A Proud Member of the Reality-Based Community

Acknowledgements

SEARCH eRiposte!

ELECTION 2004: CAVEAT EMPTOR
Flawed Poll Results Tilt Significantly and Artificially in Favor of Bush
Kerry is not Out of the Woods Yet But He Appears to be About Even* with Bush

9/20/04

SUMMARY

In recent days, widely varying poll results have been reported by different polling organizations. Some polls show large or enormous leads for George W. Bush (over John Kerry), while others show a tie or slight Kerry lead. However, analysis by seasoned poll watchers show serious methodological flaws in the former polls - especially the use of meaningless "Likely Voter" models and unproven/incorrect Party ID weighting in Registered Voter models. Adjusting for such flaws provides a picture of a race that is essentially close and unpredictable - and essentially dead even* at this time. Other factors such as the Margin of Error (MoE) in the polls, the Nader effect, the impact of Absentee Ballots, and Undecided Voters, is also not properly accounted for in some articles that provide "analysis" on polling results. There are some broad lessons here for Kerry supporters - and some feedback for the Democratic Party and Senator Kerry. 

[P.S. A less noted but significant aspect in the Gallup poll which skewed dramatically towards Bush is that Kerry is "leading" Bush amongst independents by 7% (MoE 4%)] 

ACKNOWLEDGEMENTS

The following sites are acknowledged in particular since their analysis forms the main basis for my summary:
Donkey Rising (Ruy Teixeira), MyDD, DailyKos, Left Coaster

OUTLINE

1. Brief summary of some recent polls

2. POLLING ASSUMPTIONS: What should we believe?

2.1 "Likely" Voter (LV) models

2.2 Registered Voter (RV) models

2.3 Margin of Error (MoE)

2.4 The Nader Effect

2.5 Overseas Absentee Ballots?

2.6 How do Undecideds Split Their Votes?

2.7 ASIDE: Telephone polling -- What about Cell Phone Owners?

2.8 WHAT MATTERS: The Electoral College!

3. Recent National Poll results adjusted for LV and RV model flaws

3.1 Gallup

3.2 CBS/New York Times

3.3 Other Polls: ABC, Fox, IDB, Newsweek

4. Conclusions and Lessons for Kerry supporters

4.1 Some Election 2000 History

5. Still a close, likely dead-even race: Feedback to Senator Kerry and the Democratic Party

6. APPENDIX: Why do "leading" media outlets publish stories using highly questionable results?

DETAILS

1. BRIEF SUMMARY OF RECENT POLLS

The table below is extracted from the useful site 2.004k.com, showing the NATIONAL poll results reported in the past week. The third column in the table is critical, and 2.004k.com provides important guidance on this column (at the bottom of their page) - something that is explored in greater detail in the next section.

RV = Registered Voters (most desired group)
A = All Adults (next-most desired group)
LV = Likely Voters (arbitrary and subjective; least desired group)

As the table shows, many recent polls show Kerry either tied with Bush or even slightly "ahead". Others show Bush significantly "ahead". This is what we need to explain and understand today.

CAUTION! I'm using terms like "ahead" and "behind" loosely and this is NOT statistically accurate terminology because it ignores the margin of error (MoE)! If you included MoE, the vast majority of these results suggest a close and unpredictable race. If you want to understand confidence intervals (also important) and MoE, here are some useful links: StudyWorks, CampaignDesk, Secular Blasphemy

2. POLLING ASSUMPTIONS: WHAT SHOULD WE BELIEVE?

2.1 "Likely" Voter (LV) models

The main point I highlight in this section, as Ruy Teixeira has elegantly argued for a while now, is that Likely Voter models create significant errors in projecting the horse-race results. As a clincher, he notes: 

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!

Why are LV models very likely (!) to be flawed?

Teixeira in The Gadflyer says:

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.

In this post he summarizes the case against Gallup's LV data (bold text is eRiposte emphasis):

Here is a summary of the case against Gallup's LV data:

...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.

...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!

Thus, as Teixeira says (bold text is eRiposte emphasis):

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.

Finally, this post from him (September 8th) illustrates the folly of using LV models this far ahead, more quantitatively (bold text is eRiposte emphasis):

Now consider this excellent analysis along the same lines by Professor Alan Abramowitz of Emory University, one of the leading academic analysts of American politics. (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.

The lesson? In general, don't pay much attention to LV data. 

2.2 Registered Voter (RV) models

Having dispensed with the LV model issue, we still have to address the apparent discrepancy seen in the RV results shown in Section 1. It turns out, RV results from many organizations in this election cycle also have flaws in the assumptions they have made on party identification of voters. Using the documented party ID from the 2000 election, or even in the 2004 study by Pew would reverse the results in some cases (i.e., Kerry would lead Bush rather than the other way around) and make them close-to-even in others.

Here's Teixeira again (bold text is my emphasis):

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 nine- to ten-point Republican advantage).

How realistic is it to be suddenly turning up a Republican lead on party identification, much less a large one? Not very. The weight of the academic evidence is that, while the distribution of party identification 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 identification among voters has favored and continues to favor the Democrats. In 2000, the exit polls showed Democrats with a four-point advantage over Republicans. In 1996, it was also five points; in 1996 [eRiposte note: presumably Teixeira meant 1992?], it was three points and in 1988 it was also three points.

The data also indicate that there were two shifts in party identification over the 2001–2004 period which largely canceled each other out. The first shift, in the period after September 11, shaved several points off the Democrats' lead and brought the Republicans close to even (but never ahead) in party identification. The second shift took place in late 2003 and 2004 and reconstituted the Democrats' lead on party identification to about four 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).

