Why is Gallup leaning towards Romney?

[TroyD]

Why has Gallup been leaning towards Romney so much this year?

Even Rasmussen has Obama ahead of Romney this weekend, and yet Gallup still shows a small lead for Romney.

What is going on? It's the only pollster that shows a Romney lead at the moment.

http://www.realclearpolitics.com/epolls/2012/president/us/general_election_romney_vs_obama-1171.html

[brush]

1. Nothing unusual

Gallup always leans way right. Just take whatever their results are with several grains of salt.

[TroyD]

2. No longer the Gallup of old?

It used to be a respected pollster several decades ago when the original Gallup was still alive.

Perhaps now it has been taken over by different ownership which leans more right?

[Jackpine Radical]

3. No telling. It all depends on the details of their statistical model,

how they weight responses by various categories of respondent, etc.

And, short of an election, there is no real way to verify their results.

Maybe even with an election there's no way…

[bigdarryl]

4. Because this polling shit is usless

[MFM008]

5. polling

Gallup, Rasmussan, Purple state, Fox all trend GOP, at least it drags down the average of the polls. Their job is to make this look as even as possible to hold the publics interest...........

[Marzupialis]

11. Fox drags down the average of all polls?

Fox has Obama by 9%.

[mikekohr]

6. They undersample Democrats among other things. Attached is a good link, stats and article

Alan Abramowitz adds:

It is beyond weird that Rush Limbaugh is now accusing the Gallup Poll of deliberately over-counting Democrats because the truth is that the Gallup Poll has for the past several months consistently shown a smaller Democratic advantage in party identification than other national polls.

As the table below demonstrates, of the major national polls that regularly report results on party identification, only Rasmussen shows a smaller Democratic advantage than Gallup, and only barely so.


?


The average Democratic advantage in party identification in the Gallup Poll since June, +6, is substantially smaller than the average in every other major national poll. In fact, no other major poll has shown that small a Democratic advantage even once during this time period. This is significant, of course, because party identification is very strongly correlated with opinions on other questions such as presidential approval, attitudes toward health care reform, and the generic ballot question. For example, Gallup recently showed Republicans leading on its generic ballot question for the first time this year. Of the other major polls that have asked this question, all except Rasmussen have continued to show a Democratic lead on the generic ballot question.

http://www.frumforum.com/does-gallup-poll-have-a-pro-gop-bias/

[TroyD]

7. David Frum

Thanks for the link. I'll check it out.

But David Frum is a Republican. Can his analysis be trusted?

[mikekohr]

25. Frum is not the only person to note Gallup's problems.

When people on the right recognise the same problems/issue as those on the left there is a consensus. Call me Captain Obvious.

[woolldog]

8. So do you believe they (Gallup) should re-weight their results to fit

some assumptions about what the voter ID breakdown should be?

I ask because if you believe in the effectiveness of random sampling, why would you have to re-weight the results?

[greiner3]

9. "Why is Gallup leaning towards Romney?"

DUH!

[Liberal_Stalwart71]

10. A slight bump for Robme's choice of Paul "Lyan"

[JoePhilly]

12. A close race generates Ad $$$$, even for the polling sites. The media wants a close race

for the same reason. If Obama pulls way out in front, the Ad money will shrink, and the Media Giants will lose all of that ad money.

And so, they will do what they can to keep the race close. And so will the polling firms.

On Edit: It used to be called "point shaving" in sports. In our current election coverage its called "balance".

[former9thward]

31. Not true.

The campaigns use their own private pollsters to determine where to advertise. They don't even look at the polls you and I see. Only top officials of the Obama and Romney campaigns know what the their numbers are.

[JoePhilly]

32. You seem to miss the point.

The polling sites, all of them, get paid by ad dollars (ad dollars of all kinds).

If the race is not close, no one comes to your poll site because the outcome is already known. And you lose money.

[former9thward]

33. You seem to imply the race is not close.

That must mean polling firms hired by Obama are falsifying the numbers and the Obama campaign is falling for it since they are spending hundreds of millions for ads. Who knew the top officials in the campaign are so easily suckered?

