Today's numbers - Zogby moves one point closer, Gallup shifts 3% toward McCain. Rasmussen stays the same.
Rasmussen Reports 10/25 - 10/27 3000 LV 2.0 51 46 Obama +5 Gallup (Traditional)* 10/25 - 10/27 2448 LV 2.0 49 47 Obama +2
Reuters/C-SPAN/Zogby 10/25 - 10/27 1202 LV 2.9 49 45 Obama +4
4 comments:
You're being very selective about the polls your citing to supposedly show the numbers shifting toward McCain. Why not mention the following polls showing that Obama is holding his lead: ABC/Post +7, Diago/Hotline +8, Gallup (Expanded) +10.
But my main qualm with your posts about polls is that they are national. National polls may be a good indicator of how the nation as a whole is going to vote, but if you look at the polls state by state, Obama is looking to take the election handily. You're a smart guy Todd - you realize this right?
Not to mention that pollsters are less likely to be able to reach a likely Obama voter on the phone.
touche- I'd like to cite the Pew Research Poll that has a 16-point Obama advantage. The fat lady isn't singing, but I think it's close.
I ignore Gallup expanded because they're trying to use new techniques to accomodate hypothetical turnout differences. I'll trust the standard approach from Gallup until it's proven wrong.
I've followed these pollsters consistently because I've always viewed them as the kings of pollsters, often even being the ones who perform some of the other people's polls.
Also, the big liberal media outlets, like Newsweek (12), ABC (7), CBS and NBC always show huge leads for Barack - and I can't think of any reason why their results should differ so.
I start with daily tracking polls - rolling 3 day averages that are less likely to show fluctuations based on anomalous results - and polls that are current (ie new results today including surveys from yesterday). Looking at ABC, for example, there's data from 5 days ago included. The fun of watching the polls is to see how the race is going today, not last week - since the polls aren't predicting the future but only indicating where things are today, but more reliably, showing how things are trending.
Finally - the other polls are less trustworthy, as they use smaller samples. This causes me to mostly ignore IBD and GWU - both of which show Barack leading by just 3%.
The states are lagging indicators. If McCain is leading in my polls on Monday by 3 points, we won't be expecting Barack to win New Hampshire by 10 points, will we? But the polls will likely still show that. The most recent New Hampshire poll on RCP right now started last Thursday and wrapped up Saturday, and the oldest was finished a week ago Sunday.
Pew is an outlier at 15% - to be automatically ignored. But if you take all the polls that have new results issued today and disregard other factors that I've outlined above, the average of those 6 is a 4.83% lead for Barack.
That's a close race, but more importantly, it's a close race that is trending toward McCain with a week to go.
Post a Comment