Exhibit 06 · The Contested Wing
Three famous effects that failed to replicate. The museum keeps them on display — that's the point.
failed-replication curated 2012–2016 · honesty as content
Every other exhibit in this arcade shows you a bias and dares you to doubt yourself. This wing shows the field doubting itself. Between 2011 and 2016, psychology's "replication crisis" took down a string of textbook results: when independent labs re-ran them with larger samples and pre-registered methods, the effects shrank or vanished. These three were famous enough to make TED talks and airport bookshelves. They are displayed here, respectfully, as specimens of how smart people — including scientists — get fooled by the same machinery you just watched fail you.
"Unscramble sentences containing words like elderly, wrinkle, and Florida, and you'll walk out of the lab more slowly — your body quietly enacting the stereotype."
Bargh, Chen and Burrows' 1996 study became the flagship of social priming: tiny verbal nudges, big behavioral tides. In 2012, Doyen and colleagues re-ran the elderly-walking experiment with automated timing — and found the slowness appeared only when the experimenters knew who'd been primed. The mind being nudged, they argued, was the measurer's. Large-scale replication projects since have mostly failed to rescue the genre.
"Willpower is a muscle: resist the cookies in experiment one, and you'll quit the puzzles sooner in experiment two. Self-control runs on a depletable fuel."
Baumeister's ego depletion spawned over 600 studies and a bestseller (Willpower). In 2016, a pre-registered replication across 23 labs and 2,141 participants — coordinated by Hagger and colleagues, with the original team's input on protocol — found an effect statistically indistinguishable from zero. Later meta-analyses of the broader literature kept finding publication-bias-shaped holes where the fuel gauge should be.
"Stand like Wonder Woman for two minutes before the job interview: expansive postures raise testosterone, lower cortisol, and make you bolder."
Carney, Cuddy and Yap's power-posing result rode one of the most-watched TED talks ever to global fame. In 2015, Ranehill and colleagues ran the protocol in a sample five times larger: posture changed how powerful people said they felt, but hormones and risk-taking didn't move. (Cuddy's defenders note the "felt power" piece survives; the endocrine machinery that made it a sensation does not.)
Because the failure is the exhibit. Each of these effects was published in a top journal, cited hundreds of times, taught to students, and sold to the public — and each dissolved when the methods got stricter. The machinery that fooled you in the other wings (fluency, story-telling, wanting the world to be nudgeable) fools peer review too. The correction isn't despair; it's the badge system this museum runs on: every claim here carries its replication status, and "we were wrong" gets wall space like anything else.