[JoePhilly]

34. It not that close. Not at the Electoral College where it actually matters.

What I am describing is the need, particularly for the media, to claim the race is close, and then also do what they can to cause that to be the case.

Did you know that recently, Obama stoppped running ads in PA? His lead there is strong enough. Which is bad for the local TV folks in PA. They'd love for PA to be a battleground state.

Now of course Fox News and the GOP wants Romney, so they will claim the race is closer than it actually is.

That's why when there is a poll showing Obama up 9%, it must be wrong, but if Obama is only up 2%, then Romney is surging.

Its how the media tries to manipulate people.

[former9thward]

35. I think it is much closer than you think.

But of course there is no way of proving that one way or the other. I try and look at the realclearpolitics average of polls and that has the race within the margin of error at this point. We will see.

[TroyD]

36. Gallup

Gallup has Romney ahead yet again today. Gallup says Romney has been ahead for 6 days in a row now and that this is the longest he has been ahead in months in their poll.

[JoePhilly]

37. I track realclear too ... compare Obama's TREND line now with his trend line in 2008.

What you will find is that, even though the range is close, Obama is always leading it. Always.

In statistics, a small difference can still be HUGE, if it is persistent. Obama's lead is persistent.

You are correct to watch the average of the polls, but you must also examine the trend lines of those averages. Those indicate the extent to which the numbers could change.

A small persistent lead is stronger than a large but intermittent lead.

Also ... you can examine the "paths to 270". Obama actually has a path to 270 that does not include PA, OH, or FL. Romney MUST win OH and FL, and CO, and VA, and NC and a few others where he now, persistently, trails Obama.

If Obama wins VA, or Ohio, or FL, or CO ... its over.

[TroyD]

38. Good point

If you look at 2012, Mitt Romney hasn't taken the lead once all year. It doesn't look like he has been ahead nationally since October 2011.

[JoePhilly]

39. exactly ... he might win a poll here or there, but he can't get the overall trend to change.

I think a few polls, like Rasmussen, manipulate their polls in the months prior to the election, hoping to get the undecided to "follow the crowd". Their poll is only accurate in the last week of the elections.

And if you toss out Rasmussen, Obama's persistent lead becomes stronger, and more persistent.

[developertest01]

13. Because they are Likely Voters

Because they are likely voters, they voted last time in 2010, and are likely to vote again. That being said their sample is very large so they cannot be sluffed off. They do however over factor Republicans. The polls do look good for Obama as a "Whole" but there is troubling declines in VA, not enough wiggle room in FL and Ohio is perhaps the worst, at one point this summer I hesitated to call OH a swing state, now, it may very well go to Romney. For the first time I see Romney's viable electoral path, after considering Ohio now. Before I had no belief whatsoever Romney could win. If the election were held TOMORROW, I think Romney would win. He did get a lousy bounce from the Ryan pick, (but some bounce none the less). I dont think polls are good predictors until mid to late September.

[TroyD]

14. Obama would win

If there was an election tomorrow, Obama would win, actually. He's ahead in the national polls, most of the state polls, and the Electoral College.

[developertest01]

17. Not Really

National means squat. What will count is OH, VA, NV, FL, CO and maybe the deal breaker NH. Now as Gallup slants to the right a bit, I think the other RCP polls slant to the left a bit. The National Polls mean squat because Obama has big leads in the states he will already win anyways and nobody doubts it like CA, he has a bigger lead than in say TX. Gut feeling is that Obama must have a 3 point lead in each of those true swing states to win. Obama has a 1 point lead in my state FL, but right now, I would bet money Romney takes it. The problem with Democrats that the Republicans do not see as much, is simply that they don't vote in the same numbers. Sure, give em a call they will tell you Obama is their man, but on election date its DUDE my Brah came in from Cali and we were slammed. A joke obviously but they are not as energized no doubt. Additionally you have to look at RECENT polls, some giving Obama the lead were back in July. There is still a GOP convention where there will be another bounce as well. Obama has the edge in the polls regardless. If I were a Romney supporter I would be nervous, but as an Obama supporter I want to be cautious. The recent gains by Romney in OH and VA are disturbing. If Romney could solidify a win in Ohio, its the devils game at that point, additionally it would probably signal a "movement" like we saw in 2010 giving him WI, IO as well, then it would come down to maybe CO or NH as the tie breaker. Obama looses CO in such a (Romney takes OH and FL) scenario, Romney is President. Romney takes CO only, and Obama takes NH in that scenario, Obama wins by like 4 electoral votes. OH right now is devastating, VA was like wow, Obama even has a lead in VA ( a one shot kill to Romney) but RIGHT NOW, it looks like Romney as expected may very well take VA, thats the case, OH is key. If Obama won FL, its all over though, but thats too close to call. I think its possible that the Ryan pick could swing Obama over in FL and then you can go to bed early while they count the extra 78 electoral votes in the West that are guaranteed Obama, if not, come 11 PM EST on election day, Romney will hold the lead pending CO.

