Last Monday, the New York Times published a series of articles.public opinion pollFormer President Donald Trump is shown leading Vice President Kamla Harris through three Sun Belt battleground states: Georgia, Arizona, and North Carolina.
President Trump's lead in these states iseither trump or Harris It used to be held in these same states. Additionally, Trump's leads are well within the margin of error in Arizona (+5) and Georgia (+4), and North Carolina (+2) is just a tad closer.
But just days after the New York Times released the data, Bloomberg Morning Consult published Swing States.public opinion pollour own. According to their research, Mr. Harris They lead Arizona State by three points, North Carolina State by two points, and Georgia State is tied.
It's normal for two similar polls to show slightly different results, but the variation between two polls – 8 points in Arizona and 4 points in North Carolina and Georgia – is not trivial.
I cite these examples because they highlight a fundamental problem in predicting this election. There is no way to know with any degree of accuracy who is ahead and who is behind in a presidential election.
This is especially noticeable in the differences between swing state polls and national polls. In battleground states, both candidates are trading leads in polls suggesting neither candidate can see the light of day against the other.
However, national opinion polls such as those conducted by Reuters show that recentshowing Harris with a six-point lead (50 percent to 44 percent).
As I have previously written for this publication, polls showing Harris' strength are overstated, even if to her credit Biden has made up for lost ground and improved his personal metrics. may be.
In other words, the race is much closer than some polls predict. Given the “silent Trump supporters” and the general challenges pollsters face when interviewing Republican voters, the race is likely to be a statistical tie. At best, her lead with Harris is much narrower than the 5 to 6 points suggested by Reuters and NBC.
Also, at this point, Harris is leading the way. real clear politicsThe average (49 percent vs. 47 percent) would actually be two points below President Biden's 2020 performance. Given that Biden won by fewer than 50,000 votes in key states, Harris' 2-point lead nationally may not lead to an Electoral College victory.
In theory, polls in battleground states should provide a clearer picture than national polls, but in this election the poll averages are far from clear.
RealClearPolitics looks at voting averages in each battleground state and shows Trump in the lead.georgiaIt's the only one with more than 2 points. And in four of the seven swing states, the lead is less than one point in either direction.
Individual battleground state polls only increase the challenge of making accurate predictions. Consider the polls in Michigan, Pennsylvania, and Wisconsin, the “blue wall” states that are crucial to both candidates. Quinnipiac University recent researchHarris' lead is outside the margin of error in Pennsylvania (51% to 45%) and Michigan (50% to 45%), but within the margin of error in Wisconsin (48% to 47%). It shows that there is.
At the same time, a Hill/Emerson College poll showing Mr. Trump leading Ms. Harris paints a very different picture in those states.pennsylvaniaandwisconsinThey each trail by one point, with Harris trailing by two points.michigan.
As Ashley Fields pointed out, Quinnipiac's poll results show that “Harris leads Trump by less than 1 percentage point in Pennsylvania and Michigan” and by 3 points in Wisconsin. This differs significantly from the Hill/Decision Desk Headquarters aggregate average shown.
A second, equally important challenge is making accurate predictions. The evidence shows that Republicans are systematically underrepresented in public opinion polls in both 2016 and 2020, including the very real impact of the “silent Trump vote.” Opinion polls almost always underestimate Trump's support, resulting in him far exceeding pre-election expectations.
At this point in 2020, President Trump had a seven-point lead over then-candidate Joe Biden, according to RealClearPolitics. average. And the Reuter battlefield voteOn the eve of the election, Biden was shown holding a comfortable lead in Wisconsin and Michigan (+10 each) and Pennsylvania (+7).
Once the actual votes were counted, Biden won Michigan by just 3 points, Pennsylvania by 1 point, Wisconsin by less than 1 percentage point, and won the popular vote by 4 points, about half the expected result. We won by a margin.
Similarly, three polls were conducted in October 2016.releasedIt shows Clinton with an 11 point lead.michigan9 pointspennsylvaniaand 7 pointswisconsin. Trump won all three states and ultimately won the election.
Trump will be elected for the third time in 2024, but it remains to be seen how much influence the “silent Trump vote” will have on public opinion polls. Perhaps the pressure to hide voting intentions will lessen, and given the increasingly polarized politics in our country, the pressure will increase even more.
As it turns out, this race is too close to call. Users of public opinion polls, including both candidates, must be aware that this campaign has a long way to go and it is impossible to accurately determine the true difference between the two candidates at this time.
Douglas E. Schon is a political consultant and founder and partner of Shane Cooperman Research. The latest book is “The End of Democracy? Russia and China are rising, and America is retreating. ”
