Poll: Obama holds slight edge over Romney in swing states

[technical support]

Poll: Obama holds slight edge over Romney in swing states   July 8th,2012



Posted by
CNN's Ashley Killough

(CNN) – A new poll released Sunday indicates the presidential race between Mitt Romney and President Barack Obama is still tight in a dozen major battlegrounds.

According to the latest survey by USA Today/Gallup, the president has a small advantage over his Republican rival, 47% to 45%, in 12 key states. The two-point margin falls well within the poll's sampling error.

– Follow the Ticker on Twitter: @PoliticalTicker

– Check out the CNN Electoral Map and Calculator and game out your own strategy for November.

Obama's edge over Romney represents the same two-point margin he held when the same poll was last conducted in May.

When zooming out to include voters from the other states, the gap expands to a four-point margin in favor of Obama, 48% to 44%.

However, political observers and both campaigns say the race will ultimately be decided by a select portion of swing states.

Gallup surveyed voters in Colorado, Florida, Iowa, New Hampshire, Ohio, Nevada and Virginia - all states that CNN rates as toss-ups on its Electoral Map.

The poll also included voters in Michigan, Wisconsin, Pennsylvania, and New Mexico - four states CNN considers leaning toward Obama - as well as voters in North Carolina, a state leaning toward Romney, according to a CNN analysis.

Also of note, the poll shows the barrage of attack ads on swing-state airwaves has caused a slight change in voters' minds.

One in 12 say the commercials have altered their opinion about Obama or Romney, with 76% of them now favoring the president compared to 16% who say they switched to Romney.

Both campaigns have taken out state-specific commercials in such states as Iowa, Virginia and Ohio.

For the USA TODAY survey, Gallup questioned 1,200 registered voters in the dozen swing states by telephone June 22-29, with a sampling error of plus or minus four percentage points. For non-swing states, 2,404 registered voters were interviewed during the same time period.

[timothyinMD]

Obama won the 12 states polled 53.2 to 45.5

Obama's way behind his 2008 showing

[True Federalist (진정한 연방 주의자)]

Obama won the 12 states polled 53.2 to 45.5

Obama's way behind his 2008 showing

With these numbers Obama still wins on uniform swing with either 272 or 285 EV depending on which way Virginia falls.  Romney needs to either increase the swing more than it currently is or hope that Iowa's current swing in the polls by much more than a uniform swing would call for is kept while Virginia' current less than average swing increases.  If the election were held today, Obama wins.  However, Romney still has four months to campaign and that's a lot of time and money, so he still has a reasonable shot at winning.

[pepper11]

Obama won the 12 states polled 53.2 to 45.5

Obama's way behind his 2008 showing

With these numbers Obama still wins on uniform swing with either 272 or 285 EV depending on which way Virginia falls.  Romney needs to either increase the swing more than it currently is or hope that Iowa's current swing in the polls by much more than a uniform swing would call for is kept while Virginia' current less than average swing increases.  If the election were held today, Obama wins.  However, Romney still has four months to campaign and that's a lot of time and money, so he still has a reasonable shot at winning.

Actually, considering these are registered voters this poll is a dead heat, if not a slight Romen lead with a reasonable likely voter model. That being said, I still think Romney is down in Virginia. Who knows, Iowa + Wisco makes up for Virginia.

[xavier110]

One in 12 say the commercials have altered their opinion about Obama or Romney, with 76% of them now favoring the president compared to 16% who say they switched to Romney.

The only interesting bit of data from this poll.

[krazen1211]

The GOP has taken a massive 21,000 vote registration edge in Iowa. Compare that to the 100.000 edge the Democrats had in 2008, or the 4000 edge the GOP had in 2004.

It's in the bag.

[Miles]

The GOP has taken a massive 21,000 vote registration edge in Iowa. Compare that to the 100.000 edge the Democrats had in 2008, or the 4000 edge the GOP had in 2004.

It's in the bag.

Isn't it a little late for you to be out trolling?

[krazen1211]

The GOP has taken a massive 21,000 vote registration edge in Iowa. Compare that to the 100.000 edge the Democrats had in 2008, or the 4000 edge the GOP had in 2004.

It's in the bag.

Isn't it a little late for you to be out trolling?