[mikekohr]

24. Princeton Election Consortium (Highly accurate) has President Obama at 308 electoral votes -map-

?
link: http://election.princeton.edu/electoral-college-map/

The nation is divided. This will be a repeat of 2008. All but a few percentage points of the electorate have made up their minds. That explains the "static" nature of the poll numbers over the last year. The President and the Democratic Party can make a rout of this yet. I'm predicting the President gets 332 to 348 electoral votes. But we need 25 in the House, need to hold the Senate and reform the fillibuster rule or we get at least 2 more years of deadlock and Republican obstruction. That's how empires fall.

[developertest01]

26. Interesting

I never heard of this site. It was interesting. I especially like it when they put their past election results right out there. I was taking a look back at Rasmussen and they had several states going for McCain, that ended up +4 or 5 ( a huge margin these days) for Obama.
I think swing states by and large move together, its either a good year for one or the other. I am not sure I have so much faith in IA and CO, however, Obama holds the same margin their as in FL. My concern came from Ohio, where Obama held at times this year, an 8 point lead. Now its another FL. It might not be real, it might be Purple Strategies which seems like (Red Strategies) to me but they in fact done some of the most recent polling.

[TroyD]

27. North Carolina

One of the things that is disappointing is that Obama may miss out on North Carolina this year. Obama was actually leading there until May when the Republicans stirred up the gay marriage issue in the state. Obama's numbers hit the floor right after that.

They've recovered somewhat, but the odds are in Romney's favor in NC unless he takes a hit over the next 2 months.

[CobaltBlue]

30. The numbers matter!

Numbers tell what's about to happen. And, afterward, what did happen.

Look at the margins since 1980:

1980: Ronald Reagan (R-California) unseats incumbent Jimmy Carter (D-Georgia). In 1976, Gerald Ford (R-Michgian) lost nationally to Carter by R-2.06. Reagan shifted R+11.80 to beat Carter nationally by R+9.74. 49 states shifted away from the Democrats and Carter in favor of the Republicans and Reagan. (Exception: Vermont.) When Ford lost, he still carried 27 states. (Only Ford and 1960 Richard Nixon lost their elections while having carried majority of states as is the inverse is true that only Carter and John Kennedy were elected with carriage of less than half of the available states.) So, Reagan took all 27 Ford-carried states, worth a mathematical 241 electoral votes, and, due to level of shift, managed to flip 17 states that voted in 1976 for Carter. From the 1976 Democratic to 1980 Republican column. The 17 pickups were worth 248 electoral votes to give Reagan a landslide of 489 electoral votes.

1984: Reagan gets re-elected. He had 44 states in his first election and garnered a national shift of R+8.48 to beat Walter Mondale (D-Minnesota) by R+18.22. In the electoral college, all 44 were retained and Reagan flipped five of the remaining six that were in the column for 1980 Carter. Total of 36 electoral votes between them for 525 electoral votes.

1988: Republicans get a third straight victory. Declined margin, though, with George Bush (R-Texas) winning promotion from vice president to president. He loses a good 60 percent of the 1984 Reagan margin and beat Michael Dukakis (D-Massachusetts) by R+7.73. In the Electoral College, Bush lost 9 of the 49 that voted for re-electing Reagan. (All but Tennessee shifted away from incumbency and toward opposition; party wise.) Bush lost 99 electoral votes to carry 40 states with 426 electoral votes.

