Bloomberg Poll that Boosts Obama Goes Against All Other Polls

Perhaps you've seen this poll from Bloomberg that was released recently, claiming that President Obama is leading Mitt Romney...with likely voters, mind you...by 13%. Wow really, Bloomberg?

Bloomberg Poll: Obama 53%, Romney 40% (Likely Voters, supposedly) - Obama +13%

So Obama is headed for a landslide victory? I'm sure someone else will hold THAT up, right, Bloomberg?
Let's see what other polls released within one week of Bloomberg's say:

Rassmussen Reports: Romney 47%, Obama 45% (Likely Voters) - Romney +2

Gallup: Romney 46%, Obama 45% (Registered Voters) - Romney +1

Reuters: Obama 45%, Romney 44% (Registered Voters) - Obama +1

Monmouth University: Obama 47%, Romney 46% (Likely Voters) - Obama +1 *

Tipp Online: Obama 48%, Romney 44% (Registered Voters) - Obama +4 *

Fox News: Obama 43%, Romney 43% (Registered Voters) - Tie *

CNN/Opinion Research: Obama 49%, Romney 46% (Registered Voters) - Obama +3 *

* Denotes poll that is 2 or more weeks old

So let's see, the other four polls from the last week (top three polls) give an average of Romney 45.6% to Obama 45%. These threee polls range from Romney leading by 2% to Obama leading by 1%. All four are within the margin for error. When you add in the bottom 4 polls (which date as early as the first week in June) we get an average of Obama 46%, Romney 45%. Compare that to Bloomberg's ridiculous claim that Obama is winning by 13%!

How in Heaven's name is Bloomberg arriving at this conclusion? One might guess that there was some sample stacking...and yes, there was some of that.  Based on party identification (not party registration by the way) Democrats were oversampled 6% over Republicans in direct identification (sample was 32% either straight Republican or lean Republican while 38% either identified with the Democrat Party or leaned that way), and had an additional 1% more of the Democrat leaners over Republican leaners. There were also another 26% who claim they were "true independents."(Those of you who read this regularly know that I generally disregard people who trumpet their independence to the world because 95% of these people find a way to land with a leg on each side of the political fence.)

Let's set aside what I've said over and over and over again about polling samples like this, specifically noting that 40% of Americans consider themselves conservative, which means logically that at least 40% of Americans will more closely identify with the GOP (or as Bloomberg wants you to believe, 8% of conservatives don't more closely identify with the GOP) .  You also have to believe that in addition to the 21% of Americans who consider themselves liberal, another 17% of the 35% who call themselves moderate swung to the Democrat party while the rest were "true independents." (Translation, basically no moderates are Republicans).

Let's ignore all that: Why is Bloomberg's poll so far off the mark from EVERYBODY ELSE? This includes liberal polsters like Reuters/Ipsos, Tipp Online and CNN.  Even the most generous poll for the President is the CNN/Opinion Research Poll where Obama is leading Romney 49% to 46% (and that poll was published on June 1st...nearly three weeks ago) has Obama doing 4% worse than the Bloomberg poll and Romney doing 6% better than the Bloomberg poll. That's a full 10% swing in Obama's favor!

Taking the average of the other polls published in the last week or so, (remember the four poll average was Romney 45.75% to Obama 45.5%) we see a swing of +7.5% for Obama and -4.75% for Romney, meaning a swing of 12.25% aggregate in Obama's favor!

Friends, even if this poll by Bloomberg wasn't intentionally cooked (unlikely considering the fact that somehow, despite all history, moving from registered voters to likely voters the Democrat GAINS GROUND and significantly), it should be thrown out simply based upon how ridiculously out of line it is with EVERY OTHER POLL (save for the polls I've already called out for and demonstrated that they are using skewed samples).

If nothing else, friends, ignore this poll. It's findings are so far out of line with all others that there was clearly something wrong with it. And that's the bottom line, because the numbers say so.