Investigative agencies look for explanations for erroneous predictions

Something similar is happening again: Although pollsters changed their methods after getting Donald Trump wrong in 2016 and 2020, they are wrong again now. Is the polling industry over?

Patrick Muller/ch media

No reputable pollster predicts that Donald Trump will win more votes nationwide than Kamala Harris. So it came as a surprise when the New York Times’ “predictive needle” began reporting on the sensational events of election night.

Not all votes were counted Thursday, but as things stand, Trump leads the country by 3.5 points in the so-called “popular vote.” In the last poll before the election, Harris was leading by 1.2 points. Hence the margin of error: an impressive 4.7 percentage points. It is considered the closest game in American history.

The differences in swing states are more important. Their electoral votes are crucial to the election. In recent pre-election polls, Trump led in just three of seven battleground states. However, in reality, he won all seven states, although two were not counted (but the lead was significant).

According to the Washington Post, survey averages in seven battleground states in the week before the election (October 28, 2024).

Polls do not predict that Kamala Harris will do worse than Joe Biden did four years ago in every (!) constituency in the United States. When the corresponding card popped up, the CNN anchor was stunned:

“Pollsters”—as America’s survey experts call themselves, a group imbued with professional pride—are now taking a hard look at themselves. Vanderbilt University political scientist Josh Clinton acknowledged that the Journal's prediction was incorrect. “This is due to the imprecision of the instruments we use,” he defended himself. Both Democrats and Republicans have been fairly accurate in their assessments of the 2022 midterm elections. But in all three presidential elections in which Trump has run (2016, 2020 and 2024), their results have been useless. Although after the first two failures, the method was improved.

Scientists explain this with the Trump phenomenon. Pollsters are clearly unable to capture Trump voters in a representative way. There were many among her who refused to be investigated. When an agency calls on behalf of the “mainstream media,” they hang up. Don Levy is president of the Siena Institute, which conducts investigations for The New York Times. He said someone picked up the phone, shouted “Trump!” into the receiver, then hung up. These “participants” do not fit into any survey grid. The number of dissenters is significantly higher than average among black, Latino and young voters. And Trump has made strong gains among these groups.

Can investigative agencies survive after a hat-trick of failed attempts? No. Every survey is guaranteed attention, clicks, and ratings. That's why your business will continue to thrive. But the warning signs may become clearer: The margin of error and the fact that this is a snapshot, not a forecast, will be communicated more aggressively. (aargauerzeitung.ch)

After Donald Trump won the election, US President Joe Biden promised a peaceful transfer of power on January 20. “I will fulfill my duties as president. I will fulfill my oath, and I will abide by the Constitution. On January 20, the United States will have a peaceful transfer of power.” He wants to make the most of the last few days of his term.