But if the party identification distribution is fairly stable and tends to change rather slowly, why would polls suddenly be turning up unrealistically high numbers of Republican identifiers? The best explanation, 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.

But whatever the explanation for the disproportionate number of Republicans in recent samples, if those numbers are unrealistic, they are skewing reported horse race results toward Bush. What, if anything, should be done about this?

One possible solution is to weight poll results by a more reasonable distribution of party identification. 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.

...

Given that party identification 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 recently put it in Public Opinion Watch:

[B]ecause the distribution of party identification 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 identification, 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 identification may be necessary to avoid giving the public distorted impressions of the state of the race.

What is a reasonable distribution of party identification to use in such weighting? One obvious candidate is the exit poll distribution from 2000: 39 percent Democrats, 35 percent Republicans, 26 percent independents. Moreover, the Democratic advantage in this distribution—four 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 identification numbers taken over the previous several months. This would allow the distribution of party identification 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.

Separately, pollster John Zogby has also indirectly criticized Newsweek for using highly unrealistic weightings in their RV models (bold text is eRiposte emphasis):

Two new polls came out immediately after mine (as of this writing) by the nation's leading weekly news magazines. Both Time's 52% to 41% lead among likely voters and Newsweek's 54% to 43% lead among registered voters give the President a healthy 11 point lead. I have not yet been able to get the details of Time's methodology but I have checked out Newsweek's poll. Their sample of registered voters includes 38% Republican, 31% Democrat and 31% Independent voters. If we look at the three last Presidential elections, the spread was 34% Democrats, 34% Republicans and 33% Independents (in 1992 with Ross Perot in the race); 39% Democrats, 34% Republicans, and 27% Independents in 1996; and 39% Democrats, 35% Republicans and 26% Independents in 2000. While party identification can indeed change within the electorate, there is no evidence anywhere to suggest that Democrats will only represent 31% of the total vote this year.
...
This is no small consideration. Given the fact that each candidate receives anywhere between eight in ten and nine in ten support from voters in his own party, any change in party identification trades point for point in the candidate's total support. My polls use a party weight of 39% Democrat, 35% Republican and 26% Independent. Thus in examining the Newsweek poll, add three points for Mr. Bush because of the percentage of Republicans in their poll, then add another 8% for Mr. Bush for the reduction in Democrats. It is not hard to see how we move from my two-point lead to their eleven-point lead for the President.

I will save the detailed methodological discussion for another time. But I will remind readers that my polling has come closest to the final results in both 1996 and 2000.

UPDATE 10/7/04: Chris Bowers has a significant update on Party ID distribution.

Both Pew and Mystery Pollster, who is actually Democrat Mark Blumenthal, make the same argument about Party ID. Specifically, they claim Party ID is an attitude, not a demographic, and thus is not something a poll should be weighted by.
...
Instead, after reading through the data presented by both Pew and Blumenthal, and I have come to agree with them, but only in part. Independent Party ID is an attitude. By contrast, Party ID among Democrats and Republicans is remarkably stable and does constitute a demographic. Almost all of the shifts caused in Democratic and Republican Party ID is caused by independents either shifting to one major party, or shifting away from one major Party back toward independent. For example, here is the chart Pew produced to argue that Party ID is unstable:
...
The change is not taking place among Democrats and Republicans. Instead, the change is taking place within two groups of voters, "Independent Democrats" and "Independent Republicans," that a three-way Party ID question is unable to measure. According to the chart that Pew produces to argue that Party ID is not stable, what they end up instead proving is that "independents" when considered as a block, are unstable. Of the 16% they identify as changing Party ID from October 23-26, 1988 until November 9-11, 1988, 15% were changing from independent to one of the two parties or from one of the two parties to independent. Only 1% of the people questioned shifted from one party to the other.

Pew's 1992 and 2000 graphs show similar lack of movement among partisans. From June 1992 until November 1992, only 4% of the population shifted from one party to the other, while 22% shifted from independent to one major party or from one major party to independent. In 2000, only 2% shifted from one major party to the other, while 18% shifted either from one major party to independent or from independent to one of the major parties.

Mystery Pollster goes on to cite the 2000 National Annenberg Election survey...The conclusion that NAES made from this was that Independent Party ID "was not stable over time."

Fine. Independent Party ID is not stable over time. Pew demonstrates the same. However, at the same time there is no evidence of any significant shift from Democrats to Republicans or vice-versa. Democrats are not becoming Republicans and Republicans are not becoming Democrats. Even the data used to argue that Party ID is an attitude and not a demographic demonstrates that almost every change in Democratic and Republican Party ID occurred as a result of Independents shifting one way or the other.

One could stop at this point and assume that Party ID is stable among partisans, but not among Independents. However, this would be wrong, since the shift these polls are measuring among Independents is a mirage generated by a three-way Party ID question in what is at least a five-way Party ID country. Specifically, what most polls ignore is that while around 40% of the population consider themselves to be "Independents," around 75% of Independents lean toward one party or the other and can be accurately considered "Independent Democrats" or "Independent Republicans". However, by forcing those two groups of leaners to choose between their two allegiances, especially in a dataset of 1,500 or less, the three-way question, and possibly wavering base energy, will create the appearance of a shift that has not taken place.
...
Overall, what I think this tells us about polling is, basically, that we lack adequate information to know how accurate polls are in terms of Party ID (unless, like Gallup, there is an extreme outlier).
...
More importantly, what I think this shows is the folly of trying to appeal to independent voters as a campaign tactic, instead of consolidating and energizing your base. There just are not many "Independent Independents" out there (maybe a little than 10% of the registered voter population), and they don't vote very often anyway. I wouldn't be surprised if "Independent Independents" make up only 6-7% of the vote in national elections. Why bother spending so much time appealing to such a small group when keeping your base intact is far, far more important?

