Oh gosh—so important, so powerful, so impactful—and so badly used.

Polling should be one tool among many. Instead, it’s often treated as fact. Media outlets use it to cement whatever narrative they’re building. It’s as if the poll decides the issue. Source citations have nearly vanished; all they need to chant is, “the polls say…” or “according to recent polls…”—and suddenly, speculation sounds like certainty.

Yes, polls can be useful. But only if we vet the source, examine the sampling, study the questions, and consider who’s responding. Even then, polls are always time-stamped snapshots, not conclusive evidence.

Three of the most visible political polling failures illustrate this:

“Dewey Defeats Truman” (1948) – A premature consensus that ignored late opinion shifts. Hillary Clinton defeats Donald Trump (2016) – An overconfident forecast based on flawed state-level data. Brexit Vote: Britain Will Remain (2016) – A misread of voter turnout and demographic silence.

All three were treated as foregone conclusions by the media. Each became a case study in how “moving truths” can slip just before the finish line.