Why Were Liberal Pollsters So Far Off… Again?

For all the talk of how tough it is to accurately poll Trump voters, there’s been far less discussion about whether liberal pollsters even want to.

Instead of measuring reality, most of these polling “experts” are reporting on what they want you to think reality is. That’s because Big Polling is just as much part of the Left’s disinformation complex as MSNBC, the Atlantic, or any other front group for the Democratic Party.

The Capital Research Center’s Ken Braun has done an admirable job tracking just how far off these “agents of influence” were in Trump’s latest tussle with the lefty media’s latest Democrat darling, Kamala Harris. (Spoiler alert: CNN, NBC, and the New York Times were off by as much as 9.2 percentage points in their predictions for Kamala.)

Let’s go further and stack the right-leaning—or reality-leaning—pollsters side-by-side with their competitors on the Left.

For each battleground state, we took both the final polling result (in late October or early November) and the average of all polls by a given pollster for the period after Joe Biden dropped out of the race in late July. (Most polls are from September at the latest, however.)

Pennsylvania

On the Left:

• New York Times/Siena
  1. Average: Harris +3.7
  2. Final: Tie
• Quinnipiac
  1. Average: Harris +2 (high of +5 in mid-September)
  2. Final: Trump +2
• Marist
  1. Average: Harris +1
  2. Final: Harris +2
• Selzer & Co.
  1. Average: Harris +1.75 (high of +5 in mid-September)
  2. Final: Harris +2
• CNN: Tie

On the Right:

• Atlas Intel
  1. Average: Trump +2.2
  2. Final: Trump +1
• Rasmussen Reports
  1. Average: Trump +1.5
  2. Final: Trump +2
• Trafalgar Group
  1. Average: Trump +2.2
  2. Final: Trump +1

Least accurate: Marist & Selzer &. Co.—tied at 3.8 percentage points off from the final election result.

Most accurate: Atlas Intel—0.8 percentage points off.

(ANALYSIS: Leftist Groups Spent $6 Million Fighting Tough-on-Crime Measures—and Lost)

Wisconsin

Final result: Trump +0.9

On the Left:

• New York Times/Siena
  1. Average: Harris +1.5
  2. Final: Harris +2
• Marist
  1. Average: Harris +1.5
  2. Final: Harris +2
• CNN: Harris +6
• Quinnipiac
  1. Average: Trump +0.3
  2. Final: Tie
• Selzer & Co.
  1. Average: Harris +3.25 (high of +8 in late August)
  2. Final: Tie

On the Right:

• Atlas Intel
  1. Average: Trump +0.5
  2. Final: Trump +1
• Rasmussen Reports
  1. Average: Trump +1
  2. Final: Trump +3
• Trafalgar Group
  1. Average: Trump +0.25
  2. Final: Harris +1

Least accurate pollster: CNN—5.1 percentage points off.

Most accurate: Atlas Intel—0.1 percentage points off.

Michigan

Final Result: Trump +1.4

On the Left:

• New York Times/Siena
  1. Average: Harris + 1.75
  2. Final: Tie
• Marist
  1. Average: Harris +4
  2. Final: Harris +3
• Quinnipiac
  1. Average: Harris +1.7
  2. Final: Harris +4
• Selzer & Co.
  1. Average: Harris +5 (high of +11 after Biden dropped out in July)
  2. Final: Harris +3

On the Right:

• Atlas Intel
  1. Average: Trump +2
  2. Final: Trump +2
• Rasmussen Reports
  1. Average: Harris +0.5
  2. Final: Harris +1
• Trafalgar Group
  1. Average: Trump +1.25
  2. Final: Trump +1

Least accurate pollster: Quinnipiac—2.6 percentage points off.

Most accurate: Atlas Intel—0.6 percentage points off.

North Carolina

Final result: Trump +3.2

On the Left:

• New York Times/Siena
  1. Average: Harris +0.7
  2. Final: Harris +2
• Marist
  1. Average: Trump +1
  2. Final: Trump +2
• CNN
  1. Average: Harris +0.5
  2. Final: Harris +1
• Quinnipiac
  1. Average: Harris +1.7
  2. Final: Harris +3
• Selzer & Co.
  1. Average: Harris +0.25
  2. Final: Trump +1

On the Right:

• Atlas Intel
  1. Average: Trump +0.7
  2. Final: Trump +2
• Rasmussen Reports
  1. Average: Trump +3.7
  2. Final: Trump +3
• Trafalgar Group
  1. Average: Trump +2.75
  2. Final: Trump +3

Least accurate pollster: New York Times/Siena—5.2 percentage points off.

