Aug.30

electoral-vote.com

Dem 51

   

GOP 49

Post-Convention Polls: Strange Brew

In the past week, two things have happened. First, the DNC is now over. Second, Bobby Kennedy Jr. suspended his campaign and endorsed Donald Trump. Will Kamala Harris get a bump from the convention? Will Trump inherit all or most of Kennedy's supporters? Or maybe both will happen and they will cancel each other out? Pollsters are going to really have to work overtime to brew their "secret sauce." Anyhow, polling is starting to come in now.

Emerson College polled the seven swing states over the Aug. 25-28 timeframe. In other words, the polls started 3 days after the DNC ended and 2 days after Kennedy endorsed Trump. It sometimes takes a while for people to update their opinions, even after some big event happens. They have to talk to friends, and so on. Still, this is a start. Here are the head-to-head horse-race numbers in the seven swing states:

What do we have? A really close race. Harris is ahead in three states with a total of 37 electoral votes. Trump is ahead in three states with a total of 37 electoral votes. Pennsylvania is tied. Did we mention that it is close? And did we mention that both campaigns are betting the farm on Pennsylvania? Also note that none of the totals add to 100%. Are the missing voters for Jill Stein or Cornel West? Or are they truly undecided now? Or maybe still for Kennedy?

We think there is a footnote here. In the Arizona Senate race, Rep. Ruben Gallego (D-AZ) is crushing Kari Lake. A Fox poll just out has Gallego ahead by 15 points, 56% to 41%. If this poll is true and the Emerson poll is also true, that means at least 9% of the voters support Gallego and Trump. So many people who support the fairly progressive Gallego over the Trumpy Lake also support Trump? We find that hard to swallow. Also, the abortion initiative isn't factored in here. Our suspicion is that the combination of Gallego swamping Lake and the abortion measure bringing marginal young women to the polls is going to affect the presidential race, but that may not show up when the questions are all asked separately.

Emerson also provided crosstabs by state, breaking out independents, voters under 30, women, and men. Here they are.

The big story here is the gender gap. Harris is scoring big with women and Trump is scoring big with men. The gender gap could be 20 points this time. With independents and young voters, Harris has a commanding lead, excepting independents in Nevada.

The second storyline is Trump is being crushed by unheard-of numbers among young voters. If Harris can get them to turn out in force, Trump will have a huge problem. Harris has to think carefully about how to get through to them. Making TikTok videos? Appearing at music festivals? Asking Taylor Swift what it would take to get her overt endorsement? Imagine the impact of a free concert on the National Mall where Swift welcomed Harris to the stage and endorsed her. That could draw a million people. How would Trump match that?

Emerson also asked people what their top issues were. Here is that breakdown:

By a mile, it is the economy, stupid. Democracy, immigration and housing are also important. We were surprised that abortion was not in the top three anywhere. At first we thought: "Did Emerson forget to include it in the list? So we checked the spreadsheet with the full results. It's in there, all right, but generally comes in fourth or fifth.

The message here is that Harris has to explain what she will do on the economy. Most likely by "economy," people mean "cost of living," rather than "jobs" or "stock market." One approach is to blame big companies for price gouging. Another is to be honest and say "no president can force prices down, but I will break up monopolies and oligopolies to create competition, and that will help bring prices down." Placing the blame on big greedy companies is probably a relatively easy sell. Once she gets specific, Trump will be forced to respond with his plan. Placing tariffs on all imports will raise prices, not lower them.

Yet another question is Project 2025. Is it starting to sink in? Emerson asked about it and whether: (1) it made people more likely or less likely to support Trump, (2) it made no difference or (3) they didn't know what it is. Here are those results:

We have our doubts about this one. We're the last people to insult the intelligence of the American people. OK, that's not true, but we are not the first people to insult the intelligence of the American people. Maybe fourth or fifth or so. According to these numbers, 87-90% of respondents know what Project 2025 is. We really doubt that. We suspect many people didn't want to come off as ignoramuses and said it makes no difference rather than admitting they don't have a clue what it is. As Harris continues to advertise about it, more people will genuinely learn what it is.

It is striking that among people who do know what it is, it is very unpopular. This suggests that: (1) Harris should keep talking about it and (2) she should emphasize the specific aspects that people don't like. This may require a number of focus groups to find out what people really think, rather than polling with multiple-choice questions. Harris should also emphasize that it was written by Trump's former advisers. For Trump, dealing with it could be tricky. Maybe he could talk about replacing the "deep state" (meaning the merit-based federal civil servants) with his own people, but it could be a tough sell for many people, even Republicans.

One last question was about not taxing tips. Here, 65% to 75% of respondents support the idea, even though most economists say it is a horrible idea and could lead to demands to not tax police agents, firefighters, nurses, or teachers, which could tear the entire tax system apart. A far better approach from an economic perspective would be to eliminate the tipped minimum wage and say that all workers must get the normal minimum wage and also advocate for raising that to $12 or $15 or some other amount. That would benefit all workers, not just those who get tips.

Morning Consult also polled The Big Seven. Their poll was done roughly a day earlier than Emerson's, so in theory, it could be slightly less up to date. Here are the Morning Consult results.

Except in Michigan, the results are better for Harris across the board. Does this mean Harris slipped a bit between the two polls? No. What it shows is a dirty little secret of polling. The model of the electorate matters. No poll of 1,000 people is going to have precisely the right number of young blonde left-handed white vegetarian college-educated lesbians. So all pollsters make corrections based on their model of what the electorate (the people who actually vote) will be. Different organizations have different models of the electorate. In this case, Emerson's model might have more people who are Trump supporters than Morning Consult's. Maybe more 45-65 year olds. Maybe more rural voters. Maybe more blue-collar workers. We don't know. It is also possible the difference is just due to chance since the margin of error on most polls is about 4 points. Pollsters all know that their model of the electorate is important and no doubt are constantly tinkering with it as new data become available. We're going to have to live with this uncertainty. This is why for the map and our averages page we use the most recent poll and average it with all other polls within a week of it. The hope is that effects due to incorrect models cancel out, but it is only a hope. (V)