« August 2004 | Main | October 2004 »

September 30, 2004

Harris Poll: Statistical Tie on Eve of Debate

George Bush leads John Kerry 48-46 percent among nation-wide LV's, with 3 percent for Nader, 2 percent for none of these and 2 percent not sure/refused, according to a Harris Interactive Poll conducted 9/20-26.

LA Times Poll: Bush Ahead by 4 As Debates Begin

Bush leads Kerry 49-45 percent of nation-wide RV's, with 6 percent undecided, according to a Los Angeles Times Poll conducted 9/25-28.

September 29, 2004

Kerry, Bush Tied in New Economist Poll

John Kerry and George Bush are dead even at 46 percent of nation-wide RV's, with 2 percent for Nader in a new Economist/YouGov Poll conducted 9/27-9.

New Study Targets Key Groups for Kerry Gains

John Kerry can make major gains among key demographic groups of discontented white voters: women blue and pink collar workers; rural voters; those under age 30; and senior women, according to a study of post-Labor Day polls by Democracy Corps reported 9/28. The study also identified other demographic groups Kerry should target for significant gains.

Single Women: Put Domestic Concerns First

Unmarried women RV's in 12 of the 16 swing states say Jobs and the economy, affordable health care and education are more urgent priorities than the situation in Iraq and believe that John Kerry does a better job of addressing these issues, according to a survey by Greenberg Quinlan Rosner Research for Women's Voices conducted 9/8-19.

New IBD/CSM/TIPP Shows Dead Heat in Race for White House

Kerry and Bush are tied at 44 percent of nation-wide RV's, with 11 percent unsure, according to an Investor's Business Daily/Christian Science Monitor/TIPP Poll conducted 9/22-27.

Pew Research Center Has Bush Up 3, Then 8 in Two New Polls

Bush leads Kerry 45-42 and 48-40 in two polls of nation-wide RV's conducted 9/17-21 and 9/22-26 respectively by the Pew Research Center for the People and the Press.

September 28, 2004

Even MoveOn.org Is Playing the "How Can Gallup......." Game!

I'm sure many have heard about today's full-page ad, "Gallup-ing to the Right", in The New York Times (page 5!) by MoveOn.org questioning Gallup's methodology and numbers. But if you haven't actually seen the ad, by all means click on the link and take a look. I think it's a striking and effective ad.

The numbers in the ad, which are quite eye-opening, are rock-solid. The ad says Gallup's average LV lead for Bush this month has been 10 points, while the average of all other LV polls has been 4 (they're clearly referring to 3-way LV results--which are by far the most numerous LV results--based on other data in the ad). That's correct. Even taking into account data released since 9/26 (the end-date for the ad's analysis), Gallup this month has averaged a 10 point lead for Bush among LVs in 3-way trial heats, while the other 27 3-way LV trial heats taken this month have averaged a 4 point Bush lead.

Similarly, the ad says polls released since 9/12 (that is, two weeks before the end-date of the ad's analysis), excluding Gallup, have averaged a 3 point lead for Bush in 3-way LV trial heats. Correct again, even adding in polls released since 9/26. In the 17 3-way trial heats released since 9/12 by polling organizations whose names are not "Gallup", Bush is averaging just a 3 point lead.

The "How Can Gallup......." Game (Continued)

Well, our correspondent, Alan Abramowitz, has been getting into the spirt of the "How Can Gallup...." game, so I thought I'd share some of the fun he's been having with DR readers:

It's hard to know where to begin when it comes to all of the preposterous results in [the new Gallup poll]. First of all, they've got about a 10 point Republican advantage in party ID among registered voters. I am guessing that this is one of the largest Republican party ID leads in the history of the Gallup Poll. So according to Gallup, what's happened since the Republican convention is something on the order of the New Deal realignment in reverse. If you reweight their data based on the partisan composition of the 2000 exit poll, you get something like Bush 48, Kerry 47.

Going along with the ridiculous party ID results, they've got Bush now leading by 15 points in the Midwest and by 21 points in the West. For Bush to be leading by 21 points in the West he'd have to actually be leading in California by about 10 points and cleaning up in Washington and Oregon as well. This is totally out of line with recent independent state polls showing Kerry leading in California by 15, leading in Washington, and running about even or slightly ahead in Oregon. According to this poll, Bush is doing better in the West as a whole than he is in Montana, where the most recent independent poll has him ahead by only 18. The Midwest result is totally out of line with the most recent independent polls in Illinois (Kerry +15) and Ohio (Bush +3).

Gallup's own state polls are totally out of line with this result. How can Bush be 2 points ahead among RVs in states like Florida and Nevada but 13 ahead among RVs in the entire country?

How indeed? Anyway, be sure to play "How Can Gallup...." at home. Don't let Alan and I have all the fun!

September 27, 2004

The "How Can Gallup........" Game

Say everybody, I've got a great new game to play! It's the "How Can Gallup....." game.

How do you play? It's easy! Just take the latest Gallup outlier and compare it to other publicly available data that seem to contradict it. And let the fun begin!

Today's Gallup outlier is their RV result in their latest poll. As many have no doubt heard, that result puts Bush up a whopping 13 points over Kerry in a head-to-head matchup. Maybe that 13 points sounds familiar. Well, last Gallup poll, it was their LV result that had Bush ahead by 13 and that was far away from everyone else's results. Guess they like to spread those outliers around.

Here are some examples I've come up with, but the great thing about this game is that all of you can play at home and make up your own examples, so be sure to do so.

How can Gallup......have Bush up by 13 nationwide, when he's only up by 2 points among Florida RVs?--and according to their own poll!

How can Gallup.....have Bush up by 13 nationwide, when he's only up by 2 points among Nevada RVs?--again, according to their own poll.

How can Gallup....have Bush up by 13 nationwide, when he's only up by 3 points in Ohio, according to Fox News?

How can Gallup....have Bush up by 13 nationwide, when he trailing by 5 points in Pennsylvania and 2 points in Michigan, according to Fox News?

How can Gallup.....have Bush up by 13 nationwide, when he's only leading in Montana by 18 points? (in 2000, Bush won Montana by 25 points)

How can Gallup.....have Bush up by 13 nationwide, when he's behind Kerry by 15 points in California and 20 points in New York?

How can Gallup.....have Bush tied in the solid blue states (that is, the non-battleground blue states, so WI, MN, IO etc aren't included), when he is trailing Kerry by 15 points in California and 20 points in New York?

How can Gallup.....have Bush up by 13, when he's only leading among independents by 2 (and that was exactly Bush's margin among independents in 2000 when, as you recall, he did not win the popular vote by 13 points)?

Not only that, Bush's current margin among Republicans in the Gallup poll is not too far from his 2000 margin (93-6 now vs. 91-8 then) and Kerry's margin among Democrats is identical in magnitude to Gore's (85-10 now vs. 86-11 then).

Actually, this one is kind of easy. The only way you can produce a 13 point Bush lead with these internals is if you have quite a few more Republicans than Democrats in the sample--my guess is 7-8 points more. If you re-weight their sample to the 2000 exit poll party ID distribution (and I kind of have to do this, just to drive certain pollsters and their acolytes into a frenzy), you wind up with a modest Bush lead of 2 points.

See? This game is fun! Tell a friend about the "How Can Gallup....." game and add a little zing to your election season.

WAPO-ABC News Poll: Bush Ahead by 7

Bush leads Kerry 51-44 percent among nation-wide RV's, with 2 percent for Nader, 2 percent other/neither and 2 percent no opinion, according to a Washington Post-ABC News poll conducted 9/23-6.

September 26, 2004

So, How Well Does That Gallup Likely Voter Model Really Work?

As I've repeatedly argued, Gallup's likely voter (LV) data would be inappropriate to use until right before the election even if they worked perfectly at that particular time. But it's interesting to note that their LV data lately have not been working terribly well, compared to RVs, even on election eve, which is supposed to be when theses data really shine. That kind of undercuts their pseudo-rationale for their cavalier use of these data.

After all, David Moore of Gallup has admitted: “We simply do not know, nor can we know, which model is better during the campaign itself.” But he defends the use of LV data months and months before the election because “if it is the most accurate model just before the election, it is probably the most accurate during the campaign as well.”

Oh really. Well, let's see just how accurate these data have been recently. Here are the last 4 presidential elections and the track records of the Gallup LVs and RVs in terms of predicting the final margin of the election:

2000: RVs, +1D; LVs, +2R; actual result, +.5D
1996: RVs, +16D; LVs, +11D; actual result, +8.3D
1992: RVs, +8D; LVs, +12D; actual result, +5.5D
1988: RVs, +8R; LVs, +12R; actual result, +7.7R


1. In 2000, the RVs pretty much get it on the nose and correctly call the popular vote winner, which the LV data miss.

2. In 1996, the LVs are indeed substantially better--but in 1992, the RVs are substantially closer to the final outcome. And in 1988 the RVs are pretty much are dead-on while the LVs are more than 4 points off.

That's 3 out of the last 4 elections for the don't-get-no-respect RVs.

Not too impressive. And this track record justifies subjecting us to LV data from the very beginning of the campaign and acting like the RV data are somehow inferior? Puh-leeze.

As a public service, I reproduce here the fabulous questions with which the Gallup folks produce their fabulous LV samples:

1. SALIENCE: How much thought have you given to the upcoming election for president?— quite a lot, or only a little? (“Quite a lot” or “Some” as a volunteered response score one point)

2. KNOWLEDGE: Do you happen to know where people who live in your neighborhood go to vote? (“Yes” scores one point)

3. BEHAVIOR: Have you ever voted in your precinct or election district? (“Yes” scores one point)

4. BEHAVIOR: How often would you say you vote—always, nearly always, part of the time, or seldom? (“Always” or “Nearly always” scores one point}

5. INTENTION: Do you, yourself, plan to vote in the presidential election on November (*), or not? (“Yes” scores one point)

6. BEHAVIOR: In the [last] presidential election, did you vote for (*) or (*), or did things come up to keep you from voting?

7. INTENTION: I’d like you to rate your chances of voting in the upcoming election for president on a scale of 1 to 10. If “1” represents someone who definitely will not vote, and “10” represents someone who definitely will vote, where on this scale of 1 to 10 would you place yourself?

If a voter answers each of these questions the “right” way, they get a 7, miss one and you get a 6, and so on. In practice that typically means all of the 7s—given full weight—plus some proportion of those with lower scores (usually the 6s), who are weighted down so that the size of the likely voter sample matches the projected turnout for the year (apparently 55 percent this year). All other voters are discarded from the sample.

This seems like an awful lot of work for something that (a) doesn't work that well at election time; and (b) isn't even appropriate during most of the campaign.

But Frank Newport at Gallup insists this is a "scientific" approach to take to polling. Sounds more like dogma to me.

Kerry Up 1 in NH See-Saw Race

Kerry leads Bush 47-46 among New Hampshire LV’s, with 7 percent unsure, in a head-to-head poll conducted 9/20-23 by Research 2000 for The Concord Monitor.

Kerry Gains on Bush in Time Poll

Bush leads Kerry 48-44 among nation-wide RV's in a new Time magazine Poll conducted 9/21-23. (Bush was up by 12 among RVs in Time's poll 2 weeks earlier).

September 25, 2004

Fox News Poll: WH Race in Statistical Tie

George Bush leads John Kerry 45-43 percent among nation-wide LV's, with 12 percent other in a head to head poll conducted September 21-22 by Opinion Dynamics for Fox News. (margin of error +/- 3).

Marist Poll: Statistical Tie

Bush Leads Kerry 47-45 percent of nation-wide RV's, with 3 percent for Nader and 5 percent undecided in a Marist Poll conducted 9/20-22. (margin of error +/- 3.5 percent)

CBS Poll: Bush Ahead by 8

George Bush leads John Kerry among nation-wide RV's 49-41 percent in a head-to-head CBS News Poll conducted Sept. 20-22.

AP-Ipsos Poll: Kerry Lags by 9

Bush leads Kerry 51-42 among nation-wide RV's, with 2 percent for Nader, according to an Associated Press-Ipsos Poll conducted 9/20-22.

September 24, 2004

Gallup Vs. Democracy Corps

The indefatigable Alan Abramowitz offers the following comparison of the track records of Gallup and Democracy Corps in the last presidential campaign. The results may surprise you and have some important implications for assessing recent poll results.

What should we make of the huge discrepancy in the results of recent national polls? On the one hand, the prestigious Gallup Poll has George Bush leading John Kerry by 13 points among likely voters. On the other hand, a Democracy Corps Poll released just a few days later shows that the race is a dead heat. Who should voters believe? Most people probably assume that the Gallup Poll is a lot more credible than a poll conducted by a partisan polling organization. After all, Democracy Corps is headed by two well-known Democratic political operatives: James Carville and Stan Greenberg. But if that's what you're thinking, an examination of the track record of these two polling organizations during the 2000 presidential campaign might make you change your mind.

On October 25, less than two weeks before the 2000 presidential election, Democracy Corps released a national poll showing Al Gore leading George Bush by 2 points. On the same day, Gallup's national tracking poll showed George Bush leading Al Gore by 7 points. One day later, the Gallup tracking poll had Bush up by an incredible (literally)
13 points. Score one for Democracy Corps.

Okay, you're probably thinking, maybe Gallup just had a bad day (or two or three). But six days later, on October 31, Democracy Corps released it's final pre-election poll showing a tie in the presidential race. Gallup's tracking poll that day still had Bush leading by 5 points.

Make it Democracy Corps 2, Gallup 0.

In its final poll, Gallup did have Bush's lead down to only 2 points--still not as good as Democracy Corps' poll released almost a week earlier.

Now I'm sure that Gallup and its defenders in the polling business would argue that Democracy Corps was just lucky and that the Gallup Poll was actually more accurate all along because it was picking up real shifts in voter sentiment in the final two weeks of the campaign. Maybe Al Gore really did go from a 13 point deficit on October 26 to a small lead on Election Day, maybe there really is a 6 point Republican advantage in party identification among registered voters right now, as Gallup is currently telling us, maybe Iraq will hold free and fair elections next January, and maybe the tooth fairy will pay you a visit tonight, but I wouldn't count on it. If I had to put down a bet right now on which poll will prove to be closer to the actual results of the 2004 presidential election, Gallup or Democracy Corps, I'd put my money on James Carville and Stan Greenberg.

Oh, and if you're looking for a backup poll, I'd recommend Fox News/Opinion Dynamics. Their final pre-election poll in 2000, released five days before the election, also showed a dead heat. Their latest 2004 poll has Bush leading Kerry by 2 points, right in line with the new Democracy Corps poll. And I don't think anyone would accuse Fox News of having a pro-Kerry bias.

Pretty darn interesting. And if reading this has put you in the mood for a little DCorps analysis, you really must check out their latest analysis of "The State of the 2004 Race". Highly worthwhile.

More On Those Alleged Security Moms

Yesterday, I highlighted analyses by Phillip Klinkner and Anna Greenberg critquing recent data-challenged newspaper stories about "security moms" and the vanishing gender gap. Today, it's Noam Schieber's turn over at The New Republic in his amusingly-named "Mothers of Invention" article. Here's a taste of what Noam has to say:

If you've been following the presidential campaign these last few weeks, you've probably heard a thing or two about security moms--the erstwhile soccer moms who became obsessed with terrorism after September 11, and, in the process, began tilting Republican. The typical "security mom" story--variations of which have appeared in The Washington Post (twice), The New York Times, the Chicago Tribune, and the Philadelphia Inquirer in recent weeks, as well as on CNN, ABC, and NPR--cites the hair-raising effect of the recent Russian school massacre. It mentions Laura Bush's frequent pitches to women on security matters, and notes how the Republican Convention was awash in security talk. Often the stories are larded with a testimonial by a real-live security mom, invariably a pro-choice, pro-gay rights, anti-death penalty former Gore supporter who's convinced only George W. Bush can keep her children safe. All of them conclude that security moms could cost John Kerry the election.

Oh, and the stories usually have one other thing in common: They're based on almost no empirical evidence.

...[I]t wasn't until after this summer's Republican convention that security moms became a bona fide growth industry. Suddenly, as The New York Times put it earlier this week, "an issue Mr. Bush had initially pitched as part of an overall message--which candidate would be best able to protect the United States from terrorists--has become particularly compelling for women." Except that, well, it hasn't--at least that part about "particularly compelling." The problem with most of the reporting on security moms is that it fails to distinguish between Kerry's support among women relative to men (i.e., the gender gap, which doesn't tend to fluctuate much over short periods of time) and his absolute level of support among women (which fluctuates just like it does for anyone else). In fact, while Kerry has lost ground among women since August, he's lost about the same amount of ground among men.

There's lots more. By all means, check out the whole article.

Strategy Notes:
John Belisarius

What the Public Really Thinks About Iraq - And the Challenge Facing Kerry

Now that the Kerry campaign has clearly focused on Iraq as the central issue in the 2004 presidential race, public attitudes on the subject become critically important. Kerry faces a difficult challenge because he must win support from voters who hold sharply divergent views.

There are four basic questions whose answers, taken together, provide a reasonably clear outline of the public’s general view of Iraq as a political issue.

1. Was decision to invade right or wrong?
2. Has it helped or hurt the war on terrorism
3. Should the US stay until a stable government is established or bring the troops home within a short time.
4. How well is Bush handling the situation?

While there are other important questions, these four provide a solid basic framework or outline of the key public attitudes. They have the additional value of having been asked a sufficient number of times in recent weeks to provide survey data that reflects a variety of question wordings and polling methodologies.

The central and most dramatic fact about current public opinion on Iraq is that, even after the Republican convention, public support for the Bush administration’s approach on all four of the key questions above has still not returned to the levels of last January.

According to a Sept 17th Pew Research Center survey, last January:

65% of the general public thought that launching the war in Iraq was the “right decision”. Only 30% disagreed.

55% believed that the war in Iraq “helped the war on terrorism”. Only 32% thought it had hurt it.

63% felt America should “keep military troops in Iraq until the situation has stabilized” while 32% wanted the U.S to “bring the troops home as soon as possible”.

59% approved of the way Bush was handling the situation in Iraq, in contrast to only 37% who disapproved.

In short, support for Bush and his policies hovered in the 55-65% range while contrary views did not rise above the low and mid 30’s. In the following 4-5 months, however, all of these positive numbers for the Bush administration declined dramatically under the impact of the startling military reverses in Faluja and other major cities along with the appearance of major reports critical of the Bush Administration and the revelations regarding the Abu Ghraib prison scandal.

As a result, by June the Pew data showed that the percentage believing the war was the right decision had fallen by 10% (from 65% to 55%), the percent who felt the Iraq operation helped the war on terrorism had fallen by 12% (from 55% to 43%) and the percentage wishing to keep troops in Iraq fell 12% as well (from 63% to 51%).

(Conversely, those who felt the invasion had been the wrong decision increased by 8%, those who felt it hurt the war on terrorism rose by 12% and those who wanted to bring the troops home more rapidly also increased by 12% )

Finally, the percentage approving Bush’s handling of Iraq fell a remarkable 17% (from 59% to 42%) while those who disapproved rose 14% (from 37% to 51%).

This dramatic decline in the support for Bush and his policies in Iraq played a substantial role in creating the overall lead in the polls Kerry assumed during the spring of 2004. But this lead was clearly tenuous because it was substantially rooted in a perception that the Bush administration had lost control of events and literally did not know what it was doing. It was therefore reasonable to expect that, if the administration could prevent further military reverses for several months and avoid further scandals or damaging revelations, these depressed levels of support would gradually reverse themselves and begin rising back toward their previous levels.

In fact, however, the recovery Bush enjoyed was extremely limited. By August, the percentage saying that the U.S. made the “right decision” actually fell another two points to 53% and remained at that level even after the Republican convention. The percentage saying that the incursion had helped the war on terrorism only rose 3% from June to September (from 43% to 46%) and the percentage wishing to keep the troops in Iraq until stability was achieved also rose only by 3% from June to September (from 51% to 54%).

Those approving Bush’s handling of the situation rose from 42% to 47%. Those disapproving declined from 51% to 45%.

In short, while the downward slide in support for the war and Bush’s handling of it that began in April slowed and slightly reversed during the summer, the reversal was actually rather modest and, even after the Republican convention, support was still substantially below the level it had been the winter before. Support for the view that the war was the “Right decision” was 12% below the level of January, support for the view that it had helped the war on terrorism was down 9%, support for keeping the troops in Iraq was down 9% and Bush’s approval rating was 12% below its January level.

It is important to keep this larger pattern in mind when looking more closely at the specific questions. There are important differences in the way Americans feel about the four distinct issues, but in all these cases, the key fact is that Bush and his policies in Iraq have still not returned to their former level of popularity.

In examining the first of the four questions in more detail, the most striking thing that emerges from a comparison of the major polls taken during August and September are the clear differences that appear depending on the wording of the question. Polls which ask whether the U.S. “did the right thing”, “made the right decision”, or “made a mistake” tend to show a clear majority affirming that the U.S. was indeed “right”

When the wording shifts to whether it was “worth going to war”, however, polls tend to show the respondents being much closer to evenly divided. And when the question asked is if the war was “worth the number of U.S. casualties and financial cost” or “worth the cost in lives and dollars”, respondents reply that it was not by margins of 6-9%.

Democracy Corps, whose polls often go beyond traditional polling methods to incorporate techniques derived from market research and the social sciences, in mid- September asked respondents to choose between a full-paragraph statement about Iraq that noted the 200 billion dollar cost, the 1,000 dead and the loss of control over a substantial portion of the country and two “anti-Kerry” statements challenging either Kerry’s supposed “waffling” on Iraq or his willingness to leave Saddam Hussein in power – the two main counterarguments offered by the Bush administration. In both cases from 7% to 13% more of the respondents said that they agreed with the first, “the war is not worth it” statement rather then either of the two anti-Kerry propositions.

This clearly suggests the critical importance of the language that is used during the coming weeks. Many Americans still want to consider the war as “right” even as they conclude that it is increasingly not worth the cost in lives lost and resources wasted.

Turning to the second question – whether the invasion of Iraq has helped or hurt the war on terrorism, the data again shows the degree to which opinion can be shaped by the language that is used. The most striking example can be found in two questions asked in an August 10-15 Harris poll.

When asked if the invasion of Iraq has “strengthened or weakened the war on terrorism” 50% chose the first option while only 40% chose the second. But when the same survey also asked if the invasion “helped to protect the United States from another terrorist attack, 54% said it had not, in contrast to only 43% who thought that it had. From this, it appears that more general, “macho” phraseology (“strengthened” rather then “weakened”) evokes a more pro-administration response, while more practical questions about actually reducing risk produce a less belligerent set of responses.

This is reinforced by an August 20 survey by the Program on International Policy Attitudes (PIPA) which revealed even more clearly that, when asked relatively specific and practical questions about terrorism, the American people do not think the war has actually made them more secure. One question, for example, asked if a better use of US resources would have been a) to invade Iraq and overthrow Saddam Hussein or b) to “use those same resources instead for pursuing al-Qaeda and stabilizing Afghanistan”. In this comparison, 52% chose the latter option while only 39% chose the former. In the same survey 49% said “US involvement in Iraq is creating more terrorists who are planning to attack the US” while only 25% thought the war in Iraq was eliminating such terrorists.

The third issue – how long to keep troops in Iraq – shows a rather different pattern from the first two questions. Regardless of question wording almost all surveys during August and September showed majorities of between 54 and 58% clearly in favor of keeping troops in the country until some form of stability is achieved. Moreover, unlike the other questions, this majority sentiment never fell below 50% during the spring of 2004 in most surveys or ever declined to near-parity with any “bring the troops home” alternative.

This suggests that the resistance to withdrawing U.S. troops until some form of stability is achieved is a more stable and deeply rooted opinion then some of the others. It likely reflects a common American attitude that, having committed the soldiers to combat, they should remain until “the job is done”. A similar kind of sentiment was evident during the Vietnam War in the 1960’s and also during France’s colonial occupation of Algeria and led in both cases to a “casualty paradox”. Increasing numbers of dead and wounded, rather then stimulating a desire and demand for withdrawal, led to a growing intransigence and insistence on the application of additional military force to “finish the job”.

There are a variety of conclusions that can be drawn from this data, but the most important is that Kerry cannot win the election if he only garners the support of those who think the Iraq war was “wrong”, who believe it has made us more vulnerable to terrorism and who feel that the troops should be brought back home even in the absence of a stable government.

On the contrary, to build a winning coalition Kerry also has to reach out to a substantial number of the people who, the data show, think the war is, in some meaningful sense, “right”, who believe (or at any rate would like to believe) that our attack on Iraq represented in some undefined way a retaliatory strike against a global enemy called terrorism and who believe that we should stay in Iraq until we “finish the job”

This demands an extremely difficult balancing act. Mobilizing the first group demands an aggressive style of campaigning, one that embodies and expresses many Democrats deep frustration and passionate disagreement with Bush’s actions in Iraq. At the same time, attracting a significant segment of the second group requires Kerry to present himself as more mature, thoughtful and experienced leader who will do a better job of putting together a long-range, international plan for restoring Iraq to stability.

It will be extremely difficult to simultaneously satisfy both these audiences but the degree to which Kerry succeeds in doing so will substantially determine the outcome, not only of the coming debates, but of the 2004 election.

Race Still Tight in Swing States

Kerry leads Michigan LV’s by 4% in FOX News/Opinion Dynamics Poll (Sept. 21-22)

Kerry lags by 2% Nevada RV’s in CNN/USA Today/Gallup Poll (Sept. 18-21)

Bush leads by 3% Ohio LV’s in FOX News/Opinion Dynamics Poll (Sept. 21-22)

Kerry up 5% among Pennsylvania LV's in FOX News/Opinion Dynamics Poll (Sept. 21-22)

Bush ahead by 10% West Virginia RV’s in CNN/USA Today/Gallup Poll (Sept. 17-20)

Bush up by 6% among Wisconsin RV’s in ABC News Poll (Sept. 16-19)

Kerry, Bush in Statistical Tie in Pair of Iowa Polls

Bush leads Kerry 48-45 percent among Iowa LV’s, with 1 percent Nader and 6 percent unsure, according to a FOX News/Opinion Dynamics Poll conducted Sept. 21-22.

Bush leads Kerry 47-46 of Iowa LV’s, with 6 percent unsure, according to a poll by Research 2000 for KCCI-TV conducted Sept. 19-21.

Kerry, Bush Run Close in Two Florida Polls

Bush leads Kerry 48-47 percent of Florida RV’s, with neither/unsure 6 percent, according to a CNN/USA Today/Gallup Poll conducted Sept. 18-22.

Bush leads Kerry 48-43 percent of Florida RV’s, with 2 percent someone else, 1 percent wouldn’t vote and 6 percent unsure, according to a Quinnipiac University Poll conducted Sept. 18-21.

SurveyUSA Polls Show Close WH Race in Key Swing States

A series of polls of LV's conducted September 19-22 by SurveyUSA show:
Kerry ahead by 10% in Michigan and 5% Washington State
Bush ahead by: 1% Maine; 4% Iowa; 13% Tennessee and 1% Oregon.

September 23, 2004

About That Vanishing Gender Gap

On Wednesday, Katherine Seelye solemnly informed us in The New York Times that "Kerry in a Struggle for a Democratic Base: Women". Today, Lois Romano piled on and claimed in the The Washington Post that "Female Support for Kerry Slips", which was particularly heavy with theorizing about the importance of the alleged "security mom". Both articles, however, were thin on data and logic, never stopping to ask themselves whether, if Kerry has less support than he did, say, six weeks ago, whether he has actually lost more support among women than among men. And neither bothered to look at how women stack up to men in terms of being politically motivated and supportive of the president because of security issues.

Nah. Too complicated.

First, let's take the issue of how much women's support of Kerry has changed compared to men's. Here's what Philip Klinkner, Professor of Government at Hamilton College and contributor to the political science blog, PolySigh, has to say:

[Wednesday's] NYT reports that the Kerry campaign is worried about losing support among women. The story however provides no evidence that the gender gap is closing. The only numbers cited show that Kerry is running worse among women than Gore did in 2000, but Kerry's also been running worse than men. In fact, there's no evidence that Kerry has lost any more support among women than among men. Bush, at least until recently, has been running better than Kerry, so that almost inevitably means that he'll be doing better among women as well as among men. But is there any evidence that Bush is closing up the gender gap? In fact, to paraphrase Mark Twain, reports of the demise of the gender gap are greatly exagerated.

In 2000, exit polls showed that Democratic vote for women was 54-43 and for men it was 42-53. Calculating the gender gap as the women's Democratic vote minus the men's Democratic vote divided by two, that gives a gender gap of 11 points [((54-43)-(42-53))/2]=11

In the latest (early September) WaPo poll, the Democratic vote was 49-43 among women and 39-57 among men. That's a gender gap of 12 points...a bit higher than in 2000. The latest ARG poll shows women voting Democratic 50-42 and men 42-51. That's a gender gap of 8.5 points, a bit lower than in 2000. Both polls suggest that the gender gap is running about where it was 4 years ago.

Furthermore, there's no evidence that Bush's recent surge came more from women than from men. The early August WaPo poll showed Kerry leading Bush 50-44. The early September poll had Bush up by the same 50-44 margin. That means Bush picked up a net of 12 points (he went from down 6 points to up by 6 points). In August, Kerry was running 56-39 among women and 44-50 among men. That means Bush picked up a net of 11 points among women (from down 17 to down 6) and a net of 12 among men (from up 6 to up 18). Thus, the Bush surge came equally from men and women.

Well, that takes care of that. Now on to the security moms nonsense. Pollster Anna Greenberg is all over this one, as you can see from checking out her excellent memo, "The Security Mom Myth". Here are some excerpts from the memo:

Sixty-four percent of women voters are married, but only 43 percent have children under 18 years of age. This means that only 26 percent of all women voters could be characterized as “security moms.” If we narrow the analysis to white women, this number goes down to 22 percent of all women voters. Women are diverse and trying to characterize them as a monolithic group with unified set of political views misses the mark. As we know, there are huge differences among women voters, just look at the marriage gap between married and unmarried women. Kerry currently wins unmarried women by 22 points and loses married women by 4 points.

...Women tend to be more worried about their personal and economic security than men, which is not surprising because they are more likely to be victims of crime at home and they are more likely to live on the economic margins. But this concern about personal security does not necessarily translate into political preferences. In fact, men are much more likely to make the war on terrorism and security a part of their voting calculus.

...Since 9/11, Bush has garnered strong ratings on security from everybody...It is important to note, however, that men give Bush stronger ratings on all matters of security than women: they prefer Bush on the war on terrorism to Kerry by 26 points while women only prefer Bush by 9 points; they prefer Bush on foreign policy by 9 points, while women are evenly divided between the candidates; and they prefer Bush on being respected in the world by 10 points while women prefer Kerry by 13 points.

Women are more skeptical, moreover, about the situation in Iraq than men and women (55 percent) are more likely than men (47 percent) to say that the war in Iraq has made us less secure.

Well, you know what they say: you can't believe everything you read in the papers. And the Times and the Post have provided us with a couple of fine examples.

Dead Heat in New Democracy Corps Poll

John Kerry and George Bush are tied at 49 percent of nation-wide LV's, with 1 percent other in a poll conducted Sept. 19-21 by Greenberg Quinlan Rosner Research for Democracy Corps.

Kerry Holds Narrow Lead in Economist Poll

John Kerry leads George Bush 46-45 percent of nation-wide RV's, with 1 percent for Nader in an Economist/YouGov Poll conducted Sept. 20-22.

New WSJ/NBC News Poll Shows Tight Race

Bush leads Kerry 48-45 percent among nation-wide RV's, with 2 percent for Nader, 1 percent for none/other and 4 percent unsure, according to a Wall St. Journal/NBC News poll conducted Sept. 17-19.

September 22, 2004

NDN Surveys Show Democrats Need Clarity and High Hispanic Turnout

The New Democrat Network (NDN) has two new surveys that each provide a different important insight into the dynamics of the 2004 campaign.

The two surveys are a nationwide poll of likely voters by Penn, Schoen & Berland (PSB) and a survey of Hispanic RVs by Bendixen and Associates that included 800 interviews in Florida and 600 interviews each in the three southwestern states of Arizona, New Mexico and Nevada.

The PSB poll has a number of interesting findings and I urge you to read the poll memo on the NDN site. But the most revealing finding is this: when asked whether Bush, the Republican party, the Democratic party and Kerry, respectively, “has a clear agenda for the future of the nation”, this is what voters had to say:

Bush: 57 percent yes/40 percent no
Republican party: 56 percent yes/39 percent no
Democratic party: 45 percent yes/50 percent no
Kerry: 41 percent yes/52 percent no

I detect a pattern here and it ain’t good for the Democratic campaign. It’s not enough to criticize Bush; voters need to believe Democrats have a clear and positive alternative.

The Bendixen polls of southwestern states show Kerry running even more strongly than he was in NDN's polls of the same states in April. Kerry's lead among southwestern Hispanics is now 63-30, slightly larger than Gore received in these three states in the 2000 election.

The Bendixen data also finds that southwestern Hispanics prefer the Democrats over the Republicans by wide margins on issues like: helping you and your family live a better life (+29); being committed to public education (+27); creating a large number of new jobs (+32); and supporting universal health care (+37).

And Bush and the Republicans are viewed ever more negatively by these voters. In Arizona, the Republican party image is down from 48 percent positive/44 percent negative to an abysmal 32/52 and among all southwestern Hispanics the Republican party image has declined from 42/42 to a net -16 (34/50). Similarly, Bush's job rating among Arizona Hispanics has sunk to 38/58 from 42/53 and among all southwestern Hispanics from 41/53 to 39/58.

Finally, the poll of Florida Hispanics finds Democrats and Kerry making some gains compared to earlier in the year, including among Cuban-Hispanics who typically vote very heavily Republicans. Bendixen's estimtes indicate Democrats could do better this election among Florida Hispanics than they did in 2000, which could be key to delivering the state for Kerry.

New ARG Polls Have the Race Very, Very Close

ARG is polling all 50 states and DC on the race for president and today they released the final 30 or so of these polls. Here's how ARG summarized what their data had to say (but check out all their data at their website):

How close is the race for president?

George W. Bush is at 47% and John Kerry is at 46% in the weighted national popular vote.

Bush leads outside the margin of error in 17 states with 133 electoral votes.

Kerry leads outside the margin of error in 10 states with 132 electoral votes.

Bush has any lead in 29 states with 253 electoral votes.

Kerry has any lead in 20 states with 270 electoral votes.

Bush and Kerry are tied in Wisconsin and West Virginia.

Bush needs to defend small leads in 5 states - Colorado, Iowa, Nevada, New Hampshire, and Ohio.

Kerry needs to defend small leads in 5 states - Maine, Florida, Minnesota, Oregon, and Pennsylvania.

Among men nationwide, 51% say they would vote for Bush and 42% say they would vote for Kerry.

Among women nationwide, 42% say they would vote for Bush and 50% say they would vote for Kerry.

And here's a brief take from Alan Abramowitz on these data:

The results of the new 50-state ARG poll appear to be very much in line with other recent polls in the same states. Using today's National Journal Scoreboard, I found other polls conducted by independent, nonpartisan, or bipartisan polling organizations during approximately the same time period in 22 states. On average, support for George Bush was 0.6 points lower and support for John Kerry was 1.8 points higher in the ARG poll compared with the poll conducted closest in time to the ARG poll.

ARG's conclusion that the national race is a virtual dead heat is consistent with the results of other recent state polls conducted by independent, bipartisan, and nonpartisan polling organzations.

Alan promises a more detailed analysis on Friday, so look for his further thoughts then.

Kerry Up by 8 in Washington State Poll

John Kerry leads George Bush 49-41 among Washington State RV's, with 2 percent for Nader and 8 percent unsure, according to a poll conducted Sept. 17-20 by Ipsos-Public Affairs for The Columbian.

September 21, 2004

Was the Republican Convention Bigger Than 9/11?

Those who defend the sudden tilt toward the Republicans in registered voter samples as a real political trend and not any kind of sampling problem, like to point toward the post-9/11 period as an example of a recent shift in the party ID distribution. If it happened then, they say, why shouldn't we give full credence to the shift we're seeing now?

But there is a very serious problem with this logic. After 9/11, despite the immensity of the rally effect behind the president and his party, the shift in party ID toward the Republicans was substantially less than what we're seeing now. What polls showed then was not a shift toward a 4-5 point (or more) Republican advantage in party ID--like we're seeing in some current polls--but rather a simple reduction in the Democratic edge or at best parity. Moreover, even this modest shift took place over several months, rather than over several weeks, like the shifts we've seen in some recent polls. (Note: these Republican gains were given back in a year or two, so the Democrats this year have had their a party ID edge at about the same level they had in 2000 and early 2001.)

Given this, how believable is it that we would now be getting not a gradual reduction in the Democratic party ID advantage (as we did after 9/11) but a much more sudden, much larger shift in party ID to produce an actual Republican advantage of 4-5 points or more? Are we really to believe that the GOP convention was such an earthshaking event that it had a bigger effect on the underlying sentiments of the electorate than did 9/11 and Bush's six months of 80+ approval ratings in the post-9/11 period?

And this from a convention that poll data said was viewed with a distinct lack of enthusiasm by the public! According to the Gallup poll, Bush's acceptance speech, which the media fawned over so ostentatiously, was not rated any better by the public than was Kerry's--in fact, it received slightly worse ratings. Kerry's acceptance speech was rated excellent by 25 percent and good by 27 percent; Bush's was rated excellent by 22 percent and good by 27 percent.

In terms of whether the Republican convention made voters more or less likely to vote for Bush, there were almost as many saying the convention made them less likely to vote for Bush (38 percent) as said it made them more likely (41 percent).

That was actually quite a poor performance. The Democratic convention this year had a substantially better 44 percent more likely/30 percent less likely split. In fact, looking back to 1984, which is as far back as Gallup supplies data, no candidate has ever had a more likely to vote for/less likely to vote for split even close to as bad as Bush's this year.

Well, what about the tone of the convention? Did voters think the Republicans got that one right? Nope. Just 39 percent thought the GOP maintained the right balance between criticizing the Democrats and saying positive things about themselves, compared to 50 percent who thought they spent too much time criticizing the Democrats. By contrast, in 2000, 45 percent thought the GOP maintained the right balance in their convention, compared to 38 percent who thought they spent too much time criticizing.

Can anyone seriously maintain, then, that this year's GOP convention was such a blockbuster that it could produce a surge in Republican party ID that dwarfs that produced by 9/11? It just does not compute.

Still not convinced that party-weighting should at least be considered to correct for sudden partisan imbalances in polls? I close with the words of Charlie Cook in his latest online column:

...Pollsters acknowledge variances from one poll to the next in gender, race, income and education, and they correct for it, but refuse to acknowledge that partisan numbers fluctuate just the same, and need to be corrected.

My own view is that samples should be weighted by party to the average party breakdown in a combination of the polls for the last several months, linking it to a very large sample of combined surveys to reduce sampling error. While this method might be a bit sluggish if party identification is changing dramatically, it would mean that when a candidate is gaining or dropping, it is most likely because they really are, not because of a sample that is too tilted in favor of one party or the other. If Republicans are indeed gaining in party identification, it will show up after a couple of polls in the average.

You tell 'em, Charlie.

New Polls in Iowa, Michigan, Ohio and Oregon

Bush leads Kerry 48-43 percent among Iowa RV's, with 3 percent for Nader and 6 percent for neither in a CNN/USA Today/Gallup Poll conducted Sept. 16-19, 2004.

Kerry leads Bush 48-44, with 2 perent for Nader and 6 percent unsure in a poll of Michigan LV's by EPIC/MRA conducted Sept. 15-19, 2004.

Bush leads Kerry 54-43 percent among Ohio LV's, with 2 percent for Nader and 1 percent unsure, according to the Ohio Poll conducted Sept. 12-18 by the Institute for Policy Research of the University of Cincinnati.

Kerry leads Bush 51-44 among Oregon LV's, with 5 percent unsure in a poll by Research 2000 for The Portland Tribune, et al. conducted Sept. 13-16.

New National Poll Shows Statistical Tie in WH Race

Bush leads Kerry 44-43 percent in a head-to-head match-up of nation-wide RV's, with 13 percent unsure, according to a Investor's Business Daily/Christian Science Monitor/TIPP poll conducted by TechnoMetrica Market Intelligence Sept. 14-18, 2004.

Kerry Leads in 11 of 16 Swing States In New Zogby Poll

The Zogby Interactive Poll of LV's was conducted for the Wall St. Journal Sept. 13-17.

Kerry's leads (%): Arkansas 0.1; Florida 0.5; Iowa 3.0; Michigan 6.0;
Minnesota 9.7; New Hampshire 3.6: New Mexico 12.7; Oregon 12.0; Pennsylvania 3.1; Washington 8.7; Wisconsin 2.4

Bush's leads (%): Missouri 5.4; Nevada 2.2; Ohio 3.3; Tennessee 5.5; West Virginia 12.4

September 20, 2004

ARG Vs. Gallup

This month, ARG is polling every state in the presidential contest. Recently, they released the first 20 of these polls and the results looked pretty good for Kerry. How much faith should we put in these polls? How well did their polls turn out in 2000 when they also polled all 50 states?

Alan Abramowitz has looked at ARG's track record and here's what he found:

My analysis of ARG's September, 2000 poll of all 50 states plus the District of Columbia indicates that, in general, the poll was highly accurate. On average, the state by state results yielded an average lead for Al Gore of about 1 percentage point in September. On Election Day, Gore actually lost the average state by an average margin of 3.6 percentage points even though he narrowly won the national popular vote. This is due to the equal weighting in the average of heavily populated states and sparsely populated states. Gore carried 6 of the 9 most populous states while Bush carried 15 of the 20 least populous states.

The 4.6 point shift to Bush between the ARG poll and Election Day can easily be explained by the fact that the poll was conducted at a time when Gore was leading in almost all of the national polls, before the first presidential debate. By Election Day, of course, the race had narrowed to a virtual dead heat.

The ARG September poll accurately predicted the winner of 45 of the 50 states, missing only Missouri, West Virginia, Louisiana, Kentucky, and Tennessee. I am, of course, counting ARG's poll showing Al Gore leading George Bush in Florida as a correct prediction since Gore did actually receive more votes, or at least intended votes, than Bush in Florida.

The correlation between the ARG September poll results and the actual election results is a very impressive .95. This means that the results of the ARG poll explain about 90 percent of the variation in the state-by-state election results.

The ARG 50-state poll, done well before Election Day, provided to be a much more reliable predictor of the actual results of the 2000 presidential election than the Gallup tracking poll which gyrated wildly in the weeks leading up to the election, sometimes showing sizeable leads for Bush, sometimes showing sizeable leads for Gore, and sometimes showing a close race. Indeed, only 10 days before Election Day, Gallup's tracking poll had George Bush leading Al Gore by 13 points--similar to Bush's lead over John Kerry [among likely voters] in Gallup's most recent 2004 poll.

Therefore, we can have a good deal of confidence in the accuracy of ARG's 2004 50-state poll which, based on the first 20 states that have been released thus far, seems to indicate that, contrary to some (but not all) recent national polls, we have a very close presidential contest. Kerry is leading by a comfortable margin in every blue state. More importantly, he is leading in 4 of the 5 battleground states included in the first wave and is only trailing by 1 point in Colorado--a state that George Bush carried by 9 points in the 2000 election. Another very recent newspaper poll in Colorado shows the same thing.

My advice is watch the rest of the ARG state polls as they are released....I predict that they will continue to show a very competitive battlefield in the 2004 presidential election.

ARG releases the rest of their state polls on September 22. Stay tuned.

WH Race Still Close in New National Poll

Bush leads Kerry 47-44 percent among nation-wide LV's, with 7 percent
undecided in a Zogby America poll conducted Sept. 17-19.

New State Polls

Bush Ahead In Poll of Arizona LV’s

Bush leads Kerry 50-39 percent among Arizona LV’s, with 3 percent for another party ticket and 8 percent undecided, according to a survey conducted Sept. 13-14 by Mason-Dixon Polling & Research for Knight Ridder and MSNBC.

Bush Leads By 7 Percent In Two Missouri Polls

Bush leads Kerry 48-41 percent among Missouri LV’s, with 1 percent for Nader, 1 percent for another party ticket and 9 percent undecided in a survey conducted Sept. 14-16 by Mason-Dixon Polling & Research for Knight Ridder and MSNBC.

Bush leads Kerry 49-42 percent of Missouri LV’s, with 9 percent unsure in a head-to-head survey conducted Sept. 13-16 by Research 2000 for The St. Louis Post-Dispatch and KMOV-TV.

Bush Leads Among NH LV’s

Bush leads Kerry 49-40 percent among New Hampshire LV’s, with 3 percent for Nader and 8 percent unsure, according to a survey conducted Sept. 13-15 by Mason-Dixon Polling & Research for Knight Ridder and MSNBC.

Kerry Lags By 7 Among Ohio LV’s

Bush leads Kerry 49-42 percent among Ohio LV’s, with 2 percent for Nader and 7 percent undecided, in a survey by Mason-Dixon Polling & Research for Knight Ridder and MSNBC, conducted Sept. 14-15.

Kerry and Bush Near Tie In WV Heat

Bush leads Kerry 45-44 among West Virginia LV’s, with 11 percent undecided in a survey by Mason-Dixon Polling & Research for Knight Ridder and MSNBC, conducted Sept. 13-14.

A Few More Thoughts on the New CBS News/New York Times Poll

As Chris Bowers of MyDD points out in his insightful new essay, "Rapid Poll Movement is a General Election Myth", the new CBS/NYT poll is actually a lot worse for Bush than the CBS News poll released a week earlier, even though both polls gave Bush a 50-42 lead among RVs. That's because, since the current poll is substantially more Republican than the earlier poll (which actually had a slight Democratic edge), Bush should actually have performed better than the earlier poll on the horse race and on indicators like job approval and right/direction wrong track in this poll, instead of about the same. That also means that if we adjust the current poll to correct the apparent surplus of Republicans, Bush's performance on these indicators should actually decline below the measurements of the earlier poll.

Since CBS News thoughtfully provides the overall result and the result broken down by party ID for each and every question in their survey, it is possible to estimate what Bush's ratings would have looked like if there weren't so many Republicans in the sample. Here are some examples, based on reweighting the current poll to the 2000 exit poll distribution of partisanship:

Overall job approval: 49 percent approval/44 percent disapproval
Economic job approval: 42/52
Iraq job approval: 45/51
Campaign against terrorism job approval: 57/37
Right direction/wrong track: 40/53

In every case, these ratings are worse than they were a week ago, making the idea that the race is tightening up more plausible.

Of course, Kerry needs not just a tight race, but to pull ahead. Given Bush's continued vulnerabilities, which these data highlight, Kerry's got the opening to do so. I'll address this issue in future posts.

September 19, 2004

Wall Street Journal Article Asks All the Right Questions

John Harwood has a front-page article in the Monday Wall Street Journal that does an excellent job of summarizing and discussing the debate that has been unfolding around the problems with recent polls. Of course, here on DonkeyRising I have discussed these problems in detail and defended the proposition that, once these problems are taken into account, the Presidential race is close to tied.

Or, as the article puts it:

If the CBS and Pew surveys are adjusted to reflect comparable numbers of Republicans and Democrats, their results would have been virtually identical.

Indeed that's precisely what liberal polling analyst Ruy Teixeira did on his Web log, called Emerging Democratic Majority. As the New York Times report of the poll carried the headline "Bush Opens Lead," Mr. Teixeira's blog declared, "CBS News/New York Times poll has it close to even."

...Mr. Teixeira argues that the Democratic edge Mr. Kohut [of the Pew Research Center] found is realistic, since exit polls from the 1996 and 2000 campaigns indicated that in both cases four percentage points more Democrats than Republicans showed up to vote. Slightly more self-described Democrats than Republicans voted in the 1984, 1988 and 1992 elections as well.

Nice to see some of the these very important arguments escaping the blogosphere ghetto and entering the mainstream press.

September 17, 2004

CBS News/New York Times Poll Has It Close to Even!

Well, that is if you weight their data to conform to the 4 point Democratic party ID lead which we have good reason to believe is the underlying distribution in the voting electorate. As many have already heard, the new CBS News/New York Times poll, conducted September 12-16, gives Bush an 8 point lead (50-42) among RVs--but also gives the Republicans a 4 point edge on party ID. Reweight their data to conform to an underlying Democratic 4 point edge (using the 39D/35R/26I distribution from the 2000 exit poll) and you get a nearly even race, 47 Bush/46 Kerry.

Nearly even. That goes along with the the 46-46 tie in the Pew Research Center poll (which gave the Democrats a 4 point edge on party ID without weighting) and the 48-48 tie in the Gallup poll (once weighted to reflect an underlying Democratic 4 point edge). Not to mention the two other recent national polls (Harris, Democracy Corps) that show the race within one point.

Perhaps all this is just a coincidence, but the pattern seems striking. Once you adjust for the apparent overrepresentation of Republican identifiers in some samples, the polls all seem to be saying the same thing: the race is a tie or very close to it.

Note: this entry has been revised from the original to correct the CBS reweighted horse race from 46-46 (original) to 47-46 (corrected).

Gallup Strikes Again!

Here are Bush's leads in the three national polls released before Gallup's current poll (no RV data available for DCorps and Harris; Pew and Harris matchups include Nader):

Democracy Corps, September 12-14 LVs: +1
Pew Research Center, September 11-14 RVs: tied
Harris Interactive: September 9-13 LVs: -1

Looks like a tie ball game, right? But according to the Gallup poll conducted September 13-15 and released today, Bush is up......13???

Let's just say I'm just a wee bit skeptical of this one. First, Gallup's poll only includes one day (the 15th) these three other polls do not, so it can't be Gallup's survey dates that explain the big Bush lead.

Second, this 13 point lead is an LV figure and, as I've repeatedly emphasized, Gallup's LV screening procedure produces completely untrustworthy measures of voter sentiment this far in advance of the election. Here is a summary of the case against Gallup's LV data:

Sampling likely voters is a technique Gallup developed to measure voter sentiment on the eve of an election and predict the outcome, not to track voter sentiment weeks and months before the actual election. There is simply no evidence, and no good reason to believe, that it works well for the latter purpose. In fact, the evidence and compelling arguments are on the other side: that the registered voters are the more reliable guage of voter sentiment during the course of the campaign.

Here’s why. Gallup decides who likely voters are based on 7 questions about their interest in voting, attention to the campaign and knowledge about how to vote (e.g., where their polling place is located). The interested/attentive/knowledgeable voters are designated “likely” and the rest are thrown out of the sample. But as a campaign progresses, the level of interest among voters tends to change, particularly among those with partisan inclinations whose interest level will rise when their party seems to be mobilized and doing well and fall when it is not. Because of this, partisans of the mobilized party (lately, Republicans) tend to be screened into the likely voter sample and partisans of the demobilized party (lately, Democrats) tend to get screened out. But tomorrow, of course, the Democrats could surge, in which case their partisans may be the ones over-represented in likely voter samples.

That suggests the uncomfortable possibility that observed changes in the sentiments of “likely voters” represent not actual changes in voter sentiment, but rather changes in the composition of likely voter samples as political enthusiasm waxes and wanes among the different parties’ supporters. And that is exactly what political scientists Robert Erikson, Costas Panagopoulos, and Christopher Wlezien find in their analysis of Gallup's 2000 RV/LV data in their forthcoming paper, “Likely (and Unlikely) Voters and the Assessment of Campaign Dynamics” in Public Opinion Quarterly: “shifts in voter classification as likely or unlikely account for more observed change in the preferences of likely voters than do actual changes in voters’ candidate preferences.”

That means that, instead of giving you a better picture of voter sentiment and how it is changing than conventional registered voter data, likely voter data give you a worse one since true changes in voter sentiment are swamped by changes in who is classified as a likely voter.

I think the case against the Gallup LV data looks rock solid. In my view, it's time for them to drop reporting these data because they are highly likely to give an inaccurate picture of the state of the race and, by doing so--especially given the high profile of Gallup's polls--unfairly pump up one side of the race and demoralize the other. That doesn't seem acceptable to me.

Of course they'll reply: well, our data work so well right before the election, they must be the best data to use all the time. But, for the reasons outlined above, that reasoning is completely specious. And then there's this: the LV data haven't been working so well lately even right before the actual election. In 3 of the last 4 presidential elections (including the last one), Gallup's final RV reading was actually closer to the final result than their final LV reading!

As I say, maybe it's time for a rethink down at Gallup HQ.

Throwing out the Gallup LV data, then, let's move on to their RV result: an 8 point Bush lead. Obviously pretty far off the results of the other contemporaneous polls summarized above, but....could be I suppose.

But then there's this: the Gallup internals show Kerry with a 7 point lead among independent RVs. Huh? Kerry's losing by 8 points overall, yet leading among indenpedents by 7. How is that possible? Only if there are substantially more Republicans than Democrats in the sample.

That suggests that reweighting the sample to reflect the 2000 exit poll distribution (39D/35R/26I) would give a different result. It does: the race then becomes dead-even, instead of an 8 point Bush lead. (Note: Steve Soto of The Left Coaster got Gallup to give him their party ID distributions for this poll and confirms a 5 point Republican party ID advantage in their RV sample.)

One final note: I mentioned the Pew Research Center poll had the race dead-even just like the reweighted Gallup data. And what was Pew's party ID distribution in their RV sample? You guessed it: a 4 point lead (37-33) for the Democrats, just like in the 2000 exits.

I think we've finally found out how to make these polls get along!

Kerry, Bush Running Close in 3 PA Polls

Bush leads Kerry 47-45 percent among Pennsylvania RV's in a head-to-head race, with 8 percent unsure in a Keystone Poll conducted September 8-15 for Franklin & Marshall College/Philadelphia Daily News/CN8.

Bush leads Kerry 47-44 among Pennsylvania RV's in a head-to-head match-up, with 1 percent someone else, 2 percent wouldn't vote and 7 percent don't know in a Quinnipiac University Poll conducted Sept. 11-14, 2004.

Kerry and Bush are tied at 48 percent of Pennsylvania RV's in a head-to-head contest, with 2 percent neither, 1 percent wouldn't vote and 1 percent unsure in an ABC News Poll conducted Sept. 9-12, 2004.

Bush Leads by 1 in Another MN Poll

Bush leads Kerry 47-46 among Minnesota RV's in a head-to-head match-up, with 7 percent neither/other/no opinion, according to a USA Today-CNN-Gallup Poll conducted Sept. 11-14.

Bush Has Small Lead in MN Poll

A poll of Minnesota LV's by Mason Dixon Polling & Research, Inc. for the St.
Paul Pioneer Press and Minnesotra Public Radio conducted Sept. 11-14 has Bush
ahead of Kerry 46-44 percent, with 1 percent for Nader and 9 percent undecided.

Bush Leads in New National Poll

A national USA Today/CNN/Gallup poll conducted Sept.13-15 has Bush leading among RV's 52-44 percent, with 2 percent for neither and 2 percent no opinion.

Kerry, Bush Tied in Major National Poll

A new national poll by the Pew Research Center for the People & the Press
has Kerry and Bush tied at 46 percent among RV's (September 11-14), with 1
percent for Nader and 7 percent other/don't know.

September 16, 2004

Kerry Shows Strength in New ARG Poll

An American Research Group poll of 20 states conducted September 7-13 shows John Kerry leading among LV's in 4 out of 5 designated "swing" states, ME, MN, OR, and WA, with Bush leading only in CO by 1 percent. The polls also indicate Kerry is maintaining substantial leads in "blue" states.

Kerry Seizes Lead in New National Poll

A Harris poll of nation-wide LV's conducted Sept. 9-13, has Kerry leading Bush 48-47 percent, with 2 percent for Nader and 3 percent not sure/refused.

White House Race Narrows in New National Poll

A head-to-head poll of nation-wide LV's by Greenberg Quinlan Rosner Research conducted 9/12-14 for Democracy Corps has Bush ahead of Kerry 49-48 percent with 1 percent going for other.

Persuadable Voters Still Not Persuaded

Even as Bush has opened up a small lead in the national polls--and I've tried to offer some evidence in my posts that this lead is underwhelming--evidence has been accumulating that he remains weak among the kind of independent and swing voters he needs to form an electoral majority.

In fact, the Annenberg Election Survey has just released data that indicate, while Bush made some small gains among the overall electorate when comparing the pre-GOP convention and post-GOP convention periods, he has actually lost ground among "persuadable voters" (those voters who are undecided or who say there is a "good chance" they could change their mind about the candidate they currently support).

For example, Bush's favorability rating fell from 47 percent favorable/30 percent unfavorable among persuadables in August (August 9-29) to 43/33 after the GOP convention (September 3-12). And Kerry's rating among this group actually has gone up: from 36/25 to 43/25 (now somewhat better than Bush's).

In addition, Bush's overall job rating among persuadables is now 44 percent approval/49 percent disapproval; his job rating on the economy is 32/63 and his job rating on Iraq is 34/59. Even his job rating on terrorism is only 50/41. And all of these ratings are now lower among persuadables, not higher, than they were in August

Bush has also lost significant ground among persuadables since August on some key candidate characteristics including "cares about people like me", "shares my values", "out of touch with people like me", "stubborn" and "arrogant". These voters are now more liikely, not less likely, to think the positive attributes apply to Kerry and the negative attributes to Bush.