Most accurate: Rasmussen Reports—0.2 percentage points off.

(COMMENTARY: A Fair and Balanced Look at How Trump’s Victory Destroyed Corporate Media Forever)

Georgia

Final result: Trump +2.2

On the Left:

• New York Times/Siena
  1. Average: Trump +2.5
  2. Final: Harris +1
• Marist
  1. Average: Trump +0.5
  2. Final: Tie
• CNN: Trump +1
• Quinnipiac
  1. Average: Trump +4.7
  2. Final: Trump +6
• Selzer & Co.
  1. Average: Tied
  2. Final Trump +2

On the Right:

• Atlas Intel
  1. Average: Trump +2
  2. Final: Trump +2
• Rasmussen Reports
  1. Average: Trump +4
  2. Final: Trump +5
• Trafalgar Group
  1. Average: Trump +1.3
  2. Final: Trump +2

Least accurate pollster: Quinnipiac—3.8 percentage points off.

Most accurate: Atlas Intel—0.2 percentage points off

Nevada

Final result: Trump +3.1

On the Left:

• New York Times/Siena
  1. Average: Trump +2
  2. Final: Harris +3
• CNN:
  1. Average: Tied
  2. Final: Trump +1
• Selzer & Co.:
  1. Average: Harris +3.5 (high of Harris +7 in mid-September)
  2. Final: Harris +1

On the Right:

• Atlas Intel:
  1. Average: Trump +1.83
  2. Final: Trump +3
• Rasmussen Reports:
  1. Average: Trump +1.75
  2. Final: Trump +2
• Trafalgar Group:
  1. Average: Trump +0.3
  2. Final: Tie

Least accurate pollster: New York Times/Siena—6.1 percentage points off.

Most accurate: Atlas Intel—0.1 percentage points off.

Arizona

Final result: Trump +5.5

On the Left:

• New York Times/Siena:
  1. Average: Trump +2.8
  2. Final: Trump +4
• Marist:
  1. Average: Trump +1
  2. Final: Trump +1
• CNN:
  1. Average: Trump +2
  2. Final: Harris +1
• Selzer & Co.:
  1. Average: Harris +1.25 (high of Harris +3 in mid-September)
  2. Final: Tie

On the Right:

• Atlas Intel:
  1. Average: Trump +3.3
  2. Final: Trump +5
• Rasmussen Reports:
  1. Average: Trump +2
  2. Final: Trump +2
• Trafalgar Group:
  1. Average: Trump +1.6
  2. Final: Trump +2

Least accurate pollster: CNN—6.5 percentage points off.

Most accurate: Atlas Intel—0.5 percentage points off.

Missed it By That Much

1. Atlas Intel—0.5 pts. off
2. Tie: Rasmussen Reports & Marist—each 1 pt. off
3. Trafalgar Group—1.6 pts. Off
4. New York Times/Siena—2.8 pts. Off
5. Selzer & Co.—4.25 pts. Off
6. CNN—6.5 pts. off

2024’s most accurate polling firms have one thing in common: They aimed to accurately report the views of America’s electorate.

Their competitors across the aisle, in contrast, reported figures designed to sway the election toward Harris. They shouldn’t be called pollsters at all, since they’re really partisan operatives who suppressed the Republican vote to push a pro-Harris narrative. The New York Times actually brags about weighting against “low-quality” (read: Republican-friendly) polls—that is, oversampling Democrats—while accusing their rivals of “building a narrative of unstoppable momentum for Mr. Trump.”

Why else dump a last-minute Harris +3 poll in Iowa, a ruby-red state Trump won by 8 points in 2020 and a whopping 13 points in 2024, four days before Election Day (and with no crosstabs)? The “expert” who released that suppression stat, Ann Selzer (Selzer & Co.), announced her retirement from polling two weeks later.

As someone quipped on X/Twitter, it was as if she polled the line at Starbucks. Hey, maybe they’re hiring.

(READ MORE: Trump Won Wisconsin Thanks to Conservatives Adapting to Mail-In and Early Voting)