When did you become so rude?

http://blogs.desmoinesregister.com/dmr/index.php/2012/07/05/republicans-bolster-lead-over-democrats-in-iowa-voter-registrations/

[Miles]

The GOP has taken a massive 21,000 vote registration edge in Iowa. Compare that to the 100.000 edge the Democrats had in 2008, or the 4000 edge the GOP had in 2004.

It's in the bag.

Isn't it a little late for you to be out trolling?

When did you become so rude?

http://blogs.desmoinesregister.com/dmr/index.php/2012/07/05/republicans-bolster-lead-over-democrats-in-iowa-voter-registrations/

I'm just saying what many other people on here are thinking.

[Fmr President & Senator Polnut]

The GOP also had a hotly contested caucus in IA... the Dems didn't...

But on topic, I hate multi-state polls... and I tend to ignore them. 

[pbrower2a]

Obama won the 12 states polled 53.2 to 45.5

Obama's way behind his 2008 showing

With these numbers Obama still wins on uniform swing with either 272 or 285 EV depending on which way Virginia falls.  Romney needs to either increase the swing more than it currently is or hope that Iowa's current swing in the polls by much more than a uniform swing would call for is kept while Virginia' current less than average swing increases.  If the election were held today, Obama wins.  However, Romney still has four months to campaign and that's a lot of time and money, so he still has a reasonable shot at winning.

Actually, considering these are registered voters this poll is a dead heat, if not a slight Romen lead with a reasonable likely voter model. That being said, I still think Romney is down in Virginia. Who knows, Iowa + Wisco makes up for Virginia.

The Republicans win the general election (including the Presidency) with a "likely voter" model that shows an electorate similar to that of 2010. Democrats at all levels will seek to expand the electorate as much as possible. Expect Democrats to do everything possible to register new, D-leaning voters and encourage them to vote. Expect Republican officials to make voting as difficult as possible.

The GOP has taken a massive 21,000 vote registration edge in Iowa. Compare that to the 100.000 edge the Democrats had in 2008, or the 4000 edge the GOP had in 2004.

It's in the bag.

Democratic registration was way up in Iowa in 2008 because of the contested Democratic primary, and Republican registration in 2012 is way up because of the heavily-contested Republican primary.  Fall college terms have yet to start, and if the Obama campaign has any uncertainty about Iowa it is going to double down on registering new voters and cajoling them to vote.

It's who has the edge at the end of registration that counts.

One in 12 say the commercials have altered their opinion about Obama or Romney, with 76% of them now favoring the president compared to 16% who say they switched to Romney.

The only interesting bit of data from this poll.

 

...and that is itself a portent of an R disaster. It's the marginal voter that counts most in a close election.

[pepper11]

One in 12 say the commercials have altered their opinion about Obama or Romney, with 76% of them now favoring the president compared to 16% who say they switched to Romney.

The only interesting bit of data from this poll.

Except that this data point does not make any sense. It is total junk.

Obama won the 12 states polled 53.2 to 45.5, Obama won all 50 states 52.9 to 45.6, essentially an identical margin.  The poll gives Obama a 49-45 lead overall and a 47-45 lead in swing states (which appears to be a subset of the general poll). If 8% of the poll respondants were influenced by ads (as the poll states), and 76% of these voters chose to Obama to 16% Romney, this should expand Obamas lead by 4.8%. If we assume only swing states are seeing the ads (not 100% true) and if we assume Obama should be as strong in non swing states as in swing states (as seen in 2008 results), Obama should be up by about 9% in the swing state subset group. Yet somehow this poll gives Romney some swing state magic and only puts him down by 2 in their swing state subset.  The math doesn't add up, its not even close. Considering this and the fact that the swing state margin narrowed from Obama + 4 to Obama + 2 during the last month when Obama's ads have been running, its pretty tough to make the case the case the ads are working. Despite the deceptively dramatic headline by Gallup, these TV ads are confirming existing positions, not creating new ones. Plus grouped state polls are complete rubbish anyway.

[krazen1211]

The GOP also had a hotly contested caucus in IA... the Dems didn't...

But on topic, I hate multi-state polls... and I tend to ignore them. 

Well, that's a nice theory, but the GOP gains mostly occurred in the past 3 months.

[Brittain33]

The GOP also had a hotly contested caucus in IA... the Dems didn't...

But on topic, I hate multi-state polls... and I tend to ignore them. 

Well, that's a nice theory, but the GOP gains mostly occurred in the past 3 months.

I don't see how that follows when you have a swing of 120,000 votes since the Dem peak but the Republicans only gained 20,000 since "early spring."

[krazen1211]

The GOP also had a hotly contested caucus in IA... the Dems didn't...

But on topic, I hate multi-state polls... and I tend to ignore them. 

Well, that's a nice theory, but the GOP gains mostly occurred in the past 3 months.

I don't see how that follows when you have a swing of 120,000 votes since the Dem peak but the Republicans only gained 20,000 since "early spring."

Ah, I apologize. Iowa voter registration didn't really change in the months prior to the Caucuses this January, so it is quite peculiar to point to the caucuses as a reason for the registration growth in the GOP.

The Democrats loss of their 120,000 vote edge mostly occurred in earlier years. Here are rough estimates from the SoS of Iowa:

September 2011 - 608k GOP 643K Dem
January 2012 - 613k GOP 641K Dem
Current: 616K GOP 596K Dem

As you can see, registration did not substantially change in the months prior to the caucuses! Registration was tied in March 2012, and the actual 21k lead was built primarily in the past 3 months. Since May 2012 the GOP has added  2k/1k/3k/4k in the 4 congressional districts, respectively.

[krazen1211]

Obama won the 12 states polled 53.2 to 45.5

Obama's way behind his 2008 showing

With these numbers Obama still wins on uniform swing with either 272 or 285 EV depending on which way Virginia falls.  Romney needs to either increase the swing more than it currently is or hope that Iowa's current swing in the polls by much more than a uniform swing would call for is kept while Virginia' current less than average swing increases.  If the election were held today, Obama wins.  However, Romney still has four months to campaign and that's a lot of time and money, so he still has a reasonable shot at winning.

Actually, considering these are registered voters this poll is a dead heat, if not a slight Romen lead with a reasonable likely voter model. That being said, I still think Romney is down in Virginia. Who knows, Iowa + Wisco makes up for Virginia.

The Republicans win the general election (including the Presidency) with a "likely voter" model that shows an electorate similar to that of 2010. Democrats at all levels will seek to expand the electorate as much as possible. Expect Democrats to do everything possible to register new, D-leaning voters and encourage them to vote. Expect Republican officials to make voting as difficult as possible.

The GOP has taken a massive 21,000 vote registration edge in Iowa. Compare that to the 100.000 edge the Democrats had in 2008, or the 4000 edge the GOP had in 2004.

It's in the bag.

Democratic registration was way up in Iowa in 2008 because of the contested Democratic primary, and Republican registration in 2012 is way up because of the heavily-contested Republican primary.  Fall college terms have yet to start, and if the Obama campaign has any uncertainty about Iowa it is going to double down on registering new voters and cajoling them to vote.

It's who has the edge at the end of registration that counts.

One in 12 say the commercials have altered their opinion about Obama or Romney, with 76% of them now favoring the president compared to 16% who say they switched to Romney.

The only interesting bit of data from this poll.

 

...and that is itself a portent of an R disaster. It's the marginal voter that counts most in a close election.

Certainly plausible of course, but it would not fit the pattern of what happened the last 2 election cycles.

In both 2004 and 2008, Iowa Democrats gained heavily in registration between January and June of that year. The parties ran much closer to even between June and November of that year.

[pbrower2a]

Obama won the 12 states polled 53.2 to 45.5

Obama's way behind his 2008 showing

With these numbers Obama still wins on uniform swing with either 272 or 285 EV depending on which way Virginia falls.  Romney needs to either increase the swing more than it currently is or hope that Iowa's current swing in the polls by much more than a uniform swing would call for is kept while Virginia' current less than average swing increases.  If the election were held today, Obama wins.  However, Romney still has four months to campaign and that's a lot of time and money, so he still has a reasonable shot at winning.

Actually, considering these are registered voters this poll is a dead heat, if not a slight Romen lead with a reasonable likely voter model. That being said, I still think Romney is down in Virginia. Who knows, Iowa + Wisco makes up for Virginia.

The Republicans win the general election (including the Presidency) with a "likely voter" model that shows an electorate similar to that of 2010. Democrats at all levels will seek to expand the electorate as much as possible. Expect Democrats to do everything possible to register new, D-leaning voters and encourage them to vote. Expect Republican officials to make voting as difficult as possible.

The GOP has taken a massive 21,000 vote registration edge in Iowa. Compare that to the 100.000 edge the Democrats had in 2008, or the 4000 edge the GOP had in 2004.

It's in the bag.

Democratic registration was way up in Iowa in 2008 because of the contested Democratic primary, and Republican registration in 2012 is way up because of the heavily-contested Republican primary.  Fall college terms have yet to start, and if the Obama campaign has any uncertainty about Iowa it is going to double down on registering new voters and cajoling them to vote.

It's who has the edge at the end of registration that counts.

One in 12 say the commercials have altered their opinion about Obama or Romney, with 76% of them now favoring the president compared to 16% who say they switched to Romney.

The only interesting bit of data from this poll.

 

...and that is itself a portent of an R disaster. It's the marginal voter that counts most in a close election.

Certainly plausible of course, but it would not fit the pattern of what happened the last 2 election cycles.

In both 2004 and 2008, Iowa Democrats gained heavily in registration between January and June of that year. The parties ran much closer to even between June and November of that year.

The trend from 2008 to 2010 in politics is unlikely to continue. Anyone who predicts a continuation of a trend beyond the very short term is a fool. If I were to extrapolate the Presidential result from 2004 to 2008 in Iowa  I would expect President Obama to win the state by 20%, which I just don't see happening any more than you do -- or common sense.

The Republican Party did very well in 2010 with a reduced electorate, and with a not-so-reduced electorate it won't do so well. Such is the difference between a Likely Voters model that is the reality in off-year and midterm elections. (2006 was a possible fluke because of the extreme disdain for the then-current President, something that won't be repeated this year).

You can trust that the Obama campaign will do everything possible to prevent the "Likely Voters" model from being the one that operates in 2012.   

[krazen1211]

Obama won the 12 states polled 53.2 to 45.5

Obama's way behind his 2008 showing

With these numbers Obama still wins on uniform swing with either 272 or 285 EV depending on which way Virginia falls.  Romney needs to either increase the swing more than it currently is or hope that Iowa's current swing in the polls by much more than a uniform swing would call for is kept while Virginia' current less than average swing increases.  If the election were held today, Obama wins.  However, Romney still has four months to campaign and that's a lot of time and money, so he still has a reasonable shot at winning.

Actually, considering these are registered voters this poll is a dead heat, if not a slight Romen lead with a reasonable likely voter model. That being said, I still think Romney is down in Virginia. Who knows, Iowa + Wisco makes up for Virginia.

The Republicans win the general election (including the Presidency) with a "likely voter" model that shows an electorate similar to that of 2010. Democrats at all levels will seek to expand the electorate as much as possible. Expect Democrats to do everything possible to register new, D-leaning voters and encourage them to vote. Expect Republican officials to make voting as difficult as possible.

The GOP has taken a massive 21,000 vote registration edge in Iowa. Compare that to the 100.000 edge the Democrats had in 2008, or the 4000 edge the GOP had in 2004.

It's in the bag.

Democratic registration was way up in Iowa in 2008 because of the contested Democratic primary, and Republican registration in 2012 is way up because of the heavily-contested Republican primary.  Fall college terms have yet to start, and if the Obama campaign has any uncertainty about Iowa it is going to double down on registering new voters and cajoling them to vote.

It's who has the edge at the end of registration that counts.

One in 12 say the commercials have altered their opinion about Obama or Romney, with 76% of them now favoring the president compared to 16% who say they switched to Romney.

The only interesting bit of data from this poll.

 

...and that is itself a portent of an R disaster. It's the marginal voter that counts most in a close election.

Certainly plausible of course, but it would not fit the pattern of what happened the last 2 election cycles.

In both 2004 and 2008, Iowa Democrats gained heavily in registration between January and June of that year. The parties ran much closer to even between June and November of that year.

The trend from 2008 to 2010 in politics is unlikely to continue. Anyone who predicts a continuation of a trend beyond the very short term is a fool. If I were to extrapolate the Presidential result from 2004 to 2008 in Iowa  I would expect President Obama to win the state by 20%, which I just don't see happening any more than you do -- or common sense.

The Republican Party did very well in 2010 with a reduced electorate, and with a not-so-reduced electorate it won't do so well. Such is the difference between a Likely Voters model that is the reality in off-year and midterm elections. (2006 was a possible fluke because of the extreme disdain for the then-current President, something that won't be repeated this year).

You can trust that the Obama campaign will do everything possible to prevent the "Likely Voters" model from being the one that operates in 2012.   

In this case, this statement is false. The trend of Iowa voter registration from 2008 to 2010 in fact accelerated between 2010 and 2012.

From November 2008 to November 2010 Iowa Republicans netted roughly 62,000 votes in registration. From November 2012 to present they have already netted 65,000 votes.

It is good for the GOP that they did so well in a -44,000 electorate in November 2010 in Iowa. Surely they will perform just as well in a +21,000 electorate.

[Miles]

If your theory on the correlation between partisan registration and party performance is so great, krazen, it must certainly apply to other swingy states.

So, in 2008, the Republicans had a 74,000 registration advantage in Colorado; since then, its been nearly halved to 43,000. Based on  the logic that you're construing with the Iowa numbers, I'll conjecture that Obama will perform at least as well as he did in 2008 this year.

[Miles]

If your theory on the correlation between partisan registration and party performance is so great, krazen, it must certainly apply to other swingy states.

So, in 2008, the Republicans had a 74,000 registration advantage in Colorado; since then, its been nearly halved to 43,000. Based on  the logic that you're construing with the Iowa numbers, I'll conjecture that Obama will perform at least as well as he did in 2008 this year.

[krazen1211]

If your theory on the correlation between partisan registration and party performance is so great, krazen, it must certainly apply to other swingy states.

So, in 2008, the Republicans had a 74,000 registration advantage in Colorado; since then, its been nearly halved to 43,000. Based on  the logic that you're construing with the Iowa numbers, I'll conjecture that Obama will perform at least as well as he did in 2008 this year.

More specifically, it is the deltas in partisan registration and partisan performance that is certainly meaningful as partisan registration is hard facts rather than agenda polling.


I suppose that your conjecture is more than plausible. Perhaps even reasonably likely. Why would it not be?

[Brittain33]

It looks like 50K Dems dropped off the rolls. I wonder what happened there.

[Miles]

If your theory on the correlation between partisan registration and party performance is so great, krazen, it must certainly apply to other swingy states.

So, in 2008, the Republicans had a 74,000 registration advantage in Colorado; since then, its been nearly halved to 43,000. Based on  the logic that you're construing with the Iowa numbers, I'll conjecture that Obama will perform at least as well as he did in 2008 this year.

More specifically, it is the deltas in partisan registration and partisan performance that is certainly meaningful as partisan registration is hard facts rather than agenda polling.


I suppose that your conjecture is more than plausible. Perhaps even reasonably likely. Why would it not be?

Because I also look at the polls. Obama is polling worse in Colorado than his final margin in 2008.

Being from a state like Louisiana, where Democrats make up 51% of the registered voters, I tend to be skeptical of voter registration as a foremost indicator of performance.

[krazen1211]

If your theory on the correlation between partisan registration and party performance is so great, krazen, it must certainly apply to other swingy states.

So, in 2008, the Republicans had a 74,000 registration advantage in Colorado; since then, its been nearly halved to 43,000. Based on  the logic that you're construing with the Iowa numbers, I'll conjecture that Obama will perform at least as well as he did in 2008 this year.

More specifically, it is the deltas in partisan registration and partisan performance that is certainly meaningful as partisan registration is hard facts rather than agenda polling.


I suppose that your conjecture is more than plausible. Perhaps even reasonably likely. Why would it not be?

Because I also look at the polls. Obama is polling worse in Colorado than his final margin in 2008.

Being from a state like Louisiana, where Democrats make up 51% of the registered voters, I tend to be skeptical of voter registration as a foremost indicator of performance.

In 1978, the Democrats held close to 90% of registered voters. I have read, though I am looking for a supporting source, that they held 60% of registered voters in 2000. The sizable drop in registration is very much reflected in the increasing struggles of the Democratic party in elections in Louisiana.

I would suggest that polls are less reliable than actual election results. We know very well that the GOP did not perform very well in Colorado in 2010, which of course fits with your conjecture.