In short, the persuadables aren't persuaded and appear to be ripe for Democratic gains. What's the key? One possibility is Iraq. Persuadables are now less convinced than ever that Bush has a clear plan for bringing the situation in Iraq to a successful conclusion--just 17 percent now think so. Unfortunately, only 15 percent of persuadables think Kerry has such a plan--not much of a difference and not even one in Kerry's favor.

Make that difference a big one in Kerry's favor and Bush's weakness among persuadables could translate into big gains for the Democratic ticket.

September 15, 2004

Kerry Holds Strong Lead in MN

A new Minneapolis Star-Tribune Poll conducted Sept. 10-13 has John Kerry leading George Bush 50-41 among Minnesota LV's, with 1 percent for other and 8 percent unsure.

Kerry Up by 7 in Michigan

John Kerry leads George Bush 50-43 percent among Michigan RV's, with 1 percent for Nader and 6 percent neither/unsure, according to a CNN/USA Today/Gallup Poll conducted Sept. 10-13.

Now Available: Latest Democracy Corps Analysis By Stanley Greenberg and James Carville

The latest strategy analysis by Democracy Corps is now available on the D-corps website. The report is based on Democracy Corps September 6-9 survey of 1004 likely voters and features an extended analysis by leading strategists Stanley Greenberg and James Carville of where the campaign now stands and what the polling data suggests Kerry should do. Here are a few key excerpts:

1. The president has a lead of about 5 points, if we look at the average of all the public polls done after the convention; the Democracy Corps poll completed last Thursday shows the president with a 3-point lead. In any case, Bush is at 49 percent with the former estimate and 48 in our survey. At the height of Bush’s convention bounce, he is just at the edge of electability. His position is simply not that strong. If his bounce recedes, as it did for Kerry, and if Kerry takes the race to Bush, the president could easily be endangered again.

2. A majority of the country still wants change. By 53 to 41 percent, voters think the country is headed in the wrong direction; by 51 to 46 percent, they want to go in a significantly different direction than Bush. The mood for change is even greater among independents (57 to 35 percent) and battleground state voters (55 to 40 percent). 14 percent of the electorate is comprised of Bush voters thinking things are going wrong.

3. Part of the stability of this race are the grave doubts voters feel about George Bush. These doubts about Bush are largely undiminished in the last month: Bush favors corporate interests over the public interest (60 percent serious doubts), is too ready to go to war (58 percent), spending too much abroad and neglecting home (54 percent), and made mistakes in Iraq that shortchange America at home (54 percent).

4. While Bush has gained on having plans for Iraq and it bringing more security, a stable majority says the war was not worth it. Just 43 percent believe we are making progress there, and this is before recent developments in Iraq, with increased fighting.

5. And on the economy, voters by nearly two-to-one reject Bush’s assertion that the middle class is making gains .... Bush could not be more out of sync, even as he is compelled to make the case for progress.

Bush Ahead in New Florida Poll

A SurveyUSA poll of Florida LV's for TV stations in six cities conducted Sept. 12-14 has Bush at 51 percent, Kerry at 45 percent with 3 percent undecided, in a head-to-head match-up.

Kerry Lags by 4% in Nevada

Bush has 51 percent, Kerry has 47 percent, with 2 percent undecided in a head-to-head SurveyUSA poll of Nevada LV's conducted Sept. 11-13 for KVBC-TV Las Vegas.

September 14, 2004

A Few More Thoughts on the State of the Race

Well, those cards and letters keep coming in, so I thought I'd respond to a few of the most common questions that have been posed to me.

1. How can you deny that Bush is ahead?

I don't. My view is that he is currently ahead, but only modestly, contrary to the tone of media coverage and the findings of some polls. I have tried to explain the reasoning behind this assessment, especially as it pertains to possible problems with contemporary polls.

It's worth noting that the latest poll data on RVs--ending the night of the 12th--have Kerry up by 2 (IBD/CSM/TIPP) or Bush up by 4 (ICR). That averages out to a 1 point Bush lead, even without party-weighting the data. And Rasmussen LV data for the period ending the 12th also has Bush with a one point lead.

2. How is it possible for samples of RVs to suddenly have too many Republican identifiers? Aren't voters just shifting their party identification?

It is certainly possible that we gone from, say, a 4-5 point Democrtic lead in party ID to a 4-5 point Republican lead in the space of the last month. But color me skeptical about this 8-10 point swing in a few short weeks.

A better explanation for this sudden shift in poll samples, in my view, is that when the political situation jazzes up supporters of one party, they are more likely to want to participate in a public opinion telephone poll and express their views. An increased rate of interview acceptance by that party’s supporters would then skew the sample toward that party without the underlying distribution having changed very much, if at all.

In this case, the Republican convention, coming on the heels of the Swift Boat controversy, may have helped raise political enthusiasm among Republican partisans, leading to more interview acceptances and a disproportionate number of Republicans in recent samples.

Do I know this for sure? No, I don't, because we lack direct evidence that this is happening, just as we lack direct evidence that individual voters are suddenly and massively shifting their party allegiance. But I do know which of these explanations I find more plausible and consistent with other evidence about the general stability of party ID.

3. Does your analysis mean we shouldn't be worried about the Kerry campaign?

No, by all means worry (though don't get carried away with it). Bush is ahead, if not by much, and has clearly regained the political initiative. Kerry has to push things back in the other direction. I agree with Josh Marshall that key here for Kerry is probably Iraq. The situation is going to hell in a handbasket, but Kerry so far has not been able to convince voters he has a strong alternative. That needs to change.

Now you may return to your regularly scheduled panicking.

Bush Leads in Wisconsin

In a head-to-head CNN/USA Today/Gallup Poll of Wisconsin RV's conducted Sept. 9-12, 2004, Bush leads Kerry 50-45 percent, with 5 percent neither/unsure.

Bush Leads in ICR Poll

In a head-to-head ICR/International Communications Research poll of nation-wide RV's conducted Sept. 8-12, 2004 Bush leads Kerry 49-45, with 5 percent neither/unsure.

Kerry Takes Lead in New National Poll

John Kerry leads George Bush 46-44 percent in a head-to-head match-up among nation-wide RV's, with 10 percent not sure, according to an Investor's Business Daily/Christian Science Monitor/TIPP poll conducted by TechnoMetrica Market Intelligence Sept. 7-12, 2004.

September 13, 2004

Why the Race Is Closer Than People Think

Is Bush ahead by a little or a lot? Is it close to a tie ball game or has Bush surged to a commanding lead?

The conventional wisdom inclines to the latter not the former. The reason has a great deal to do with two persistent problems with contemporary polls that--at least at this point in time--tend to considerably inflate Bush's apparent lead. But once you dissect the available data with these problems in mind, a truer picture of the race comes into focus which suggests that the race continues to be very close.

The two problems are: (1) samples that have an unrealistic number of Republican identifiers and hence tend to favor Bush; and (2) the widespread and highly questionable practice of using likely voters (LVs) instead of registered voters (RVs) to measure voter sentiment this far before the election.

First, the issue of partisan distribution in samples. Lately, and very suddenly, many polls have been turning up more Republican identifiers than Democratic identifiers in their samples--in some cases, many more (as high as a 9-10 point Republican advantage).

How realistic is it to be suddenly turning up a Republican lead on party ID, much less a large one? Not very. The weight of the academic evidence is that, while the distribution of party ID among voters can and does change over time, it changes slowly, not in big lurches from week to week.

And the weight of the empirical evidence is that the distribution of party ID among voters has favored and continues to favor the Democrats. In 2000, the exit polls showed Democrats with a 4 point advantage over Republicans. In 1996, it was also 5 points; in 1996, it was 3 points and in 1988 it was also 3 points.

The data also indicate that there were two shifts in party ID over the 2001-04 period which largely cancelled each other out. The first shift, in the period after 9/11, shaved several points off the Democrats' lead and brought the Republicans close to even (but never ahead) in party ID. The second shift tooks place in late 2003 and 2004 and reconstituted the Democrats' lead on party ID to about 4 points, exactly where it was in the 2000 election according to the exit polls (see this useful study "Democrats Gain Edge in Party Identification" by the Pew Research Center for more details).

So, if polls are suddenly turning up too many Republican identifiers and that is unrealistic and skews reported horse race results toward Bush, what, if anything, should be done?

One possible solution is to weight poll results by a more reasonable distribution of party ID. The issue of whether to use this approach to the problem is well-summarized by Alan Reifman in his invaluable essay "Weighting Pre-Election Polls for Party Composition: Should Pollsters Do It or Not?" on his website.

As Reifman puts it:

One factor (among many) that may contribute to discrepancies between different outfits' polls in their Bush-Kerry margins....is polling firms' different philosophies as to whether it's advisable to mathematically adjust their samples -- after all the interviews have been completed -- to make the percentages of D's and R's in their survey sample match the partisan composition that is likely to be evident at the polls on Election Day. The latter can be estimated from exit polls from previous elections, party registration figures (in states where citizens declare a party ID when registering to vote), and surveys.

(Another issue that often comes up in evaluating pre-election surveys, with which many of you may be familiar, is whether results are reported for "registered" or "likely" voters. That is a different issue from what is being discussed [in this essay]. Whether a pollster reports results for registered voters, likely voters, or both, weighting by party ID is a separate, independent decision.)

Note well Reifman's point that the issue of whether and how to use LVs, not RVs, to report results is separate from the issue of whether and how to do party-weighting. I discuss the LV issue below after the party-weighting discussion.

Given that party ID does shift some over time, my instinct has generally been to avoid party-weighting if possible and promote a full-disclosure approach. This is how I put it in a recent post:

[B]ecause the distribution of party ID does shift some over time....polls should be able to capture this. What I do favor is release and prominent display of sample compositions by party ID, as well as basic demographics, whenever a poll comes out. Consumers of poll data should not have to ferret out this information from obscure places--it should be given out-front by the polling organizations or sponsors themselves. Then people can use this information to make judgements about whether and to what extent they find the results of the poll plausible.

But this approach increasingly seems unrealistic to me. The polling organizations and sponsors do not routinely release the data I call for and certainly do not prominently display them. And even if they did, the typical consumer of polling data lacks the time and skills to use these data to re-weight or adjust reported results. The fact of the matter is that people pay attention to reported results period; therefore they are at the mercy of whichever results are reported and emphasized (an issue that also looms large in the LVs vs. RVs issue, discussed below).

This suggests that weighting poll results by a reasonable distribution of party ID may be necessary to avoid giving the public distorted impressions of the state of the race.

What is a reasonable distribution of party ID to use in such weighting? One obvious candidate is the exit poll distribution from 2000: 39D/35R/26I. Moreover, the Democratic advantage in this distribution--4 points--closely matches the average Democratic advantage in 2004, as measured by the Pew Research Center (see above) and other polling organizations, making it an even more attractive option.

But political analyst Charlie Cook probably has the best idea, even though it can really only be implemented by the polling organizations themselves: "dynamic party identification weighting". Cook's idea is that polls should weight their samples by a rolling average of their unweighted party ID numbers taken over the previous several months. This would allow the distribution of party ID to change some over time, but eliminate the effects of sudden spikes in partisan identifiers in samples (such as we are experiencing now).

Lacking such a dynamic weighting, however, the best we can probably do at this point is to use the exit poll distribution mentioned above. How much difference would this make if we applied it to recent polls?

Quite a bit. Here are Bush's leads in a number of recent polls, ordered by size of his lead, once the horse race question is weighted by the 2000 exit poll distribution (note: not all recent polls can be included because you need the horse race figures among Democrats, Republicans and independents separately to do this procedure and not all polls release these figures; in addition Zogby and Rasmussen results are party-weighted to begin with and therefore do not have to be re-weighted; RV results used unless only LV results available):

CBS News, September 6-8 RVs: +5
Zogby, September 8-9 LVs: +2
Rasmussen: September 10-12 LVs: +1
Fox News: September 7-8 LVs: +1
Washington Post, September 6-8 RVs: +1
Newsweek, September 9-10 RVs, -2
Gallup, September 3-5 RVs: -4

These data present a clear picture of a tight race, with Bush likely running a small lead, but not the solid--and even large--advantage that has been conveyed to the public.

The other problem that is afflicting the polls and considerably inflating perceptions of Bush's lead is the widespread, and highly questionable, use of LVs, instead of RVs, to report horse race results far in advance of the actual election. The reason why using LVs instead of RVs is a bad idea is simple: the LV approach is being asked to do a job--gauge voter sentiment and how it changes from week-to-week (and even day-to-day)--that it was never designed to do. What the LV approach was designed to do was measure voter sentiment on the eve of an election and predict the outcome. That was, and remains, an appropriate application of the LV approach.

But applied as many polling organizations currently do, it is highly inappropriate and frequently very misleading. As political scientists Robert Erikson, Costas Panagopoulos and Christopher Wlezien put in in their important forthcoming paper, "Likely (and Unlikely) Voters and the Assessment of Campaign Dynamics" in Public Opinion Quarterly:

[E]stimates of who may be likely voters in the weeks and months prior to Election Day in large part reflect transient political interest on the day of the poll, which might have little bearing on voter interests on the day of the election. Likely voters early in the campaign do not necessarily represent likely voters on Election Day. Early likely voter samples might well represent the pool of potential voters sufficiently excited to vote if a snap election were to be called on the day of the poll. But these are not necessarily the same people motivated to vote on Election Day.

And of course, since the group of people "sufficiently excited to vote if a snap election were to be called on the day of the poll" changes from poll to poll, it raises the uncomfortable possibility that observed changes in the sentiments of "likely voters" represent not actual changes in voter sentiment, but rather changes in the composition of likely voter samples as political enthusiasm waxes and wanes among the different parties' supporters. Or, as Erikson et. al. put it:

At one time, Democratic voters may be excited and therefore appear more likely to vote than usual. The next period the Republicans may appear more excited and eager to vote. As Gallup’s likely voter screen absorbs these signals of partisan energy, the party with the surging interest gains in the likely-voter vote. As compensation, the party with sagging interest must decline in the likely-voter totals.

And this is exactly what their analysis of Gallup data from the 2000 election finds--"shifts in voter classification as likely or unlikely account for more observed change in the preferences of likely voters than do actual changes in voters’ candidate preferences".

This is an important result and helps nail down what has always been disturbing about the use of likely voter methods far in advance of the actual election. Instead of giving you a better picture of voter sentiment and how it is changing than conventional RV data, it gives you a worse one since true changes in voter sentiment are swamped by changes in who is classified as a likely voter.