1992: Bill Clinton (D-Arkansas) unseats George W. Bush. The four-term Arkansas governor realigns map and a counter to Republicans realigning the South; see longterm voting patters of Deep South duo Alabama/Mississippi vs. northern and liberal bastion Vermont. (The two major parties realigned … their party.) Clinton shifts Dukakis' national margin of D-7.73 and gains D+13.29 to unseat Bush in the popular vote by D+5.56. 49 states shift away from Republicans and Bush Sr. and toward Democrats and Clinton. (Exception: Iowa, which voted for Dukakis by over 10 points but carried for Clinton by about 6.) All ten states, plus District of Columbia, in the column for Dukakis carry for Clinton. With reallocations of the electoral map (due to 1990 results of U.S. Census Bureau), they add up to 105 electoral votes. Clinton ends up flipping 22 states, including his home state of Ark. and vice-presidential running mate Al Gore's Tenn., and gets a pickup of 265 to win a total of 32 states + D.C. for 370 electoral votes.

1996: Bob Dole challenges Clinton and loses, as the 42nd president wins re-election by a margin of D+8.52. That was a gain of D+2.96 over his first election. Some states scale back support; some increase. Typical in incumbent years, but also consistent is the gain in electoral votes. Clinton lost his 1992 carriages of Colorado, Georgia, and Montana (combining for 24 electors) but wins pickups in 1992 GOP holds Arizona and Florida (worth 33 votes). He goes from 370 (1992) to 379 (1996).

2000: George W. Bush (R-Texas) wins controversial first-term election. National shift is R+8.00 which, factored into Bob Dole's R-8.52 loss, results in Bush failing to win over the popular vote as incumbent v.p. Al Gore (D-Tennessee) prevails there with D+0.52. But 49 states shifted in the direction of the opposition party, leaving D.C. and Maryland the exceptions, and it ends up flipping 11 states worth 112 electoral votes to deliver the Electoral College for Bush at 271.

2004: Bush gets re-elected by a margin of R+2.46. It's historically the lowest for an incumbent re-elected to a second term, and with a gain in the popular-vote margin (R+2.98). He loses New Hampshire and flips New Mexico and Iowa for a total 286 electoral votes.

2008: Democratic pickup for the White House with the first-ever African-American president. Barack Obama (D-Illinois) defeats John McCain (R-Arizona), who was screwed thanks to Bush's low-approval job numbers, the economic meltdown of September 2008, the housing crisis, and unemployment. Obama gets shift of D+9.72 which, added to John Kerry's (D-Mass.) 2004 loss of D-2.46, accounts for the 44th president's first-election victory by D+7.26. 45 states shift away from Republicans and McCain and toward Democrats and Obama. (Exceptions: Arkansas and Louisiana deepen their redness as they probably would not have had the nominee been Hillary. Oklahoma, Tennessee, and West Virginia are near status quo. All but the Sooner State supported Bill Clinton in both his elections.) All the Kerry states (19 of them) + D.C. naturally carry, and Obama gets a pickup of seven states carried by Bush in single digits, and he flips two other states (plus the 2nd Congressional District in Nebraska) which voted Bush by double digits. Obama takes Kerry's 252 and flips an additional 113 electoral votes to end up with 365.


The point of all this summary is to illustrate how and why these numbers matter. They correlate with the results of past elections. They do so with all elections. And this is why I give a damn and would prefer Obama to win re-election with a gain in the popular-vote margin. (Even at a modest level experienced by Clinton and Bush Jr.)

[pinto]

15. Single sample survey - What's your vote?

[Warren DeMontague]

16. I sense your concern.

[developertest01]

18. Here is Romney's VIABLE Path

[link:https://docs.google.com/document/d/1tx3eLZuCCYm5h91Aq8b4PzgBiZ_Mcf7rNlyuR3de-gw/edit|

[TroyD]

19. So. . .

That Map shows Romney winning Ohio, Virginia, Colorado, Iowa, & Florida.

