Ipsos tracking poll (thursday): Obama "solidifies lead" over Romney

[Blue Yorker]

Paging Gallup. Obama leads Romney by 5% among Likely Voters in this poll; yet in Gallup's more Democratic friendly Registered Voters poll, we are led to believe they're tied.

Ipsos' pollster (Julia Clark) "gives Obama a 70 percent to 80 percent chance of winning".


http://www.chicagotribune.com/news/politics/sns-rt-us-usa-campaign-pollbre88i1e9-20120919,0,5438220.story

[TroyD]

1. Ipsos "gives Obama a 70 percent to 80 percent chance of winning"

Sounds like someone has been reading Nate Silver.

Nate Silver says 75% for Obama right now.

[DemocratSinceBirth]

3. That's Essentially What Intrade, the Iowa Elections Market, And Betfair Are Indicating

I would link them if I wasn't about to leave.

[TroyD]

2. Wonder if RCP will put this one in their average?

They haven't put all of the Reuters/Ipsos tracking polls in there lately for some reason.

[WI_DEM]

4. Gallup is more 'democratically friendly'?

[Blue Yorker]

6. didn't say that

I said RV sample is more Democratic-friendly than LV sample. That's regardless of polls, in almost all cases;

Yet we are led to believe that Obama performs worse in an RV poll, despite doing well in LV polls, where he should be performing worse.

[BlueStreak]

5. They all have margis for error. Statisticians play CYA big time

And remember the margin is actually twice what it seems because you have to apply the margin to BOTH numbers (Romney and Obama).

A lot of these polls sample just a few hundred people nationally. They talk a big story about how scientific it is, but a lot of it is BS. Any given poll can be influenced by the people asking the questions, when the calls were made, what shows were ON TV at the time, etc. You can't really react to one poll. When a bunch of them say the same thing then you can start to have some confidence.