Chabris caught Gladwell's font claim on a 40-student study that never replicated

In October 2013, Malcolm Gladwell's new book David and Goliath opened with a striking premise: facing overwhelming odds produces greatness and beauty. One of its examples, drawn from a 2007 study, claimed that people who read reasoning problems in a hard-to-read typeface scored 29 percent higher — a counterintuitive result with immediate practical appeal.

Christopher Chabris, a psychology professor reviewing the book for The Wall Street Journal, did not argue with the story; he opened the study it rested on. The paper involved 40 people, 20 per typeface, all Princeton University students — facts Gladwell did not mention. When other researchers had tried the same test with nearly 300 students at a Canadian public university, they could not replicate the effect. The flagship example of the book's thesis collapsed on inspection.

The deconstruction generalized because the pattern was familiar. Chabris wrote that this flaw permeates Gladwell's writings: he excels at telling just-so stories and cherry-picking science to back them. Four years earlier, Steven Pinker's New York Times review had coined the Igon Value Problem for a writer whose education on a topic consists of interviewing an expert. Even admirers conceded the facts came second; Gladwell himself told interviewers he was a storyteller who looks to research to augment storytelling, and that readers who find his books oversimplified are not the audience.

The case is a transferable check. Surprising statistics should be traced to their primary paper, where sample size, sample population and replication history sit in plain sight. Forty people cannot carry a law of human behavior, and an effect that fails to reproduce with nearly 300 is not a law at all.

Do not argue with the story — open the cited study; sample size, who was sampled and whether anyone replicated are the tells that expose cherry-picked science

A primary paper answers the two questions a compelling retelling dodges: how many people were studied and who they were.

A failed replication is decisive evidence that an effect is not robust, whatever the original paper claimed.

Because the flaw sat in the book's flagship example and matched a pattern documented years earlier, one anecdote exposed a method rather than a one-off error.

The check cost nothing but reading: no new experiment was needed to show that 20 Princeton students per font cannot support a grand claim.

Before repeating a surprising research claim, trace it to the primary paper and check the sample, population and replications; one exposed cherry-pick stains every other claim the storyteller makes.

Chabris expanded the review into a blog post, Why Malcolm Gladwell Matters (And Why That's Unfortunate). The Knight Science Journalism Tracker at MIT amplified the case, and the argument reached Hacker News with 370 points and 222 comments. Gladwell's own quotes — a storyteller who uses research to augment storytelling, and books oversimplified for some readers — were read by critics as confirmation. The durable result was a demonstrated technique: when a bestseller cites a surprising study, checking its sample and replications takes minutes and can collapse the story.