It's possible, and you can never take your eye off Republicans when it comes to the possibility of electoral fraud, but Obama is ahead in almost all of those states right now.

[developertest01]

20. My Thoughts

You have to know your enemies plan for victory. The way I see it, is that in those states, Obama's polls show a lead that is super within the margin of error. If the polls are one or two points in favor of Obama, very believable, those could all be in Romney's favor at this point maybe with the exception of the TRUE SWING STATE, OH. With that, I think you will not see a mix here, if OH goes to Romney, it can be considered a bit of a national movement that, same movement giving him VA (which he should by reason stand to take), FL, CO, and certainly FL. I dont think you see a scenario where Obama wins FL and Romney wins OH, look at the historical RCP polls, its generally some sort of a sweep. In the same vain it could give Obama a landslide. Saying Obama takes FL, he will take OH, NV, CO, NH, WI and will have a "landslide" by a razor thin margin in each state. An Obama victory depends on being 5+ points up in recent polling in OH. In OH, he has lost ground. If you see a big Republican blitz to the polls to get rid of Obama, and that razor thin margin can deliver OH, the same National feel or margin will also deliver CO, FL, VA which would be enough. I picked this map giving Romney a slant. I wanted to see what is viable, he cannot win CA we know that. Given that I looked at where more recent polls gave Romney the edge, sure Slantmussen was taken into consideration but so was Leftberg. Its not scientific, but its a picture of how in states where Obama may hold a 1 or 2 point lead aggregate, they are wrong. Gallup is the scariest right now, because they have a large sample, and have been consistent. They are not like Slantmussen where one day Romney is +5 and the next -2. Look at the polls from say after 8/12 (state polls) the states in red I have outlined are certainly up for grabs, especially if there is even a razor thin (lets get Obama out wave).

[developertest01]

23. 2008 Results

If you look at the polling from 2008, noone had their shit together until 9/21. We can say that definitively in retrospect. Rasmussen had FL way off and OH at a tie with McCain, while the other polls placed accurate bets. Gallup had Obama over McCain by a margin way more than he actually got. So Gallup, straight up confusion for me. It is irrefutable, that Rasmussen is 2-3 (and in 2008 as much as 6) in favor of the GOP. Whats up with Gallup?

[developertest01]

21. The Battle Plan on Election Night

Its all about timezones. If FL goes blue, drink you champagne and go to bed early. If FL is Red, wait an hour an a half for OH. If OH goes blue, drink the other half of that bottle, if it goes Red, it means Romney will take IO, and WI as well. Then call into work because now you will need to wait for CO. NV called a swing state, is pretty solid Obama, but if CO goes Red, take a valium and go to sleep because you will wake up to President Romney. FL could go Red whilst OH goes blue, very possible. If FL goes blue, it means no Romney wave and OH will go blue, and even if it doesnt, CO will, NH will, WI will. So really although not the dealbreaker, a Blue FL is a clear signal of an Obama victory. Right now I think FL will go for Romney, and OH is looking scary, so my CO prediction is (RIGHT NOW AS IT STANDS) the dealbreaker. Again Polls, they dont mean doo doo until mid-late Sept. Even McCain had a lead in august 2008 for awhile, and I did not even bother to stay up because I knew Obama would be elected.

[developertest01]

22. Old Polls

Also, many of the polls averaging into Obama's lead are like 7/26 which might as well be 7/26/1976. The map is based on 8/12 or greater giving Romney the margin of error.

[Capt. Obvious]

28. I enjoyed your sober analysis in this thread

Thanks

[Sunlei]

29. because GOP will use early voting to their extreme advantage

County by county across America, they seem poised (primary practiced) to gather thousands of early votes. All it takes to make a difference is a couple hundred thousand votes in some states.

[chloes1]

40. NewsMax shows Romney too

At least Newsmax will let people vote, so that could change... [link:http://www.newsmax.com/surveys/RomneyvsObama/Mitt-Romney-vs-Barack-Obama/id/48/kw/default?PROMO_CODE=FBAC-1|