The latest Democracy Corps survey, conducted Oct 23-25 Shows John Kerry leading George Bush 49-47 in their national sample and 52-45 in the battleground states. The poll also found that Kerry is ahead by 22 pts among new voters and includes substantial additional information on the latest trends among population subgroups and target voters.
Thanks for signing up!
TDS Strategy Memos
Latest Research from:
tds-wwc-roundtable
Editor’s Corner
By Ed Kilgore
-
September 6: Things That Make Joyful Democrats Jittery
Despite the recent return of Democratic optimism associated with the Harris-Walz ticket, there are a few stubborn fears that keep partisans awake at night. Here’s a review of four of them that I wrote at New York:
Democrats are in a vastly better state of mind today than they were a couple of months ago, when Joe Biden was their presidential candidate and his advocates were spending half their time trying to convince voters they were wrong about the economy and the other half reminding people about how bad life was under President Trump. While it’s possible this would have worked in the end when swing voters and disgruntled Democrats alike took a long look at Trump 2.0, confidence in Biden’s success in November was low.
Now that the Biden-Harris ticket has morphed into Harris-Walz, there’s all sorts of evidence from polls, donor accounts, and the ranks of volunteers that Democrats can indeed win the 2024 election. But at the same time, as Barack Obama and others warned during the Democratic National Convention, the idea that Kamala Harris can simply float on a wave of joy and memes to victory is misguided. She did not get much, if any, polling bounce from a successful convention, and there are abundant signs the Harris-Trump contest is settling into a genuine nail-biter.
While the September 10 debate and other campaign events could change the trajectory of the race, it’s more likely to remain a toss-up to the bitter end. And many fear, for various reasons, that in this scenario, Trump is likelier to prevail. Here’s a look at which of these concerns are legitimate, and which we can chalk up to superstition and the long tradition of Democratic defeatism.
Republicans’ perceived Electoral College advantage
One reason a lot of Democrats favor abolition of the Electoral College is their belief that the system inherently favors a GOP that has a lock on overrepresented rural states. That certainly seemed to be the case in the two 21st-century elections in which Republicans won the presidency while losing the national popular vote (George W. Bush in 2000 and Donald Trump in 2016). And in 2020, Joe Biden won the popular vote by a robust 4.5 percent but barely scraped by in the Electoral College (a shift of just 44,000 votes in three states could have produced a tie in electoral votes).
However, any bias in the Electoral College is the product not of some national tilt, but of a landscape in which the very closest states are more Republican or Democratic than the country as a whole. In 2000, 2016, and 2020, that helped Republicans, but as recently as 2012 there was a distinct Electoral College bias favoring Democrats.
To make a very long story short, there will probably again be an Electoral College bias favoring Trump; one bit of evidence is that Harris is leading in the national polling averages, but is in a dead heat in the seven battleground states that will decide the election. However, it’s entirely unclear how large it will be. In any event, it helps explain why Democrats won’t feel the least bit comfortable with anything less than a solid national polling advantage for Harris going into the home stretch, and why staring at state polls may be a good idea.
Recent polling errors
For reasons that remain a subject of great controversy, pollsters underestimated Donald Trump’s support in both 2016 and in 2020. But the two elections should not be conflated. In 2016, national polls actually came reasonably close to reflecting Hillary Clinton’s national popular-vote advantage over Trump (in the final RealClearPolitics polling averages, Clinton led by 3.2 percent; she actually won by 2.1 percent). But far less abundant 2016 state polling missed Trump’s wafer-thin upset wins in Michigan, Pennsylvania, and Wisconsin, largely due to an under-sampling of white non-college-educated voters. The legend of massive 2016 polling error is probably based on how many highly confident forecasts of a Clinton win were published, which is a different animal altogether.
There’s no question, however, that both national and state polling were off in 2020, which is why the narrow Biden win surprised so many people. Two very different explanations for the 2020 polling error have been batted around: One is that the COVID pandemic skewed polling significantly, with Democrats more likely to be self-isolated at home and responding to pollsters; the other is that the supposed anti-Trump bias of 2020 polls simply intensified. The fact that polls in the 2018 and 2022 midterm elections were quite accurate is consistent with either interpretation.
So we really don’t know if polling error is a given in 2024, or which candidate will do better than expected. A FiveThirtyEight analysis of polling error since 1998 shows a very small overestimation of the Democratic vote across 12 election cycles. It might be prudent, then, to expect that Trump might exceed his polling numbers by a bit, but not necessarily by a lot.
Fundamentals in election forecasts
A lot of election forecasts (or model-based projections) incorporate, to varying degrees, what are known as “fundamentals,” i.e., objective factors that are highly correlated historically with particular outcomes. There are models circulating in political-science circles that project presidential-election results based mostly or even entirely on macroeconomic indicators like GDP or unemployment rates. Others take into account presidential approval ratings, the positive or negative implications of incumbency, or historical patterns.
While forecasts vary in how to combine “fundamentals” with polling data, most include them to some extent, and for the most part in 2024 these factors have favored Trump. Obviously the substitution of Harris for Biden has called into question some of these dynamics — particularly those based on Biden’s status as an unpopular incumbent at a time of great unhappiness with the economy — but they still affect perceptions of how late-deciding voters will “break” in November.
The high chances of a chaotic overtime
A final source of wracked Democratic nerves is the very real possibility — even a likelihood — that if defeated, Trump will again reject and seek to overturn the results. Indeed, some MAGA folk seem determined to interfere with vote-counting on and beyond Election Night in a manner that may make it difficult to know who won in the first place. Having a plan B that extends into an election overtime is a unique advantage for Trump; for all his endless talk about Democrats “rigging” and “stealing” elections, you don’t hear Harris or her supporters talking about refusing to acknowledge state-certified results (or indeed, large batches of ballots) as illegitimate. It’s yet another reason Democrats won’t be satisfied with anything other than a very big Harris lead in national and battleground-state polls as November 5 grows nigh.
Mady-
Yesterday afternoon, I looked at all polls with post dates of October at race2004.net. I excluded those with either a GOP or Dem designation. I used 3 way data, if available, when Nader was on the ballot, 2 way data, if available, when Nader was not on the ballot. If a given organization did more than one poll, I used the most recent.
Of the 11 polls in Florida, 4 had Kerry in the lead, 4 Bush, 3 tied. The unweighted (by sample size) average was Bush up by 0.5%.
Of the 11 polls in Ohio, 7 had Kerry in the lead, 4 had Bush in the lead. The average was Kerry up .8%
Of the 7 polls in Wisconsin, 2 had Kerry in the lead, 3 Bush, and 2 were dead ties. The average was dead even.
Of the 8 polls in Iowa, 2 had Kerry up, 5 Bush, with 1 tie, average Bush lead of 2%.
Of the 5 in New Mexico, 2 had Kerry up, 3 Bush, average Bush lead of 2%.
Of the 4 in Minnesota, 2 had Kerry up, 2 Bush, with an average Kerry lead of 0.75%.
Of the 9 in New Hampshire, 6 had Kerry up, 3 Bush, with an average Kerry lead of 3%.
So…each of those polls has both sides winning some of the time this past month. It’s hard for me to think of any of them as solid. And Wisconsin is probably the least solid of any of them.
Those 7 seem the key battlegrounds now, with the possibility of an Arkansas or Hawaii creeping in.
[Others states for which at least one independent poll has shown both both sides in the lead of at least 1% at some point last month, using only the most recent poll from the organization…:
Colorado, Bush 5, Kerry 1
Oregon, Kerry 7, Bush 1
That’s it. Hawaii, Arkansas, Michigan, New Jersey, and Nevada seem to have been close here or there…]
If all but the 7 listed above are considered solid, you’re looking at 228 EV’s for Kerry, 227 for Bush. Kerry can get the other 42 with Ohio, New Hampshire, Minnesota, and either Florida, Wisconsin, or [Iowa and New Mexico].
Looking at today’s state polls, I think Wisconsin is up for grabs, while Iowa seems to lean slightly–but not irrecoverably–to Bush. Giving Nevada and especially Arkansas to Kerry seems a little overoptimistic, but not impossible. Probably the most important–Florida and Ohio–are split down the middle (if you discount what I hope is the out-of-whack Gallup result in Florida), with new polls divided on who is ahead, usually within the margin of error.
I think general polls of battleground states weight their sample so that more people are sampled from the largest states, with many more Floridians than Oregonians, for instance. But the samples from each state are almost certainly too small to reliably extrapolate results from individual states.
Zogby’s rolling tracking is showing a Bush lead on the basis of one day’s result having Bush +7. He’s up 3 which means the other days in the survey have to net a +4 to Kerry. Once that Bush +7 drops out in 2 days, it will be back where it was most likely.
I don’t know what’s going on with Tipp.
WaPo seems to be going Kerry’s way as is Rasmussen a bit (It was B+3) last week. So things are looking better.
Mady: There’s nothing addressing an analysis of the early votes that I know of. Are California and New York having early voting this year? I know many of the Southern States do and some of the northern states do not. Could that be the discrepancy? Also, I don’t know if that figure is an estimate based on exit polling, or an actual count.
In the state polls, Race.com has taken PA, NH, NJ, ME from undecided to Weak Kerry.
Mady: I checked most of the recent polls in WI. They’re mostly from SV, a republican outfit. I’ll wait for some more credible polls there.
Same is true of Iowa.
Mady,
Steve Soto at Left Coast has an analysis of early voters and it’s great for Kerry. Typically, early voters are elderly women in western states (most early voting states are in the West) – and trend heavily Republican. I believe Bush beat Gore among early voters in 2000 by 55-44 or something. So if Bush only leads by 4 now among early voters (and with a sample size of 170 voters anyway) then Bush is doing poorly.
http://www.theleftcoaster.com/
Also, a new Gallup poll (yes, Gallup) shows Kerry up 1 in Iowa. I have less confidence about Nevada but with OH and FL Nevada doesn’t matter.
Can anyone tell me a poll on how many 2000 bush supporters are not voting for him this time. Are the polls weighted in such a way that fails to this into account. There are many who simply will not vote for him this time. There are some who may not vote for Kerry but they are definitely not voting for Bush. This seems like it would be a significant thing to know.
I’m not sure, looking at the polls tonight most of which show Wisconsin, Iowa, and Nevada as pretty firmly Bush, where you find the information that those states will go to Kerry. Is it the potential for Democratic turnout, or some way of reading polling that I’m not doing? I’d love to believe with you, but I really think it is totally up in the air right now.
Also, no one addressed a question I had posted earlier–why is the voting that’s been done already showing a Repub lead? Any information on patterns, where people are voting early, which groups are?
probably Kerry
Wisconsin
New Mexico
Nevada
Florida
Ohio
Arkansas
Minnesota
Michigan
Pennsylvania
Oregon
New Hampshire
probably Bush
Colorado
Missouri
West Virginia
Virginia
Tennessee
Kerry is going to win BIG.
According to ABC (Noted Now) both campaigns are now putting resources into Hawaii, which seems to be genuinely in play. That is really not a good thing.
Kudos, commendations, and additional crown jewels to Ruy for being the best Donkey Rising, hands down, thumbs up!!
To Ruy and staff …. thanks a bunch! Many are greatful for your efforts. Including me.
Is it possible to give us some feedback on early voting? Either turnout or exit polling, or anything concrete? Anything more concrete on election turnout in general?
Again, thanks.
Do we know what Demo Corps. considers battlegrounds still? If it’s the entire 16 or so that originally existed (including WA, OR, etc), then this is not surprising or exciting.
Numbers that would be meaningful at this point should essentially be confined to PA, FL, OH, IA, WI, MN, NM, NV and possibly AR and CO. If these numbers include OR and WA and perhaps some others, then these aren’t that useful, since OR and WA will obviously inflate Kerry’s numbers since these are barely battlegrounds now. In fact, I’d argue that we could take PA off the list too to get a really good number of how Kerry is doing.
Here’s hoping that similar numbers exist for the states I mentioned.
-Jeff
Personally, I find the “battleground states” polls unenlightening.
America does not elect presidents nationally; it elects them on a state-by-state basis through the Electoral College. As we learned in 2000, it does not matter if a candidate racks up big vote-margins in one state (and win the national vote), only to lose a couple of others by mere handfuls and therefore lose the EC.
Subsequently, I honestly don’t care if John Kerry is building up a big vote-total in all 15 battleground states. Does this mean Kerry is doing well in all 15 battlegrounds? Or does it mean Oregon has gone whole-hog for Kerry while Kerry is losing the rest?
You can’t tell. Therefore, these polls are completely uninformative to me, and don’t give me any strategic information whatsoever.
I wish polling firms would stop doing them, and either do state-by-state polls which would give us real data by which to predict EC outcomes, or divert the resources to states (like Arkansas) which are seeing little polling but which might just well be coming into play (as the last Arkansas poll indicated).
Zogby has Bush up by 3 nationwide and also ahead in most of the battleground states…The brief lead Kerry had in the Rasmussen poll is gone and it is a dead heat again. Boy, I sure hope that all the good polls are the ones listed on EDM, but I can’t help but be a little nervous at this point. At least EDM gives me some hope though!