olitics and the New Machine
The core argument
The essay argues that polling has become less reliable at the same time that it has become more powerful, and that this combination distorts democratic politics.
Polls:
increasingly fail to accurately measure public opinion
yet increasingly determine who gets attention, legitimacy, money, debate access, and media coverage
How Trump fits in
The piece opens with Donald Trump claiming he has no pollster and doesn’t tailor his message to polls. Lepore calls this disingenuous:
Trump may not have had a traditional campaign pollster
but his rise depended heavily on polls for visibility and validation
polls got him into debates, dictated stage placement, and fueled media coverage
So Trump is described as “a creature of the Sea of Polls,” not above it
Why modern polls are broken
The article explains in detail why polling has deteriorated:
- People don’t answer anymore
Response rates used to be 60–90%
Now they’re often in the single digits
Most Americans refuse poll calls, creating non-response bias
- Technology & law made it worse
Fewer landlines
Cell-phone autodialing is illegal
Internet polls are self-selected and skew younger and more liberal
Mixed-method polling still doesn’t work well
- Samples are tiny and fragile
National election polls often rely on ~1,000–2,000 people
Statistical “weighting” tries to fix bias, but the lower the response rate, the shakier the results
Why polls now matter more than ever
Despite being unreliable, polls are used to:
decide who qualifies for debates
determine media attention
shape fundraising and momentum
create “winners” and “losers” long before anyone votes
Fox News using polls to select debate participants is presented as a major example of polling replacing democratic processes.
Historical background
The essay gives a history of polling:
Early “straw polls” by newspapers
The rise of George Gallup in the 1930s
Polling claimed to represent “the will of the people” scientifically
But:
Early polls systematically excluded Black Americans, the poor, and the disenfranchised
Polling mirrored and amplified existing inequalities
What was presented as “public opinion” was often the opinion of a privileged subset
Deeper philosophical critique
Lepore raises a fundamental question:
What if measuring public opinion isn’t good for democracy at all?
Key ideas:
Polls treat public opinion as the sum of individual answers, ignoring how opinions are formed socially
Polls can create opinion rather than measure it
Constant polling shifts politics from deliberation and leadership to reacting to numbers
Bottom line
The piece isn’t just saying “polls are inaccurate.”
It’s saying:
Polls shape reality instead of describing it
They weaken representative democracy
They reward spectacle, momentum, and media attention over governance
And they increasingly substitute statistical artifacts for actual voting