Does this matter? You bet it does. When Gallup told the world on September 6 that Bush was leading Kerry by 7 points among LVs, the world listened and absorbed that figure as a trustworthy indicator of where the race was. Completely lost, except to those who bother to look at such things, was the Gallup finding that Bush only led by single point among RVs--in other words, that the race was about tied. Gallup and its sponsoring organizations implicitly and explicitly encouraged people to treat the LV finding as the real story and the RV finding as an unreliable afterthought (after all, those voters aren't "likely"!). The incredible irony, of course, is that the real situation was exactly the reverse: as the Erikson et. al. findings suggest, it was the RV data that provided the best gauge of voter sentiment and the LV data that should have been an unreliable afterthought.

Or take the Gallup data gathered in Ohio in the last two months, perhaps the key state in this election and the subject of endless media stories about "the battle for Ohio". On September 8, Gallup released data showing Bush ahead of Kerry by 8 points among LVs in Ohio, a 14 swing from late July when Kerry led by 6. Again, completely lost in the Gallup, newspaper and television reports on the poll was the poll's finding that Bush had just a 1 point lead among RVs in the state, representing a much more modest swing of 6 points since late July.

Guess which figures are still with us as coverage of the battle for Ohio continues? That's right: Bush's 8 point lead among LVs and 14 point swing. In fact, just this Sunday, The New York Times practically built their Ohio campaign story around these figures which showed just how well Bush is doing! and just how much the situation has changed!.

In short, these LV figures, especially from Gallup, are contributing mightily to the impression that Bush has built a substantial lead and is even surging ahead in some of the key swing states. But, as we have seen, these LV data are fundamentally inappropriate for measuring the state of the race, and how it is changing, this far ahead of election day. For that, you need the RV data and they suggest something far different: the race is damn close and Bush's substantial lead is a myth.

September 12, 2004

Fox Poll Shows Kerry Lead in Battleground States

The recent Fox News poll of LVs conducted Sept. 7-8 shows Kerry leading Bush by 5 points (48-43) in key "battleground states" (AZ, AR, FL, IA, MI, MN, MO, NV, NH, NM, OH, OR, PA, WV, WI).

Update: The above entry has been corrected from the original as per comments.

Dead Heat in Maine

A September 9th Zogby International Poll of Maine LV's for the Portland Press Herald and the Maine Sunday Telegram has Kerry tied with Bush at 43 percent, with 3 percent for Nader and 10 percent undecided.

September 11, 2004

Strategy Notes:
John Belisarius

How much damage did the Swift Boat attacks really do - and what can Dems learn from them?

It would be hard to find a Democrat right now who doesn't think that the attacks during August by the Swift Boat critics were extremely damaging to John Kerry and that a much earlier and more combative response to the accusations would have substantially reduced their impact. With reports of new smear attacks on the horizon, many democrats consider the two generalizations above as the unquestioned basis for planning the democratic response.

But, surprisingly, the evidence for both of these conclusions is not as clear as it might seem. The analyses that have advanced this view have generally said something like the following: "Well, Kerry's poll support fell during August and the Swift Boat attacks were the biggest news story that month so the one must have caused the other"

For example, Newsweek's summary of its most recent poll argued,
Kerry's campaign, on the other hand, may have been hurt by the television ad campaign launched against him by Vietnam veterans who question his record. Just 45 percent of all voters view Kerry favorably (down from 53 percent in late July) and 46 percent view him unfavorably (up from 37 percent).

This direct jump from the swift boat ads to the level of Kerry's popularity is, to put it mildly, a pretty breathtaking leap of statistical inference (and one that is probably not a good idea to try out in Stat 101). But, if you look closely, many commentators argue along pretty much the same kinds of lines.

Commentators get away with this kind of "seat of the pants" thinking because, like most people, they rely heavily on gut, instinct and experience to reach conclusions and frequently use polling data more like decoration for their ideas then as actual proof.

But if Democrats want to make serious use of public opinion data in planning their political strategy, they need to begin by examining the data in a different way, focusing on results that can actually inform choices between alternative strategies.

In the case of the swift boat ads there are three key questions a practical political strategist would want answered: how effective actually was the attack, how much of an effect did the delayed response of the Kerry campaign have, and what impact, if any, did the attacks and the controversy have on the image and reputation of Bush and the Republican party.

1. How effective were the ads?

On the first question - how successful were the attacks - it is important to distinguish between public opinion about Kerry's participation in the anti-war movement after he returned from Vietnam and the specific challenges the Swift Boat critics raised about his military service. As noted previously, Kerry's anti-war activities during the early 1970's had always been certain to anger and alienate a significant number of veterans and families of men who had died in Vietnam. According to an August 23-26 poll by the Annenberg Center for Public Policy, some 60% of the veterans who were familiar with John Kerry's anti-war statements disapproved of them.

But in regard to the specific allegations raised by the Swift Boat critics that Kerry did not actually deserve his medals, the Annenberg study found that only 24% of the sample agreed in contrast to 55% who thought Kerry did indeed deserve them. In fact, even a majority of traditionally conservative groups such as men and veterans agreed with Kerry rather then his critics on this issue. It was only among Republican partisans and in the closed conservative media environment created by talk radio and cable TV that the percentage of those who thought Kerry did not deserve his medals ever rose above 30%.

Even after the Republican convention this negative view of the ads has persisted. In a Sept 3-5 Gallup survey, only 21% of the respondents felt the ads were "generally accurate" in contrast to 40% who felt that they "distort the truth" (another 40% were either unfamiliar with the issue or had not formed an opinion)

It is unquestionably disturbing that the Fox/Talk Radio/Republican Party media machine can convince 20% of the American people of charges that are rejected by most other Americans. But, while the attacks clearly twisted media coverage of Kerry in a negative direction during August and prevented Kerry from focusing attention on "his" issues, as an attempt to actually damage John Kerry's image and reputation with voters outside the loyal Republican base, the data from the Annenberg survey (and other polls as well) indicate that the smear campaign was essentially a failure.

2. Would a faster response by the Kerry campaign have prevented the smear from taking hold?

The Annenberg survey also calculated day by day data during mid-August which showed that the number of people who doubted that Kerry deserved his medals gradually rose from about 20% on August 10, when the advertisements had been in the news for about 6 days, up to 30% by August 18th when the Kerry campaign first forcefully responded. After that, the percentage of doubters then sank back down to about 20% by August 26th as an increasing number of eyewitnesses, documents and editorials in major newspapers appeared supporting Kerry's version of events.

Given the 10% decline in the percentage of people who doubted Kerry's earned his medals that occurred once the Kerry campaign and the media began firmly challenging the attacks, it would indeed appear that an earlier response might have prevented the smear from gaining traction in the first place. But, before accepting this conclusion it is necessary to consider that significant sympathy for Kerry was created by the perception that he had been the victim of a two-week long, totally one-sided attack. Had the Kerry campaign responded furiously to the charges the same day they were launched, the counterattack would almost certainly have been spun by the critics as an attempt to "hide the truth" "create a cover-up", "bully Vietnam veterans" or "prevent an honest debate". This accusation, in turn, would then have been used to demand that the mainstream media criticize both camps for "negative politics" rather then just the Republicans and could easily have weakened the very categorical rejection of the accusations that finally appeared in the editorial comments of major newspapers like the L.A. Times and Chicago Tribune.

This does not mean that the slow response of the Kerry campaign was the right strategy. But it does illustrate that, in responding to smears, there are often significant trade-off's between conflicting objectives that have to be taken into consideration.

3. Did the controversy damage the Republican Party's image and reputation?

Regarding this final issue, the Annenberg study found in late August that 46% of the respondents believed the Bush campaign was behind the ads, in contrast to 37% who did not. By the time of the Republican convention an increasing perception of the party's behavior as generally unfair had clearly taken hold. In the September 3-5 Gallup poll noted above, a majority of 52% of the respondents felt that Republicans had "attacked John Kerry unfairly" in contrast to only 42% who did not. (The democrats, in contrast, were seen by most voters as not having attacked Bush unfairly). Similarly, 50% of the Gallup sample felt the Republicans had spent too much time criticizing the Democrats during the convention in contrast to only 39% who thought they had maintained the right balance.

Lessons for next time

There are several conclusions suggested by this data, conclusions that go beyond the currently popular view that democrats should respond to any future smear attacks as rapidly, forcefully and aggressively as possible.

First, it is probably impossible to prevent smears from taking hold within the conservative "echo chamber" of Fox and talk radio and it may be a misuse of resources to attempt to achieve that goal. The more important and achievable goal - preventing the smear from spreading beyond that audience - is probably best pursued by energetically demanding that the mainstream media fulfill their journalistic obligations by emphatically and categorically labeling false accusations as baseless on their editorial pages rather then attempting to debate the issues directly with the smear group itself.

Second, while a very rapid and aggressive response to new accusations can clearly be desirable, it must still be balanced with the need to appear fair, unruffled and unafraid of open and honest debate. A shrill or intemperate counterattack, even if launched at the earliest possible moment, can have little effect or even be counterproductive.

Finally, Republicans have significantly damaged their image and reputation among many moderates and opinion leaders by embracing an essentially dishonest, "win at any cost" approach during this campaign. This tarnished reputation is an asset democrats should energetically exploit. Not only does it reduce the appeal and legitimacy of Republicanism in general, but it makes it easier for Dems to successfully deflect future smear campaigns. Ronald Reagan's famous response, "There you go again", with which he portrayed Jimmy Carter's repeated challenges to his character as tiresome evidence of unfairness, provides one model of how such a strategy can be successfully executed.

Kerry Leads in New PA Poll

In a SurveyUSA poll of Pennsylvania LV's for WCAU-TV Philadelphia, WNEP-TV Wilkes-Barre and KDKA-TV Pittsburgh conducted Sept.7-9, Kerry Leads Bush 49-47 percent, with 5 percent undecided.

Kerry Rips Into Bush Lead in Missouri

A new SurveyUSA poll of Missouri LV's conducted Sept.7-9 for KSDK-TV St. Louis and KOMU-TV Columbia shows Bush now leading Kerry by just 2 percentage points, 48-46 percent, with 5 percent undecided.

New National Poll Shows Neck and Neck Race for White House

A new Zogby America poll of nationwide LV's conducted Sept. 8-9 has Bush at 47 percent and Kerry at 45 percent in a head to head match-up, within a 3.1 percent m.o.e.

September 10, 2004

So Where Are We?

Well, those horse race data keep coming in so let's see what they tell us about the state of the race. Here are the Bush leads in the polls released since the end of the GOP convention. I use Kerry-Bush results and, in the one case where the 2-way race is not available (the Washington Post poll), I estimate a result based on their 3-way race margin and how the 3-way and 2-way races compared in their previous poll.

I use RVs intead of LVs where both are supplied for reasons I have dwelt on extensively. But where only LVs are available I use LVs. In these cases, based on what we know about the relation between LV and RV results this year, the results should actually be better for Bush than if the RV results had been available (though with Zogby and Rasmussen, since they party-weight, it's hard to know how the two factors (LVs vs. party-weighting) net out).

Zogby, September 8-9 LVs: +2
Rasmussen: September 7-9 LVs: +2
Democracy Corps, September 6-9 LVs: +3
Fox News: September 7-8 LVs: +2
CBS News, September 6-8 RVs: +8
Washington Post, September 6-8 RVs: +4
Gallup, September 3-5 RVs: +1

Conclusion: Bush is ahead but probably not by much. September 6-8 was apparently a good period for him, relative to the immediate post-convention period covered by the Gallup poll, but he appears to be fading a bit based on the pattern of surveys covering slightly later time periods.

As for the idea that Bush has surged into a commanding lead: not in this universe anyway.

More on Labor Day Leads

I did a recent post arguing that Gallup's piece on labor day leads was misleading since they compared Bush's LV lead this year to (mostly) RV leads from previous races. They should have, I argued, compared apples-to-apples, RVs to RVs.

Frank Newport of Gallup was sporting enough to print an edited version of my comments on their editors' blog, along with his reply. In Newport's reply, his key rationale for conducting their analysis the way he did was he wished to "[use] Gallup’s best available estimates at Labor Day for each year for which we have data." (But by all means read his argument in full through the link.)

To further this discussion, here are some additional remarks on the issue replying to Newport's argument. I should add that I don't believe that Gallup has any particular axe to grind in how they did this analysis--I just think in this case they got it wrong.

Thanks for your thoughtful reply. But I still don't buy it. You surely must see that it makes a difference when people read these analyses with "7 point deficit to overcome" in mind rather than "1 point deficit to overcome"? And in fact that's how your analysis was written, focusing reader attention on the 7 point LV deficit.

And the fact remains that apples-to-apples comparisons are far preferable to apples-to-oranges comparisons. Therefore the proper comparison is between this year's RV labor day results and previous years'. Otherwise, you are not analyzing the same change (RV labor day gap vs. final gap) across years. (By the way, thanks for drawing my attention to the national adults samples prior to 1952. In the same apples-to-apples spirit, I would drop these cases from the analysis.)

Using a consistent time series would make a difference to your analysis.

Instead of:

"In summary, the history of presidential elections since 1936 suggests that in about half of the cases, the type of gap change that would be necessary for Kerry to tie or move ahead of Bush has occurred. About half the time it has not. If a gap change does occur, the odds are higher than 50-50 that it would be in Kerry's direction (i.e., a shrinkage rather than an expansion of Bush's current lead). "

You would have:

"In summary, the history of presidential elections since 1952 suggests that in all cases, the type of gap change that would be necessary for Kerry to tie or move ahead of Bush has occurred. If a gap change does occur, the odds are very strong (11 out of 13) that it would be in Kerry's direction (i.e., a shrinkage rather than an expansion of Bush's current lead)."

This clearly sounds quite a bit different. And thinking Kerry is behind by one point, rather than 7 points, clearly makes a big difference when considering elections like 1960 and 1980, which loom large in your analysis. Kennedy was behind by a point in 1960 among RVs--the same as Kerry--and Reagan was behind by 4 points in 1980--more than Kerry. If you're thinking 7 points behind, those races look a lot different.

In short, lacking a consistent time series of any length on LVs, you just shouldn't use 'em in an analysis like this.

NC Trending Purple?

In a poll of NC LV's taken by Survey USA for WBTV-TV Charlotte and WTVD-TV
Raleigh Durham Sept. 6-8, Bush leads Kerry 50-46 percent( moe 4.2 percent ).

Kerry, Bush in Statistical Tie in New Ohio Poll

A SurveyUSA poll of LV's in Ohio for WCPO-TV Cincinnatti and WKYC-TV
Cleveland conducted Sept. 6-8 has Bush with 50 percent vs. 47 percent for
Kerry ( moe 3.8 percent ).

Bush Has Narrow Lead in Democracy Corps Poll

A poll of nation-wide LV's conducted Sept. 6-9 by Democracy Corps has Bush at 50 percent, Kerry at 47 percent with 1 percent for "other."

New Economist Poll Shows Race Nearly Tied

A poll of nation-wide RV's conducted Sept 6-8 by YouGov for the Economist has Bush leading Kerry 46-45 percent, with 1 percent for Nader.

More on Those (Un)likely Gallup Voters

Another nugget from Alan Abramowitz:

If you assume that Democrats, Republicans, and independents in their LV sample voted about the same way as Dems, Reps, and indies in their RV sample (see their "dissecting the vote" analysis on the Gallup website), which was 90-7 Kerry for Dems, 49-46 Kerry for indies, and 90-7 Bush for Reps, in order to have the overall result come out 52-45 Bush there would have to be about a TEN point Republican advantage in party id among LVs. (If you assume 30 percent Dems, 40 percent Reps, and 30 percent indies, for example, with those Bush and Kerry percentages, you end up with almost exactly a 52-45 Bush lead.) Now that is ridiculous. Does anyone really believe that Republicans are going to have a 10 point advantage among 2004 voters?

I sure don't and you shouldn't either. In 2000, the Democrats had a 4 point advantage over the Republicans. That advantage, in my view, is likely to remain stable in 2004, though it's certainly possible that it might diminish some (or increase!). But turn into a 10 point GOP advantage? No way. The fact must be faced: Gallup likely voters look pretty darn unlikely and give a distorted picture of political reality.

Perhaps it's time for a re-think on this one over at Gallup headquarters.

September 9, 2004

Maybe September 6 Was a Really Good Day for Bush

OK, horse race fans, when last we picked up the story (September 3-5 Gallup poll), Bush had a one point lead on Kerry among RVs.

Today we have three different polls released covering about the same period with three pretty different horse race results.

1. CBS News, polling September 6-8 among RVs, gives Bush an 8 point lead.

2. ABC News/Washington Post, polling September 6-8 among RVs, gives Bush an estimated 4 point lead. Why do I say "estimated"? Because, now that Nader is becoming an ever-less-viable candidate, the WP poll has decided only to ask the three-way horse race question and not the followup for Nader voters that allows you to construct a 2-way race. Makes a lot of sense, right?

Since, in their last poll, the 2 way race knocked 2 points off Bush's lead, I do the same thing in this poll and estimate that Bush's 6 point lead in the 3 way race translates into a 4 point lead in the 2-way race.

3. Finally, Fox News, polling September 7-8 among LVs (no RV data available), gives Bush a 2 point lead.

Kinda confusing, huh? Why would September 6-8 be a better period for Bush than September 3-5, right after the convention? And could Fox's polling period, which does not include September 6, mean they missed Bush's best day and he was starting to go downhill a bit?

Stay tuned! There'll be more polls coming at us shortly, I'm sure.

Watch Out for Those Gallup LVs in Ohio

As Alan Abramowitz has pointed out to me in another missive:

Gallup's new poll in Ohio appears to have the same problem as their latest national poll--there is again a huge discrepancy between their results for registered voters (Bush 48, Kerry 47) and their results for likely voters (Bush 52, Kerry 44). Based on the numbers of registered and likely voters in the Gallup sample, this means that they are projecting that 90 percent of Bush supporters will vote but only 78 percent of Kerry supporters will vote. Again, this seems way out of line with evidence from previous elections and with other polls. In contrast, in Pennsylvania and Washington, Gallup's results for registered and likely voters are much more similar, and in line with other recent polls.

New Gallup State Polls

The CNN/USA Today/Gallup Poll of Washington state conducted September 3-7, 2004 has Kerry at 51 percent, Bush at 43 percent, Nader at 2 percent, and neither/unsure at 4 percent.

The CNN/USA Today/Gallup Poll of Pennsylvania RV's conducted Sept. 4-7, 2004 shows a tie, with 47 percent each for John Kerry and George Bush, with 6 percent neither/unsure.

The CNN/USA Today/Gallup Poll of Ohio RV's conducted Sept. 4-7, 2004 has Bush at 48 percent, Kerry at 47 percent and neither/unsure at 5 percent.

A CNN/USA Today/Gallup Poll of Missouri RV's conducted Sept. 3-6, 2004 reports Bush leads with 53 percent, Kerry 42 percent with 5 percent neither/unsure.

Guess They Skipped the Apples-to-Apples Thing in Gallup Training School

This is really unbelievable. Gallup posted an analysis on their site yesterday about estimating election probabilities based on labor day poll data that is almost completely worthless. The reason is that they focus on Kerry's 7 point deficit among LVs on labor day (can he overcome it?), while basing their analysis almost entirely on data about RVs on labor day.

How do I know their labor day poll data is almost entirely (prior to 1996) based on RVs? Because they published these data, clearly marked as being from RVs prior to that date, in an analysis on their own site just 6 days ago (September 3)!

Don't they read their own stuff? Clearly it makes no sense to analyze a lead among LVs this labor day, and its possible relation to the final outcome this year, on the basis of historical data about RV leads on labor day and how much they changed by election day.

Thus, the question Gallup should have been asking is: can Kerry overcome his 1 point deficit among RVs by election day, based on historical patterns? Turns out the answer to this question--really, the only question that their data can properly answer--looks pretty darn good for Kerry.

In 17 of 17 cases, going back to 1936, the labor day margin between the candidates changed enough for Kerry to tie or surpass Bush in the popular vote and, in 12 of 17 of those cases, the change was in Kerry's direction (i.e., that is, in the direction of the candidate who was behind among RVs on labor day).

Moreover, if you compare Bush's position to the position of incumbent presidents who won their campaigns for re-election, it doesn't look auspicious. In 9 cases, going back to 1936, winning incumbent presidents on labor day had an average lead of 12 points and a median lead of 11 points among RVs. The only winning incumbent president who was in a worse position than Bush is this year was Harry Truman in 1948.

Maybe I'm biased, but I have a really hard time seeing George W. Bush as Harry Truman.

September 8, 2004

Does Bush Really Have a 7 Point Lead?

I've certainly made no secret of my skepticism. Now consider this excellent analysis along the same lines by Professor Alan Abramowitz of Emory University, one of the leading academic analysts of American politiics. (He sent this to me in an email and graciously agreed to allow me to share it with readers of this blog.)

1. The latest Gallup Poll has Bush ahead of Kerry by 52-45 percent among likely voters but by only 49-48 percent among registered voters. Based on the numbers of registered and likely voters in their sample, this means that Gallup is projecting that 89 percent of Bush supporters will vote but only 79 percent of Kerry supporters will vote. That seems unrealistic. It is way out of line with data from the American National Election Studies on turnout among registered Dems and Republicans in recent elections. For the past three presidential elections, the turnout gap between Republicans and Democrats has averaged 3 percentage points and was never larger than 4 percentage points. The smallest gap was in 1992 (1 point), the election with the highest overall turnout. Assuming that 2004 will be another relatively high turnout election, we should probably expect a relatively small turnout gap, similar to 1992.

2. Among registered voters, Gallup shows Bush leading by one point overall, with Kerry leading 90-7 among Democrats, Bush leading 90-7 among Republicans, and Kerry leading 49-46 among independents. This means that Gallup's sample of registered voters includes more Republican identifiers than Democratic identifiers. But in 2000, according to the VNS national exit poll (which hits the overall percentages for Bush and Gore right on the nose), Democrats made up 40.3 percent of the electorate while Republicans made up only 36.5 percent of the electorate. If you apply Gallup's trial heat results among Democrats, independents, and Republicans to the VNS 2000 electorate, Kerry comes out with with a four point lead: 50.3 percent to Bush's 46.4 percent.

Food for thought, eh?

New ICR Poll Has Kerry Over Bush by a Point

A new poll, conducted September 1-5 by International Communications Research, has a 48-47 lead for Kerry among RVs, consistent with the recently-released Gallup poll and further calling into question the results of the Time and Newsweek polls.

September 7, 2004

Kerry Widens Lead in Battleground States!

Now that's a headline you're not likely to see in the mainstream media, consumed as they are with the storyline du jour about Bush's Big Mo' from the convention.

But that's what the internals of the latest Gallup poll tell us. Prior to the Republican convention, Kerry had a one point lead among RVs (47-46) in the battleground states. After the Republican convention, now that battleground voters have had a chance to take a closer look at what Bush and his party really stand for, Kerry leads by 5 in these same states (50-45)! Note that Kerry gained three points among battleground voters, while Bush actually got a negative one point bounce.

And wait--there's more! The Gallup poll's internals also show that Kerry continues to lead among independents (49-46) and that both parties' partisans are equally polarized for their respective candidates (90-7). Note that these findings directly contradict the results of the recent Newsweek poll, which showed Bush doing much better among Republican partisans than Kerry was doing among Democratic partisans. Note also that, given the equal polarization of partisans and Kerry's lead among independents, the only possible reason Bush has any lead at all among Gallup's RVs must be because their sample has a GOP advantage on party ID (my guess is 5 points) that is inconsistent with almost all other polling data from this campaign season (see my recent post on the Newsweek poll for more discussion of this issue).

Indeed, if equal polarization of partisans continues and Kerry carries a 3 point lead on independents into the election, he'll win fairly easily, since the Democratic proportion of voters in presidential elections is always higher, not lower, than the Republican proportion. In 2000, after all, Bush carried independents by 2 points and received stronger support from his partisans than Gore did from his--but still lost the popular vote by half a point.

Now that's another storyline you're unlikely to see in the mainstream media.

September 6, 2004

Gallup Poll Gives Bush Only a 2 Point Bounce

I think those of us who have expressed skepticism about the results of the Time and Newsweek polls can consider ourselves vindicated. The new Gallup poll, conducted entirely after the GOP convention and therefore the first poll that truly measures Bush's bounce, shows Bush with a very modest bounce indeed: 2 points, whether you look at RVs or LVs. His support among RVs has risen from 47 percent before to 49 percent after the convention, so that he now leads Kerry by a single point (49-48) rather than trailing by a point.

But that's it. Contrast Bush's 49-48 lead among RVs in this poll to Time's 50-42 Bush lead and, especially, Newsweek's 54-43 Bush lead in the same matchup. Quite a difference.

Note also that Bush's 2 point bounce from his convention (which, remember, is defined as the change in a candidate's level of support, not in margin) is the worst ever received by an incumbent president, regardless of party, and the worst ever received by a Republican candidate, whether incumbent or not (see this Gallup analysis for all the relevant historical data). In 2000, Bush received an 8 point bounce. And even his hapless father received a 5 point bounce in 1992.

So that's the big story, right--Bush got a disappointingly small bounce and the earlier Time/Newsweek polls got it wrong about the bounce and how well Bush is doing. Nope, not if you're writing stories at USA Today. You dasn't contravene the current CW about the campaign (Bush surges ahead!) no matter what your own data says.

That's why we get a story like this one, "Bush leads Kerry by 7 points", which prominently features the LV results (where Bush does have a 7 point lead) and resolutely refuses to dwell on Bush's historically poor result from his convention or on his almost non-existent lead among RVs.

Instead, the article goes on to discuss some results from the poll that look pretty good for Bush and, of course, allow Matthew Dowd to spin the poll's results in the GOP's direction.

As usual, of course, Dowd does a pretty good job of spnning the poll (we gained more than we expected!), which is then followed by a very weak reply from Mark Mellman where he essentially says the GOP's gains from the convention will fade. That's not the right reply. The right reply is what gains and and how very disappointed the GOP must be in their historically poor performance.

But this is a persistent problem: Dowd and the people behind him relentlessly spin every poll and feed journalists various mini-analyses (can we call them "analysisoids"?) that purport to show how great Bush is doing relative to expectations, historical patterns, etc. and how bogus any poll is that shows Kerry doing well. Where are the Democrats on this one? The occasional lame quote from Mellman is not enough to outgun Dowd in this particular part of the political debate.

I don't know whether Mellman just can't matchup with Dowd in this department or if he simply doesn't have the time to come up with good stuff or whether he needs a team of people monitoring the polls and coming up with analysisoids that he (or someone) can then retail to the media. Whatever the problem, it's time the Democrats found a solution so that Dowd no longer has this particular field all to himself.

End of rant. Let me mention a few other results from the Gallup poll that suggest the relative ineffectiveness of the GOP convention.

Bush's acceptance speech, which the media fawned over so ostentatiously, was not rated any better by the public than was Kerry's--in fact, it received slightly worse ratings. Kerry's acceptance speech was rated excellent by 25 percent and good by 27 percent; Bush's was rated excellent by 22 percent and good by 27 percent.

In terms of whether the Republican convention made voters more or less likely to vote for Bush--the real point of the convention after all--there were almost as many saying the convention made them less likely to vote for Bush (38 percent) as said it made them more likely (41 percent).

This is actually quite a poor performance. The Democratic convention this year had a substantially better 44 percent more likely/30 percent less likely split. In fact, looking back to 1984, which is as far back as Gallup supplies data, no candidate has ever had a more likely to vote for/less likely to vote for split even close to as bad as Bush's this year.

Well, what about the tone of the convention? Do voters think the Republicans got that one right? Nope. Just 39 percent think the GOP maintained the right balance between criticizing the Democrats and saying positive things about themselves, compared to 50 percent who think they spent too much time criticizing the Democrats. By contrast, in 2000, 45 percent thought the GOP maintained the right balance in their convention, compared to 38 percent who thought they spent too much time criticizing.

But this unfavorable judgement on tone for the GOP this year is not without precedent. In 1992, just 26 percent thought the Republicans maintained the right balance in their convention, compared to 56 percent who thought they spent too much time criticizing.

Sounds like W is going down the same road trod by his father. Let's hope it produces the same result on election day.

Bush, Kerry Running Close In New Mexico

An Albuquerque Journal poll conducted by Research & Polling Aug. 27-Sept. 1, 2004 had Bush at 45 percent, Kerry at 42 percent, 10 percent undecided/wouldn't say, Nader 1 percent, Badnarik 1 percent, Cobb 1 percent.

September 5, 2004

Could Everyone Please Just Calm Down Out There?

It's a full-time job trying to keep Democrats from hyperventilating whenever a new poll comes out! Friday, I tried to calm folks down about the new Time poll; today I'll try to do the same about the new Newsweek poll, conducted 9/2-3, which has Bush ahead 54-43 among RVs.