Howard Dean may have had some image problems, but at least he had the right strategy: appeal to the base. Register the base. Energize the base. Consolidate the base. This is especially true for Democrats, who have a larger base than Republicans. Our problem has historically not been with independents, but with not holding our own base together and with turning out at a lower rate than Republicans. When has Democratic turnout ever equaled Republican turnout? When has a Democratic nominee ever held his base together as well as the Republican nominee? 1964 is probably the last election that meets both criteria.

Howard Dean had it right, and Karl Rove has it right as well. Woe be unto us if we continue to mock the way Bush is only going after his base, while Kerry seeks "swing voters." Bush has a solid strategy, but I worry that Kerry has been chasing a mythical beast.

Also see Dailykos.

The lesson? RV results are largely meaningless unless weighted accurately by party ID. Democratic and Republican Party ID appears to be far more stable that Independent Party ID. This seems to be less due to people switching all over the place than how the polls are conducted and questions are asked. 

2.3 Margin of Error (MoE)

I already mentioned MoE briefly in Section 1 and will simply reiterate here what I said there. Terms like "leading"/"ahead" and "trailing"/"behind" are OFTEN NOT statistically accurate terminology because they ignore the margin of error (MoE). If you included MoE, the vast majority of these results suggest a close and unpredictable race. If you want to understand how to interpret MoE (and also confidence intervals which are rarely mentioned even though they are relevant), here are some useful links: 
StudyWorks, CampaignDesk, Secular Blasphemy

2.4 The Nader Effect

Some polls continue to report results with Republican, er Reform Party, er Green Party, er Independent candidate Ralph Nader in the mix. Of course, to their credit most of them seem to be reporting results with and without Nader. The point though is that results from national polls that include Nader need to be viewed with some amount of skepticism because Nader is not going to make it to the ballots in many states. 

GoKeever at DailyKos did a systematic Nader survey on 8/5/04 which provides a window into the Nader effect. Since then, Chris Bowers at MyDD provided a simple summary of the Nader effect as of 8/30/04, which I reproduce here (the highlight is eRiposte emphasis).

Now that Nader is off the PA ballot, I recalculated the percentage of people who can and cannot vote for him. Considering the success of legal efforts either to have Nader removed from ballot or have his ballot petitions rejected, and the abject failure of legal challenges to have Nader placed on the ballot, Nader's ballot status looks bleak indeed. Unless he has actually been certified on a state ballot and no legal challenges remain, his status should be considered TBD. If his petition has been rejected, then he should be considered off the ballot.

	Off	On	TBD
% of US pop over 18	48.25	5.83	45.92
% of 2000 Nader Vote	36.44	5.90	57.66
Electoral Votes	239	37	262

Nader has been verified on the ballot in six states where there no longer remain legal hurdles for him to clear: AR, DE, IA, MT, NJ and SD. Nader is off the ballot in 14 states: AZ, CA, GA, ID, IN, IL, MD, MI, MO, OK, PA, NC, SC and TX. (source)

Quite a lot of territory remains to be determined, but so far Greens and Democrats have joined forces to rout Nader off the ballot. This is just one wing of a huge Democratic effort this year--an effort that seems to be dwarfing what took place in 2000--and I for one am enjoying watching it succeed. The state with the final ballot deadline is Minnesota, on 9/14. Much more will be known by then.

Incidentally, [Libertarian Candidate] Badnarik will be on every single state ballot, with the possible exception of Oklahoma.

The bottom line? Nader's impact will be seen only in some states, and although that includes some battleground states, there are other battleground states where he will be off the ballot. Including his name in National Poll results (without Badnarik) could provide an additional, inaccurate skew towards Bush.

2.5 Overseas Absentee Ballots

The website Electoral Vote suggests overseas voter registration may be above average this election (see below). Considering how overseas absentee ballots gave Bush an edge in Florida in 2000, and how close the races are this year, we should not forget these ballots! They could swing some states in a manner not captured by the national or state polls. Given the significant support for Kerry (over Bush) outside the U.S., it is possible that Kerry may benefit at least to a small extent from these votes in 2004, but it is hard to say right now.

Here's Electoral Vote:

I have it on good authority that overseas voters are registering in huge numbers this time, maybe double or triple 2000. I was told that the number of people who showed up at the Democratic party caucus in England earlier this year was 10 times what it was in 2000, ditto in other countries. Americans overseas vote in the state they last lived in, even if that was decades ago. There are about 7 million overseas Americans and probably about 5 million are over 18. In Florida, it was the overseas absentee ballots that swung the election. I believe that something like 8% are military, but the rest are students, teachers, artists, government workers, business executives, spouses of foreign nationals, missionaries, retirees, and more. What is significant here is that these people represent a lot of votes and are not included in any of the polls. Nobody knows if they are largely Democrats or Republicans, but their votes could be one of the big surprises of this election.

Bottom line? Don't forget absentee ballots and their ability to swing close races. If you know people outside the U.S. ask them to vote!

2.6 How do Undecideds Split Their Votes?

No one I know on the web has done more work to understand this aspect than Chris Bowers at MyDD. Chris outlined his thinking and superb analysis in a number of recent posts: here, here and here. The last of these articles provides the rule that Chris is now using (bold text is eRiposte emphasis):

Lately, I have been spouting off about how undecideds [Und.] tend to break about 60% for the challenger [Chal.] and 40% for the incumbent [Inc.] in the final week of the campaign. I was basing this off of a quick estimate from research I had done five weeks ago. However, I had never actually sat down and crunched the numbers to truly determine the average undecided swing in the final week of a campaign in elections involving an incumbent. Now I have...

Using my own research as well as research sent to me by Nick Panagakis of the National Council on Public Polls, I have gone through 451 poll results since 1976. In all 451 cases, the poll was in the field for at least one day that was within seven days of the election. In every case, it was the final poll taken by the polling firm for the campaign in question. Also, I do not believe that any internal or partisan poll results were used. Unfortunately, outside of the Presidential race, I was severely lacking in data from 1996.

...The results were as follows:

Year	Polls	Und.	Inc.	Chal.
President	28	2.4	14%	86%
1976-88	155	11.8	20%	80%
1994	101	11.2	35%	65%
1998	76	10.1	27%	73%
2000	31	8.6	40%	60%
2002-4	60	7.5	42%	58%
1992-04	283	8.9	34%	66%
Total	451	9.7	28%	72%

Here are my first thoughts on this data:

• Is the incumbent rule weakening? With the exception of 1998, where Democrats kicked ass on GOTV, it has been a slow progression toward parity. Then again, this chart has holes, such as a complete lack of polls from 1996, which prevents one from drawing such a conclusion with confidence.
• The number of undecided voters is clearly going down. This year, I see no reason not to expect more of the same. As the two parties slowly become primarily ideological, rather than regional and ethnic coalitions, the difference to voters is becoming starker. As a result of this, more people have made up their minds going into the booth. We really are becoming a polarized nation--that isn't just pundit bullshit.
• The Presidential sample stands out for its extremely small movement from final polls until election night. Even though undecideds break overwhelmingly--better than 6 to 1--in favor of the challenger in a Presidential race, pollsters seem particularly adept at national trial heats in Presidential races. Probably because of the extreme amount of national attention given to the Presidential race, far more people have made up their minds going into the booth than in other elections. While we should not expect significant movement from the final polls on November 1 to the final results on November 2, whatever small movement there is will be almost entirely for Kerry.
• Considering how accurate polling firms tend to be in national elections, in my presidential projections I will now include likely voter models from polling firms that do not include registered voter models. Where available, I will still use the registered voter models until the final week, however. Also, I will only include trial heats that push leaners, and give 80% of the remaining undecided to Kerry. There is simply no way that even 5% of the country will be truly undecided between Bush and Kerry going into the booth.
• Should the 1992-2004 average of 66 to 34 in favor of the challenger, or 2 to 1, be considered a benchmark at which point GOTV efforts make the rest of the difference? At least for now, I'm going to run with this possibility.

It seems like it would be fine to use these numbers as a means of estimating how undecideds will break in any given non-presidential election, as long as it is sold purely as an estimation. These figures are averages and invariably there will be results that break sharply from that average. 66-34 is where the smart money, the house money, should be. You will win over the long term betting on that sort of a split, but in any given wager you might get burned.

UPDATE 10/7/04: Chris Bowers turns in more analysis! Bold text is my emphasis.

Last month I reported an update to my research Incumbent Rule. Through an examination of 451 final pre-election polls in campaigns with an incumbent from 1976 until today, 72% of all undecideds broke toward the challenger while 28% of all undecideds broke toward the incumbent. From 1992-2004, that percentage had dropped slightly. Over 283 polls, 66% of the undecideds broke for the challenger while 34% broke for the incumbent.

However, these figures are the average of a wide variety of congressional, gubernatorial and Senatorial races. Even in these campaigns, while fairly well known, the incumbents rarely suffered from 100% name recognition and widespread crystallization of opinion. When looking at only final pre-election polls in the ultimate incumbent versus challenger campaigns--Presidential campaigns--the percentage of undecideds to break for the challenger increases sharply. In 28 final pre-election polls in Presidential elections featuring an incumbent from 1976-1996, a whopping 86% of undecideds broke for the challenger, while only 14% went for the incumbent.
...
As Guy Molyneux shows in the American prospect, incumbent Presidents finish almost precisely where their final poll results project they will finish:

There have been four incumbent presidential elections in the past quarter-century. If we take an average of the final surveys conducted by the three major networks and their partners, we find that in three of these the incumbent fell short of or merely matched his final poll number, while exceeding it only once, and then by just a single point (Ronald Reagan). On average, the incumbent comes in half a point below his final poll result.

Year  Incumbent  Final Poll %  Actual Vote %
1996  Clinton	   51	       49
1992  Bush	   37	       37
1984  Reagan	   58	       59
1980  Carter	   42	       41

The numbers for challengers look quite different. In every case, the challenger(s) -- I include Ross Perot in 1992 and 1996 -- exceed their final poll result by at least 2 points, and the average gain is 4 points. In 1980, Ronald Reagan received 51 percent, fully 6 percentage points above his final poll results.

Looking at just Gallup, Mystery Pollster delivers even more bad news for incumbent Presidents:

[T]he final Gallup projections (sans undecided) show an intriguing pattern: In the presidential elections since 1956 that featured an incumbent, Gallup's final projection of the incumbent's vote exceeded the incumbent's actual vote six of eight times. The only exceptions were Ronald Reagan in 1984 and George H.W. Bush in 1992, and then by only 0.2% and 0.7% respectively. On average, Gallup's projection of the incumbent's vote has averaged 1.3 percentage points greater than the actual result. Obviously, without seeing the raw results we can only speculate, but this pattern suggests that Gallup has allocated too many of the undecided over the years to incumbents.

It is quite clear that incumbent Presidents receive very few undecideds from the final polls until the final result. Considering this, Molyneux offers one way to track the Presidential race:

Think of it this way: The percentage that Bush receives in polls represents his ceiling of support; he may get a little less, but won't get more. In contrast, Kerry's percentage represents his floor, and he will almost certainly do better on election day. Assuming that Ralph Nader and other minor candidates will receive about 2 percent -- which is what current surveys suggest -- 49 percent becomes the critical line of demarcation in this election. If Bush can get to 50 percent or above in the polls, he should be able to win. At 49 percent -- where he is today -- we're probably looking at another photo finish, lots of recounts, and narrow state-by-state victories dictating the Electoral College outcome. And below 49 percent, Bush is almost certain to lose.

The bottom line? It appears pretty conservative to assume that undecideds will probably split 2-1 in favor of the challenger. If Kerry builds a good case with Americans, he can potentially get the vast majority of the undecideds.

2.7 ASIDE: Telephone polling -- What about Cell Phone Owners?

Recently, a Jimmy Breslin column in Newsday questioned the relevance of land-line telephone polling today, given the profusion of cell phone use. Electoral-Vote and John Zogby weigh in on this.

Here's Electoral Vote:

How do pollsters decide who to call? The simplest way to get a random sample is to pick the area code and exchange to be sampled and have the computer dial the last four digits at random. In the trade this is called RDD--Random Digit Dialing. It is very easy to do but unfortunately has some problems. Many numbers will be business phones, hospitals, police stations, teenagers, and other undesirables. Also, with number portability, the pollster may be getting Manhattan, KS instead of Manhattan, NY. Consequently, RDD is not used much any more.

Instead, pollsters buy lists of phone numbers. In some states, the government sells voter lists. In other states, lists of residential customers can be purchased from telephone companies. Commercial companies also sell lists. There are many sources. These are commonly used. Pollsters do not want to call cell phones for several reasons: First, cell phone users may be in meetings, restaurants, cars, or other places where they are not free to talk. Second, since cell phone users are usually paying for air time, they don't want to waste it and might break off half way, resulting in a useless interview. Third, that number in Colorado the pollster called may actually be a student who took her cell phone to college in California. Fourth, sound quality may be poor and questions or answers may be misunderstood.

Fifth, and most important, it might cost the pollster a lot of money: automated calling of cell phones is illegal in the United States and there is a substantial fine for breaking the law. US Code Title 47, Chapter 5, Subchapter II, Part I, Sec. 227 reads:

It shall be unlawful for any person within the United States--
(A) to make any call (other than a call made for emergency purposes or made with the prior express consent of the called party) using any automatic telephone dialing system or an artificial or prerecorded voice--

(iii) to any telephone number assigned to a paging service, cellular telephone service, specialized mobile radio service, or other radio common carrier service, or service for which the called party is charged for the call.

On the positive side, pollsters are exempt from the "Do not call" list, but most people do not know that.

But using land lines is no bed of roses either. Many people use caller ID or their answering machine to do call screening. Busy young professionals are rarely home from work before 11 p.m. whereas lonely old people are only too happy to talk to the nice young girl who seems to care about what they think. Calls made at 2 p.m. are going to oversample housewives, and so on. All these effects lead to biases. To correct for them, pollsters conduct exit polls of voters leaving the polling place on election day to get a good idea of the statistical make up of the electorate for next time. These data are used to correct the polls. For example, if the exit polls show that 10% of the voters in some state are African Americans and in a state poll of 600 people by accident only 30 are African Americans (5%), the pollster can just count each African American twice. This process is called normalization. It was proposed for use to correct undercounts in the 2000 census but was rejected by the courts on the gronds that the constitution calls for an "enumeration" of the population, not a statistical model of the population.
...
Now getting back to Breslin's column, in theory he is right that pollsters will miss people who have only a cell phone, but of the 169 million cell phones, most have a land line as well. It is estimated that 5% of the population is cell only. And most of these don't live in battleground states. And then an effect occurs only if cell-only users differ from land line users in their political preference. Thus the error introduced by missing the cell-only customers is probably smaller than the error introduced by missing the overseas voters. But it is there and in a close election, it could matter. Here is Zogby's response to Breslin's column. Robert Landauer wrote a good piece on the accuracy of polling is his Aug. 31 column. Worth a look.

More from Zogby's response:

First of all, I still conduct telephone polls.  The reality is that polling on the telephone is becoming more difficult; caller id and the widespread use of cell phones are affecting response rates.  That said, I feel that representative samples can still be achieved on the phone.

Second, I stand by both my telephone and interactive results.  I have yet to see evidence that the situation has gotten to the point where telephone surveys are unusable, and I am equally confident that my interactive surveys have reached a point where they are valid.

Third, cell phones do pose a problem for the polling industry, but not to the level Mr. Breslin feels.  It is illegal for polling firms to call cell phones, coupling that with the rapidly increasing rate of cell phone use and the gradual decrease of land lines, the polling industry will face a crisis within a decade.  For now, the 170 million cell phones are largely duplicates and triplicates of landlines.  Also, many of the people Mr. Breslin cites as missing because of cell phones, are notoriously difficult to each, no matter the circumstance.

To clarify, my statements to Mr. Breslin were aimed at pointing to growing problems in the industry.  As an industry, we must adapt to the future or face extinction, because the telephone will not always be a valid method of conducting random samples.

The bottom line? None really (at least right now). The polling industry has to keep pace with technology.

2.8 WHAT MATTERS: The Electoral College!

Needless to say, National Polls are useful more to identify broad trends and situations where one or the other candidate has a commanding lead. In close races such as this year's what really matters are the state-by-state results which determine the electoral college apportioning of the Presidential vote.

Although the website Electoral Vote is relied upon by many to predict the race, I prefer Chris Bowers' approach since his methodology, I believe, is better sanitized to remove the kinds of errors I have highlighted earlier in this page. Chris' basic assumptions are outlined here:

• Trial heats are generally accurate projections of the Presidential Election.
• The most accurate trial heats weight data by Party ID.
• Turnout by Party ID for this election will be 39%D, 36%R and 25%I/O.
• Likely voter models are more accurate than registered voter models in the final week of an election. Before then, registered voter models are better at gauging the current state of the campaign.
• In elections with incumbents, undecideds generally break 66-34 in favor of the challenger.
• The third-party vote in the election will total 2%. (that’s my guess)

The page I most often refer to to assess the state of the Electoral Vote is Chris' General Election Cattle Call on the Swing State Project.

3. RECENT POLL RESULTS ADJUSTED FOR FLAWS

3.1 Gallup

Let's start with Gallup and Ruy Teixeira (9/17/04) (bold text is eRiposte emphasis):

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 RVs: +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.
...
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!

Chris Bowers has more on "Gallup's Shame", where he points out:

It is pathetic and unacceptable for a "non-partisan" polling firm to be produce the outlying poll in favor of Bush in fourteen of its last sixteen polls. The odds of this happening at random are around one in 14,000. Considering those odds, the far more likely explanation for all these outliers is that Gallup's polling methodology is inherently structured in favor of Bush. Whether or not it is intentional, I do not know. However, I do know that Gallup's polls are connected to the largest news outlets in America of any poll, both in terms of print (USA Today is the largest circulation newspaper in the country) and cable news (CNN has more viewers than Fox, they just watch for shorter periods of time). I also know that sensational headlines sell. I further know that Gallup's chairman is a Republican donor.

This is a shameful state for the oldest and most respected polling organization in the country. Shame on you Gallup.

UPDATE on 9/20/04:

Via Ruy Teixeira, I see that John Harwood has an article out in the Wall Street Journal trying to understand the discrepancies in the polling results. Gallup's Editor-in-Chief of Polling is quoted here defending Gallup's Likely Voter model:

"We're open to any scientific evidence that would point to our modifying our likely-voter model," responds Frank Newport, editor-in-chief of the Gallup Poll. Mr. Newport says so far he hasn't seen any.

Well, could there be more nonsense from the head of a firm that once had a good reputation? I'm glad Harwood points this out in the very next sentence (bold text), which elicited a weak defense from Newport. 

In 2000, Gallup's election-eve sample of likely voters showed Mr. Bush leading by two percentage points over Al Gore. Its registered-voter sample, showing Messrs. Bush and Gore neck and neck, was closer to the actual Election Day results. But Mr. Newport notes that in 1996 the likely-voter model more accurately forecast the size of Bill Clinton's victory over former Senate Republican Leader Bob Dole. 

Ha ha! Newport gets the "scientific evidence" he asked for and still clings to their discredited methodology based on 1 data point. Why discredited? Well, as Teixeira has pointed out: 

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!

3.2 CBS/New York Times

Here's Teixeira on the CBS/New York Times poll (9/17/04):

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.

Jerome Armstrong has more at MyDD:

On the heels of Gallup. Another poll that says there's been a 5-10% partisan swing toward Republican turnout for 2004, with what proof? Regarding the NYT/CBS poll, it's composed of 36% Republicans and 31% Demoratic voters. However, that is not representative of the actual turnout:

	Democrats		Republicans	Independents
1992	34%		34%		33%
1996	39%		34%		27%
2000	39%		35%		26%

And from Atrios, on the NYT/CBS Poll results. It includes a question asking who the respondent voted for in 2000.

Answers:
Gore - 28
Bush - 36
Buchanan - 1
Nader - 1
Voted, won't say - 1
Didn't vote - 32

There's something happening here, and what it is aint exactly clear. One thing is clear, if you look at the poll numbers prior to the Republican-weighted sampling done by Gallup and CBS/NYT's the race is even.

Bush 36 Gore 28 Didn't Vote 32! What a sample! Good grief.

UPDATE 9/20/04:

Via Ruy Teixeira, I see that John Harwood has an article out in the Wall Street Journal where he points out that the large Bush lead in the latest CBS/NYT poll is because of the high percentage of Republicans in this poll:

Last week's CBS sample, in a mirror image of Pew's, contained four percentage points more Republicans than Democrats. Because this polarized contest has left roughly nine in 10 adherents of each party supporting its nominee, such variation in the number of Republicans and Democrats surveyed has an unusually large impact on polling outcomes.

In a close race, in fact, that can make the difference between an apparent dead heat and a solid lead for one candidate. 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."

3.3 Other Polls: ABC, Fox, IDB, Newsweek

Chris Bowers of MyDD has, as usual, done some astonishing work examining the history of Party ID in polling and applied recent Party ID stats to polls that do not currently weight by Party ID - such as ABC, Fox, IBD and Newsweek. No prizes for guessing the result! Here's Chris (and he starts here with the CBS/NYT poll to show how Kerry's position has actually improved if you fix the Party ID ratios), and bold text is eRiposte emphasis:

On September 9th, CBS News released results from its post-RNC poll, conducted September 6th-8th, that was very good news for Bush, and very bad news for Kerry. In fact, with the possible exception of the June Harris poll, this was the worst poll for Kerry this entire election cycle. In a poll that had a Party ID weighting of 33.5% (354) Democrats, 32.1% (340) Republicans, and 34.5% (365) Independents, John Kerry was quite simply not where he needed to be in order to win the election. Kerry polled at only 80% among Democrats, while Bush polled at 90% among Republicans. Even more worrying for Kerry, Bush led 48-39 among independents. While not overwhelming, this poll clearly showed a Bush lead. Even if the Party ID figures were re-weighted in this poll to reflect either 1996 or 2000 turnout, Bush still led by five points.

This week, the CBS poll, conducted from September 12th until September 16th, showed the exact same overall result: Bush leading Kerry 50-42 among registered voters in a two-way trial heat. On the surface, this may not seem like good news for Kerry. It may even appear as though Bush's convention bounce is sticking. However, this poll was actually very good for Kerry and, as we shall see, highly unusual because it showed real movement. While Bush leads Kerry by exactly the same 50-42 margin among registered voters, among all three different Party ID groups, Kerry actually improved his position. Among Republicans, Bush fell from 90 to 87. Among Democrats, Kerry moved up to 83 from 80. Among independents, Kerry went from nine points down to being down by only a single point, 43-42. While the overall results look exactly the same as those form last week, CBS admits to having a potentially spoiled sample...What happened was that the 33.5% Democrat, 32.1% Republican, 34.5% Independent Party ID sample changed to a 36% Republican, 32% Democrat, 32% Independent Party ID sample. Kerry's gains within the Party ID groups were wiped out by a shift of the relative importance of the Party ID groups within the poll as a whole.
...
Two polling firms, Harris and Pew, have conducted Party ID surveys over a span of thirty-five and seventeen years respectively...

Pew's 2004 survey was conducted in June and July, and contained over 19,000 people in their sample (under 0.5 MoE). Although I do not know for certain, it is reasonable to assume that their previous surveys were conducted among similar sample sizes. Since 1990, Republican self-identifiers have varied between 27 and 30 percent of the population, except once in the immediate post-9/11 survey when they rose to 31%. Democratic self-identifiers have varied between 30 and 34% of the population, except once, immediately before 9/11, when it rose to 35%. In eight of the sixteen surveys since 1990, self-identifying Democrats made up 33% of the population. This chart does not just show slow movement in Party ID--it shows glacial movement.
...
In 2003, [the Harris] poll included more than 6,000 registered voters, and thus has a slightly higher Margin of Error than Pew, but the MoE is still under plus or minus one. According to Harris, since 1990 Republicans have varied from between 28 and 33% of the population, while Democrats varied from between 36 and 38% of the population before dropping to 34 and 33 the last two years respectively. In those two years, Republicans also dropped while Independents / Not Affiliated / Other saw an increase. Further, with the exception of 2002, the Democratic edge in Party ID ranged between 5 and 9 points--a mere four-point swing. Overall, this chart also shows extremely gradual movement, never once in a year seeing the same kind of movement CBS recorded in a single week. Considering this, there is simply no circumstance under which the shift recorded by CBS, a poll with a Margin of Error of plus or minus, 3.5 can be understood as accurate.

Party ID is indeed a variable rather than an absolute constant. However, even over periods of years in surveys with miniscule margins of error, it is a slow moving, incrementally shifting variable. Considering this, from the perspective of week to week polling in the final two months of a Presidential election, in order for such polling to truly gauge the state of the campaign, it should be treated as a constant. Given this, it is my position that polls that weight their results according to a fixed Party ID model offer a more accurate picture of the campaign than those that do not weight their results by Party ID (or other demographics).
...
I have gathered together Party ID data from twelve recent polls. In six of the polls, the Economist, Harris, ICR, Pew, Rasmussen and Zogby, the samples were weighted to fit demographic and / or previous turnout models.
...
The similarity between the results in these six polls is remarkable. The race varies from Bush up four to Kerry up one, with no two polls disagreeing about Bush's raw score by more than three points or Kerry's raw score by more than four points. On average, Bush leads by less than two points (47.3-45.5).

For the sake of comparison, I have also been able to track down the internals of six recent polls that do not weight their results according to Party ID or other demographics: ABC, CBS, Fox, Gallup, IBD / CSM and Newsweek.
...
These polls show significantly more variance than the other six. But what is perhaps even more remarkable is the variance they show among themselves.
...
Interestingly, had Party been weighted in the six most recent unweighted polls, they would look almost exactly like the six recent weighted polls:

Table Five
* = likely voters
      Bush Kerry Date
ABC    48    47   9/8
CBS    47    46   9/16
Fox*   46    47   9/8
Gallup 48    48   9/14
IDB    47    47   9/12
News   46    47   9/10

Yikes!

Bottomline? Unweighted polls over short time periods are largely useless, if there are significant shifts in the Party affiliation ratios of those polled. 

4. CONCLUSIONS AND LESSONS FOR KERRY SUPPORTERS

As you can tell from the data presented here, it is very clear that polls currently showing huge Bush leads have deep methodological flaws (either from an LV standpoint, RV  standpoint, or both). The fact of the matter is that nationally, properly weighted polls show the race to be essentially close and dead even*. This is evident in Chris Bowers' always useful General Election Cattle Call as of 9/17/04.

Lessons for Kerry supporters?

1. Stop the suicide watch and the hand-wringing. 

2. Stop panicking over every other poll. Do not trust the mainstream/conventional media or their nonsensical reproduction of bogus polling data. Remember this?

October 27, 2000
...
Today's CNN/USA Today/Gallup tracking poll continues to give George W. Bush an advantage over Vice President Al Gore...

While not a prediction of the voters' choice in November, Friday's results show Bush garnering 52 percent of the vote and Gore drawing 39 percent. The survey of 851 likely voters was conducted October 24-26 and has a 3.5 percentage point margin of error.

A CNN/Time poll also released today gives Bush a 49 percent to 43 percent edge over Gore, which is statistically in agreement with today's CNN/USA Today/Gallup tracking poll, given the polls' margin of sampling error.

However, do contact the media and keep correcting them as politely as possible, for the nonsense they publish or show on TV - whether it is parading the Swift Boat Liars for Bush or ignoring Bush's countless flip-flops (both staples at Faux News and CNN) or building nonsensical stories using flawed polls, etc. 

3. Keep your eye on the ball: Participate in Get-Out-The-Vote (GOTV) efforts and donate to the DNC.

4. Stop complaining incessantly about Kerry's every move (or lack thereof). Kerry knows what he is doing. He is just up against the most powerful and corrupt political machinery in the history of the United States and a mainstream media that is largely compliant with the corruption or willing to play he-said-she-said to propagate GOP fakery unchallenged to the electorate (not to mention talking heads and columnists who are largely inept or corrupt - read the Daily Howler if you want proof or visit this page).

5. Liberal Oasis has some additional advice:

Unfortunately, Kerry's own messages were hemmed in by the mind-numbing stupidity of the Beltway Dems, whining to the press about strategy throughout the long weekend.

Instead helping with strategy by amplifying Kerry's messages.

(See LO on 4/28 for the most recent time this page called on whiny Beltway Dems to shut up.)

For the last time (this year) Establishment people: this is all-out combat, and we have one leader.

You have a problem with the campaign, then tell it to the campaign.

Telling the media is for your ego. Not for the cause. Not for the country.

There are 56 days left. Get your damn game faces on, and get behind the big guy.

4.1 Some relevant Election 2000 History

a. Remember this?

October 27, 2000
...
Today's CNN/USA Today/Gallup tracking poll continues to give George W. Bush an advantage over Vice President Al Gore...

While not a prediction of the voters' choice in November, Friday's results show Bush garnering 52 percent of the vote and Gore drawing 39 percent. The survey of 851 likely voters was conducted October 24-26 and has a 3.5 percentage point margin of error.

A CNN/Time poll also released today gives Bush a 49 percent to 43 percent edge over Gore, which is statistically in agreement with today's CNN/USA Today/Gallup tracking poll, given the polls' margin of sampling error.

b. Don't forget that Gore was "trailing" Bush in the polls in October 2000, literally until Election Day:

The following charts show the difference between the votes in favor of Gore vs. those in favor of Bush in the polls conducted before the 2000 election. Note that the votes corresponding to "Other/Unsure" are excluded - so keep those in mind, as well as the error bars, when you interpret the data. My comments on the data appear below the charts.

 

5. A DEAD-EVEN RACE: FEEDBACK TO SENATOR KERRY AND THE DEMOCRATIC PARTY

Kerry and the DNC of course need to translate the dead-even race to a win for Kerry - this is going to be a challenge especially if Bush springs an "October Surprise" or two. But, I will not try to second-guess the Kerry camp's strategy at this point (as Liberal Oasis has wisely advised) since they know the internals of all these polls and the pulse of the electorate better than anyone else. What is clear is that although Kerry seems to be drawing ahead of Bush when it comes to Independents, the percentage of Democrats who plan to vote for him seem smaller than the percentage of Republicans who plan to vote for Bush. In other words, it appears Kerry needs to shore up his base substantially and motivate them to come out to the polls on Election Day, which makes GOTV efforts all the more critical. (There is at least some evidence that Kerry's lead among women voters, traditionally the Democratic base, has slipped - the latest Pew poll shows his lead down to 5% amongst women, down from 10% in August).

As I wrap up this piece, there is one piece of advice for the Kerry campaign though. They HAVE to replace their absolutely inept media representatives (who only add to the damage caused by the inept and lazy talking heads themselves) with competent ones. The nonsense from some of their representatives appearing on TV has gone on far too long and the Kerry campaign is making Kerry's life more difficult by not choosing the right people to appear in these shows - people who can clearly rebut/refute GOP talking points/spin/lies and articulate Kerry's position/platform rather than question it.

6. APPENDIX: Why do "leading" media outlets continue to publish stories using highly questionable results?

Well, for the same reasons that they continue to "stand by" highly discredited or fake stories on other topics. For self-promotion and to protect their "reputation". No surprise here. 

I'll let Steve Lovelady at the Columbia Journalism Review's Campaign Desk handle this through the example of the CBS/NYT poll:

Campaign Desk took a look at this subject -- when polls go bad -- yesterday, but today's brazen display by The New York Times compels us to revisit it.

The Times brings us yet another poll measuring public preferences in the campaign for the presidency (following four other wildly-varying polls this week), and decides that this one deserves play as its lead story on page one.

Why?

As nearly as we can tell, the only "why" is this: The Times itself (along with CBS) commissioned this particular poll.

A little history: Earlier this week, a Pew Research Center poll taken between September 8 and September 10 gave President Bush a whopping 15-point lead among likely voters. Then on Thursday, Pew reversed itself by announcing that a poll taken from September 11 to September 14 found Bush leading among likely voters by all of one point. And a Harris Interactive poll released the same day found the presidential race essentially tied. Further compounding matters, a Gallup poll released the same day gave Bush an eight-point lead among registered voters and a 13-point lead among likely voters.

Finally comes the new Times-CBS poll, indicating Bush has a nine-point lead among likely voters.

Got it? It's Bush by 15 points; no, make that one point; oops, call it a dead heat; no, it's actually a 13-point lead; or is it nine points?

To its credit, The Times does run a sidebar by Carl Hulse inside the paper, offering up pollsters' explanations as to why nothing they say makes any sense. Nonetheless, there The Times is, offering up column one of page one to tout its own poll.

Running a sidebar on page A10 that says that all of these contradictory polls should be taken with a grain of salt the size of a bowling ball, but still using your lead story to shill for the particular poll that you happened to pay for, just doesn't cut it.

This is not journalism; it's self-promotion, and we say to hell with it. [eRiposte emphasis]

Amen.

FOOTNOTE added on 9/23/04:

Sometimes I use the words dead even or "tie" or "tied" in this article to represent a "tossup". In hindsight I realize I should have used the word tossup, but I don't think "tie" or "dead even" is too far off the mark.