose - there is a long time to go, and perhaps the latent racist vote will be pleasantly in line with what I expect, i.e., very small.
Today’s rumor has McCain announcing his running mate on the 29th of August.
I’ll believe this one.
- Take the focus off Obama’s speech the night before in front of 70,000 people. We know how good of a speech Obama can give.
- It is his 72nd birthday. McCain doesn’t want the second story on the Sunday news talk programs to be about him turning 72.
However, right before Labor Day weekend might not be the smartest time either. Plus, he was rumored to do it in Ohio in a 10,000 seat arena. Uh, how will that look compared to Obama filling a football stadium?
Those all sound like good reasons, but it’ll be a day that nobody is happy with the coverage. Obama supporters will say that it’s a historic day AND the “I Had A Dream” speech day. McCain people will say what you said.