The Polls Did Pretty Well!

Predictably, I fouynd this at Real Clear Politics this morning: Were the polls right? It’s complicated.

One of the most frequent questions I get about this election is some version of, “why were the polls so wrong?”

. . . RCP’s final poll average gave Harris a lead of .1 (that is, one-tenth of a point) nationally. Trump is now winning by 1.7 points, yielding a miss of 1.8 for the national average — pretty close, by any reckoning. And it could get even closer when the final tranche of ballots is tallied.

The biggest miss in the swing states was Arizona, where the poll average missed by 2.7. In four of the seven swing states, the difference between the average poll margin and the vote count was 1.7 points — even better than the national results. 

Turn now to the seven battleground Senate contests: In three of them, the miss was less than 1 point. It averages 1.6 points. The biggest miss was in Nevada, where the polls had Jacky Rosen ahead of Sam Brown by 4.9 points in a race she won by 1.6.

So the polls were pretty darn close, huh? Less than two points off! The problem was that most of them, including the RCP average, had Harris winning. On top of that, the consensus across all media throkughout the election season, even before Biden withdrew, was it was going to be "razor thin", "extremely close", and so forth.

This site gives the electoral college totals for all presidential elections since 2000:

• 2000: George W. Bush wins the Electoral College 271-266 (with one faithless elector not voting for Al Gore)
• 2004: George W. Bush wins the Electoral College 286-251 (with one faithless elector not voting for John Kerry)
• 2008: Barack Obama wins the Electoral College 365-173
• 2012: Obama wins the Electoral College 332-206
• 2016: Donald Trump wins the Electoral College 304-227 (Trump lost 2 electoral votes to faithless electors while Hillary Clinton lost 5)
• 2020: Joe Biden wins the Electoral College 306-232

It concludes,

So the last 6 elections give us some parameters in assessing whether and how 2024 ends up being a “close” election.

As noted above, it would be very hard for the election to be closer than 2000, given the hair’s width of separation between Bush and Gore that decided the pivotal state of Florida.

Something along the lines of 2016 and 2020, where the decisive margin was 5 figures’ worth of raw vote margin and less than a percentage point’s worth of difference in the key states, would also qualify.

Even something like 2004, where the difference between victory and defeat was a bit over 100,000 votes or 2 points in margin in a single key state, probably would qualify as “close,” although it actually would not rank as one of the top 3 closest elections of the 2000s.

The 2024 Electoral College total was 312-226, with only Obama's two victories surpassing Trump's margin. This was nothing like a close election, especially with the recent example of 2020 for comparison. This piece neverthneless insists the polls did pretty well:

The final Yahoo News/YouGov survey had Trump and Harris at 47% of the vote apiece among likely voters. That left 6% of likely voters who were either backing third-party candidates or who were still undecided — giving ample room for either candidate to take a small lead.

But the popular vote total had Trump at 50% to Harris at 48.3%, which means that that poll had Trump's total at 3% too low. All it did was reassure everyone that the election would be close, when it wasn't. The only definite thing this does is convince me that reading Real Clear Politics is a waste of time.