Nobody has a crystal ball and can say with 100% certainty how the election next week will turn out. We're all basically guessing based on polling, betting markets, and wind patterns. The latest polls to come out have been pretty darn weird — so weird, in fact, that Nate Silver is accusing pollsters of putting their thumbs on the scale.
Last month, AtlasIntel, which was the most accurate pollster of the 2020 election released the results of their survey, and, as I said at the time, their results were weird. The poll showed Donald Trump leading Kamala Harris nationally by three points, with Trump at 50.7% and Harris at 47.6% with a full field. In a head-to-head match-up, Trump leads 50.8% to Kamala’s 48.1%. Their state results felt a little off those, as a three-point lead nationally ought to have translated into an Electoral College landslide, but instead, theirs pointed to a literal 269-269.
Recommended: Nate Silver Claims Pollsters Are 'Putting Their Finger on the Scale' to Make It Appear the Race Is Close
AtlasIntel has some new numbers out, and they appear to make more sense.
Nationally, AtlasIntel shows Trump leading Harris by 49.6% to 48.2% in a head-to-head matchup. When factoring in other third-party candidates, he still holds the edge, leading 49.1% to 47.2%. That may be a smaller margin than their previous poll, but a national lead is still pretty huge for Trump, who has yet to win the national popular vote.
HEAD-TO-HEAD SCENARIO
— AtlasIntel (@atlas_intel) October 31, 2024
In a head-to-head race between Harris and Trump, the difference in favor of the Republican in voters' preference at the national level is reduced to 1.4%. The absence of other candidates benefits Harris, who grows by exactly 1 percentage point. pic.twitter.com/oKpiuSJlyI
Here's how the battleground state numbers look. In North Carolina, Trump is ahead with 50.7% of the vote compared to Harris's 47%, giving him a 3.7-point advantage. Georgia sees a similar pattern, with Trump narrowly leading Harris by 2.3 points, 50.1% to her 47.8%. Arizona reflects an even wider gap, where Trump secures 51% of the vote, putting him 4.2 points above Harris, who holds 46.8%. The margin is comparable in Nevada, with Trump at 50.6% to Harris’s 47%, a 3.6-point lead.
Recommended: EXCLUSIVE: Pollster Explains Why This Election Feels Like Reagan vs. Carter
However, in Wisconsin, Harris edges slightly ahead, capturing 49% to Trump’s 48.8%, a slim lead of just 0.2 points. Despite this, Trump regains a slight advantage in Michigan, leading by 0.6 points with 49.3% against Harris’s 48.7%. In Pennsylvania, Trump holds a narrow 1.5-point lead, standing at 49.4% compared to Harris’s 47.9%.
📊ATLAS POLLS - SWING STATES
— AtlasIntel (@atlas_intel) October 31, 2024
Harris leads in only one of the 7 decisive swing states for the Electoral College outcome: Wisconsin. Trump's advantage over his opponent is most significant in states such as AZ, NV and NC. The race is tight in the Rust Belt swing states. pic.twitter.com/H0AqUuraoL
At face value, that gives Trump an Electoral College victory of 302-236. However, AtlasIntel considers Wisconsin and Michigan as toss-ups. So they say this has Trump winning 287-226, with the results of those states essentially not impacting the outcome.
Note: In the above projection, states that were not polled use publicly available ratings from the University of Virginia Center for Politics.
— AtlasIntel (@atlas_intel) October 31, 2024
AtlasIntel is expected to release one or two more polls before Election Day.
These results are more sensible than the previous poll, though they have some curiosities. Even if Trump wins Nevada, I expect him to win by one or two points, not more than three. I also suspect he’d likely take Wisconsin if he wins both Pennsylvania and Michigan. More importantly, these state-by-state results better reflect a national environment where Trump is up about two points than the previous poll.
Hopefully, this one is right.
