The Houston airport baggage claim and Schiphol urinal fly stories are unverified

These stories persist because their shape lets them be cited without a source. A single round number, a one-step cause, and a conclusion that flatters intuition all travel well — especially into training decks, where citation isn’t expected.

Trace the Houston baggage-claim story and it stops at a 2012 New York Times opinion piece. Trace Schiphol’s urinal-fly “80 percent reduction” and it stops at an interviewee’s memory. Neither has an airport record or measured dataset behind it.

The Houston bag-claim story stops where it starts

Follow the Houston story backward and it ends at one opinion column — Alex Stone’s, in The New York Times, dated August 18, 2012. The anecdote has everything a keynote needs. Baggage-claim complaints are high.

The average wait is eight minutes; the walk from gate to carousel takes one minute. Executives rework the layout so passengers cover six times more ground — and complaints supposedly fall to near zero.

What the column leaves out is the entire factual skeleton. No airport is ever named. No terminal, no year, no complaint baseline, no measured after number.

Not a single airport document of any kind. Later business blogs and consulting decks preserve the arithmetic while dropping more context, yet each retelling still leads back to that single column, never to an airport’s own files. No airport-authority publication reporting complaint volumes for the intervention has turned up.

That’s the striking part. The story has repeated for years without once gaining a load-bearing fact.

It’s not that a covered-up dataset contradicts it — it’s that no dataset exists that can be checked.

Schiphol’s fly is real; its 80 percent number isn’t

At Schiphol the founding details are actually documented — kind of. The fly goes back to the early 1990s, when cleaning-department manager Jos van Bedaf raised the idea after remembering similar targets from his military days.

Aad Kieboom, the executive who oversaw terminal extensions and renovations, helped win management approval. The fly was an actual object attached to the men’s urinals in the early 1990s, not a metaphor invented for training decks.

Then comes the credibility gap. The famous 80 percent reduction in spillage has no trial record behind it. No measured dataset, no original measurement window.

Kieboom himself called the figure “very empirical” in an interview with Works That Work — a description that’s honest and also exactly the problem. Where the 80 percent first appeared in print remains unresolved.

In other words, the fix itself has a firmer paper trail than the number attached to it.

Two failure modes, one repeatable pattern

The airport cases aren’t isolated. Seven claims outside aviation show the identical failure mode, and they split into two groups worth keeping separate.

Some have no study or measurement behind them at all — Houston’s complaint drop, the urinal fly, the “stories are 22 times more memorable” statistic attributed to Jerome Bruner, the “angry customers tell 10 people” line. The rest sit on real studies whose findings got stretched as they passed from hand to hand.

That second group is the subtle one. The 1964 Kitty Genovese murder was real, but the famous reporting misstated how many witnesses there were and whether they could see the attack — yet it still launches most bystander effect lectures. Rosenhan’s 1973 study existed, but later investigators reported missing evidence and contradictions in its own records.

Walter Mischel’s marshmallow experiments studied strategies for resisting temptation, never a general forecast of adult success. Oxytocin’s “trust chemical” reputation did not survive later replication attempts, including a critical review. And the 55 percent body language, 38 percent tone, 7 percent words formula gets used far outside the narrow conditions of the research usually cited for it.

The table below lays out the evidentiary trail for each claim. Note the distinction: absent documentation and downstream distortion are separate problems.

Unverified isn’t the same as false — a null result is simply harder to publish than a satisfying story.

The Asia-Pacific comparison turns up quieter than expected. Changi, Incheon, Haneda and Sydney all came up empty — none produced a verified behavioral design parable that clears the same bar.

Changi’s official history confirms only a relocation: the airport left Paya Lebar for its current site on June 30, 1981. That is a genuine development milestone, not an unsourced design claim.

Why the ