The stat that a reporter fact-checked into a scandal

The stat that a reporter fact-checked into a scandal

A proptech firm pushed a headline finding: "73% of renters regret their city." It got picked up fast — then a data journalist pulled the underlying file and noticed the question only offered negative options. The follow-up wasn't coverage. It was a piece titled "How one survey engineered the outcome it wanted."

The cherry-pick didn't just fail; it became the story. The brand earned links — the worst kind, the ones pointing at a takedown.

We rebuilt their research standard around survivability: every question shown to reporters in full, neutral wording, results released even when boring or off-narrative. We stopped letting marketing pre-decide the headline before fieldwork ran.

The next study found something genuinely odd but defensible, with balanced question design and full crosstabs (the breakdown tables) attached. The same data journalist who'd written the takedown covered it straight this time — and noted, in print, "the methodology here is unusually transparent." That single line did more for credibility than any pitch could.

Lesson: if your data only works when no one checks the questions, a journalist will check the questions. Build studies to survive the fact-check, not just to win the headline.