Here are some important points to keep in mind about the poll:

1. It is still not a true bounce poll; only one night of the two covered by the poll actually took place after the GOP convention was over. That night is highly likely to be Bush's best post-convention night, since it was right after his big speech and the huge media splash the next day. And, in fact, Newsweek's data show that Bush led by 16 points in their poll on this night and by only 6 the night before. Don't forget that Kerry did very well in polls the night right after his speech then fell off rapidly in the next few days.

So why do Newsweek and Time insist on doing their bounce polls wrong so they're almost guaranteed to get misleading results? Simple: their publication schedule. They've got to have to data in time to dump it into their print publication. If they waited to do it right the poll would be too old to put in their magazine the subsequent week.

This is especially egregious since even a poll conducted entirely after the convention needs to be viewed with caution. As Charlie Cook points out:

A week or 10 days after the GOP convention, the electorate should have stopped bouncing and settled back down enough for horse race poll results to once again have some real meaning.

2. Aside from the timing, there are other reasons to be skeptical of the Newsweek poll. As has been widely reported in various blogs, the partisan distribution of the RVs in the Newsweek poll is quite startling: 38 percent Republican, 31 percent Democratic and 31 percent independent. This 7 point lead for the GOP on party ID does not comport well with other data on partisan distribution this campaign season--which have consistently shown the Democrats leading by at least several points--and can't be blamed on a likely voter screen since there was none.

As Chris Bowers of MyDD shows, if you assume a more reasonable distribution of party ID, Bush's lead is about cut in half. Moreover, if you assume that the differential in partisan support rates in the poll--94-4 for Bush and only 82-14 for Kerry--is, if not overstated now, highly likely to converge toward parity in the near future (as it has been for most of the campaign), even a Bush lead of 5-6 points looks very unstable.

So how did Newsweek manage to pull a sample with a 7 point GOP lead on party ID? It is certainly possible that there has been a sudden, large shift in party ID to the Republicans; the distribution of party ID is not completely stable and does indeed change over time. But a shift of this magnitude so suddenly and so off-trend (which has been toward the Democrats) strikes me as quite unlikely. I find it more plausible that there was differential interest in being interviewed by Democratic and Republican voters over the time period and that produced a skewed distribution of partisan identifiers in their RV sample.

Does that mean I favor polls like this weighting their samples by party ID? No, I don't, because the distribution of party ID does shift some over time and polls should be able to capture this. What I do favor is release and prominent display of sample compostions by party ID, as well as basic demographics, whenever a poll comes out. Consumers of poll data should not have to ferret out this information from obscure places--it should be given out-front by the polling organizations or sponsors themselves. Then people can use this information to make judgements about whether and to what extent they find the results of the poll plausible.

3. It's still a long time 'til election day. People should resist the urge to push the panic button and insist that Kerry launch an incendiary campaign against Bush's and his surrogate's personal attacks. As John Judis points out, there are interesting similarities between this campaign and the Reagan-Carter campaign of 1980. These similarities suggest that:

....just as Bush might be wise to avoid Carter's mistakes, Kerry might be wise to consider Reagan's successes in 1980. He is certainly going to have to answer some of the Bush campaign's personal attacks, just as Reagan occasionally responded to Carter--although Reagan did so in a disarming manner ("there you go again") that put the onus of disagreeability directly onto his opponent. But Kerry needs to direct the public's attention, like Reagan did, to the underlying reality of the economy, the Iraq war, and the threat of Al Qaeda; and he needs to propose ways to deal with each that are at least plausible, if not preferable to those adopted by Bush. If he does that, and if he shows himself to be the equal of Bush in the debates, he could discover, like Reagan did in 1980, that the voters are ready to put someone new in the White House.

Amen. End of sermon.

September 3, 2004

How High the Bounce?

It must be stressed that at this point: we don't know. Measurements of a candidate's bounce should be based on polls taken before and after a convention. So far we have no such data--in fact, tonight will be the first night where polling can be conducted that is truly after the completeion of the GOP convention. Therefore, we won't have real bounce data for several days.

That said, let us consider the results of polls taken during the GOP convention. The one that seems to be freaking out some Democrats is the just-released Time poll. (I continue to be amazed at how easily many Democrats are panicked by the release of an unfavorable poll; there's been a lot of talk about whether John Kerry is tough enough--I'm more worried about whether regular old Democrats are tough enough. Sheesh.)

The Time poll, conducted 8/31-9/2, has Bush ahead by 11, 52-41 in a 3-way LV matchup that includes Nader. (Time presumably will eventually release the 2-way LV matchup. I'm doubtful we'll ever see RV results.) How plausible is this result?

Well, it's certainly possible that Bush was as far ahead during the convention as this poll suggests. But all other available polls taken during the convention contradict this result.

In an attempt to compare apples to apples, here are Bush-Kerry results from contemporaneous 3-way LV matchups (except Rasmussen, where only a 2-way LV result is available), with Bush's margin in parentheses:

Zogby, 8/30-9/2: 46 Bush-43 Kerry (+3)
ARG, 8/30-9/1: 47 Bush-47 Kerry (tie)
Rasmussen, 8/31-9/2: 49 Bush-45 Kerry (+4)

In this company, 52 Bush-41 Kerry (+11) certainly sticks out. Could it have anything to do with the different dates included in these surveys, even though they are very close? Well, the Rasmussen data are from exactly same period as the Time data (8/31-9/2).

But if you are skeptical of the Rasmussen data, consider the Zogby data. The Zogby data only include an additional day (8/30) when compared to the Time data. But perhaps 8/30 was a very pro-Kerry day since the Republican convention had just started. However, for Zogby and Time to matchup (have Bush leading by 11) for the three days they share, Kerry would have to be leading by about 21 points in Zogby on the day (8/30) they do not share. I rather doubt that is the case.

The simplest hypothesis then is that the Time poll, for this period, is exceptionally pro-Bush and therefore should be viewed with skepticism.

In the meantime, we will await the release of data that actually measure the convention bounce, defined, just to be clear, as the change in a candidate's level of support (not the margin) from the period before to the period after the convention. And while we're waiting, here are some interesting observations that are worth keeping in mind from a just-released Gallup analysis of the bounce issue:

Based solely on history, the Bush-Cheney ticket could expect to gain five to six points among registered voters after this week's convention. That would result in a 52% to 53% support level for Bush among registered voters, up from 47% in the pre-convention poll.

However, the results from Gallup's post-Democratic convention poll showed that history might not apply in 2004, a year in which the electorate was activated long before the conventions (usually the conventions serve to activate voters), and a year in which relatively small proportions of undecided and swing voters are available to the two presidential tickets. Also, the post-Democratic convention poll suggested that the Democratic convention might have helped energize Republican voters. It is unclear whether the Republican convention could have a similar paradoxical effect on Democrats, or if Republicans will be activated, as is typically the case.

So, stay tuned. And don't forget that even when we see the real bounce data, the pattern after the Democratic convention was for Kerry's increase in support to dissipate quickly. We shall see if the same thing happens to Bush, whatever his bounce level.

Update: The Polling Report now has the Time poll's results for the 2-way race and for RVs. For what it's worth, the Bush-Kerry RV result is 50-42 Bush, for an 8 point lead. But the anomalous nature of the Time poll's results remain.

Zogby Poll of Likely Voters Shows Bush 46%, Kerry 44%

A Zogby poll conducted August 30-September 2 showed George W. Bush leading John Kerry among likely voters by 2%. The poll also found, however, that, by 48% to 46%, the respondents wanted "someone new" rather then agreeing that Bush "deserves to be reelected"

September 2, 2004

Adventures in Likely Voter Land

It has not escaped my notice that many people are puzzled as to how exactly polls go about determining "likely voters" (LVs). There's a good reason for this: polling firms or sponsors rarely put much effort into explaining, clearly and precisely, the mechanics of how they select these LVs.

So, as a public service, here's how they do it. Let's start with Gallup. According to David Moore of Gallup:

Gallup asks each [RV] respondent seven LV screening questions, and gives each person an LV score of 0 to 7. [Assuming a turnout of 55 percent], the top 55% are classified as likely voters.

In practice that typically means all of the "7"s--given full weight--plus some proportion of those with lower scores (usually the "6"s), who are weighted down so that the size of the likely voter sample matches the projected turnout for the year (apparently 55 percent this year). All other voters are discarded from the sample.

What are the Gallup likely voter questions? Unfortunately, the exact questions and their wording are not released by Gallup along with their polling data, but the questions apparently involve past voting behavior, interest in the election, intention to vote in the election and knowledge of things like the location of the local polling place.

That's how Gallup does it. What about other organizations--do they select likely voters in the same way? Nope, they don't. CBS News doesn't use a cut-off model, where low-scoring respondents are thrown out altogether, but instead includes everyone in their RV sample, in some form, in their LV sample. They do this by asking respondents a series of voting-related questions and then assigning each respondent a weight based on their score on these questions, from very high weights for high-scoring respondents to very low weights for low-scoring respondents.

Finally, by far the most common way is simply to ask a few screening questions and then terminate the interview with those respondents who give the "wrong" answers. Or only one question; some likely voter screens are as simple as asking an RV how likely they are to vote in the upcoming election; if they don't say "almost certain" or "probably", out they go.

So that's how they get the likely voters in the polls you read about. How do they know that likely voters, months before the election, are actually the voters who will show up on election day? They don't.

Here's David Moore from Gallup again: "We simply do not know, nor can we know, which model is better during the campaign itself. " Exactly. So why does he think the Gallup LV model works so well months and months before the election. Because "if it is the most accurate model just before the election, it is probably the most accurate during the campaign as well".

But that doesn't follow at all. The Gallup LV model could work perfectly right before the election (not that it really does, but that's another discussion) and still be quite a biased instrument earlier in the campaign. Pretty much by definition, Gallup's LVs months before the election are not the same voters as Gallup's LVs right before the election, since voters answer the LV questions differently at different stages of the campaign. And if there is any kind of partisan dimension to "tune-in", so that, say, Democratic partisans or groups that lean strongly Democratic (like minorities) tend to tune in later, that means the LV model will have a systematic tendency to, on average, favor the party (the Republicans) whose partisans or groups tune in the earliest.

Of course, my hypothesis here about Gallup LV bias might be completely wrong. But to evaluate it, Gallup would have to make available the demographics and partisan breakdown of the both its RV and LV samples for the polls it releases plus, ideally, the results (including demographics and partisan breakdowns) of the various screening questions it uses. I'm not holding my breath.

Welcome to the New Donkey!

I just wanted to urge everyone to point their browsers to the just-started "unofficial" DLC blog, NewDonkey, written by my good friend, Ed Kilgore, one of the most astute political analysts around. His blog is invariably well-written and full of valuable insight (his coverage of the Republican convention has been particularly good). So visit often. You'll be glad you did.

Kerry 48%, Bush 46% in New ARG poll

John Kerry receives 48% and George W. Bush 46% in a new American Research Group survey of registered voters conducted Aug. 30-Sept 1. The poll also shows Kerry with a 5% lead among independents.

September 1, 2004

So What's Going on in Pennsylvania?

There's been a shocking amount of weeping and wailing and gnashing of teeth in Democratic circles considering that most polls still show Kerry ahead. Indeed, the just-released ICR poll (see below) has Kerry ahead by 7 points among RVs, a greater margin than the same poll had three weeks ago.

But it's hard to cheer people up when they're determined to be gloomy. Indeed, there's a tendency to seize on "evidence" that supports a gloomy viewpoint without looking at it very closely.

Take the case of Pennsylvania and recent polling data. Democrats seem to be convinced the race there has seriously tightened (see this post by Josh Marshall which is otherwise terrific and level-headed but buys into this particular meme).

What's the evidence for this tightened Pennsylvania race? Disregarding the recent poll by the very Republican Strategic Vision firm, the most recent reputable poll in PA is the Gallup poll of 8/23-26. That poll shows Kerry trailing Bush by a point among likely voters (LVs). I emphasize that this result is among LVs because most PA polls that people have heard about have been conducted among registered voters (RVs). Therefore, comparing these earlier PA polls to Gallup's current LV result is not an apples to apples comparison and tells you nothing about whether and how the race has changed.

Especially when we note the following: the same Gallup poll that has Bush ahead by a point among LVs in the Bush-Kerry matchup has him behind by 5 points among RVs!

In fact, check out the last three reputable poll results from PA among RVs:

8/23-26 (Gallup): Kerry, +5
8/13-21 (Issues PA): Kerry, +2
8/11-16 (Quinnipiac): Kerry, +5

Conclusion: there is no tightening in the PA race once we do an apples to apples comparison.

This is only the latest example of confusion being sown by Gallup's LV model, which has been producing consistently pro-Bush results lately. Indeed, Gallup did a WI poll at the same time as their PA poll and found Bush ahead by 3 points among LVs in WI, but Kerry ahead by the same margin among RVs.

I continue to believe these LV results should be de-emphasized until later in the campaign. It is still too early to put much faith in likely voter screens/models as representing very accurately the voters who will actually show up on election day. There is reasonable evidence that careful likely voter methodologies work well when it is quite close to an election and do fairly accurately capture that pool of voters. But there is no such evidence for LV samples drawn this far out.

Indeed, my understanding is that Gallup, in particular, does LVs throughout the campaign not so much because they believe they are capturing election day voters all along, but more so that they can avoid having to explain sudden shifts in the horse race question as LV data replaces RV data. There have apparently been some problems with this in the past, so reporting both from the very beginning of the campaign eliminates any potential embarrassments along these lines. But that doesn't mean the LV data is any better at this particular point in time--it merely means they're providing it.

In fact, since the sample size for LVs is smaller and since the composition of the LV sample will shift depending on how political developments are affecting interest and intensity levels among different groups of voters, additional volatility and potential bias is built into the LV samples that is not there with the RV samples.

The issue of bias can particularly be a problem if it's true that, say, Democratic-leaning voters tend to tune into campaigns later and are therefore more likely to be screened out of the likely voter sample until close to the election (especially with a tight screening procedure, which Gallup definitely has). If that's true, then an LV sample could perform reasonably well close to election day, as a greater proportion of Democratic-leaning voters finally get screened in, but be quite biased toward the Republicans until then.

That could be what's happening today.

So take a tip from me: always check those RV results. It'll help keep a smile on your lips and song in your heart!

ICR poll has Kerry up by 7 points

An August 25-29 poll by ICR/EXCEL shows John Kerry with a substantial 51-44 lead over George W. Bush among registered voters and a lead roughly double that among independents.

Note: ICR is International Communications Research. The poll was conducted over the telephone as part of ICR’s twice weekly consumer omnibus study, EXCEL. ICR also polled the presidential race in 2000 and did about as well as Gallup in predicting the final margin (both organizations were off by 2 points).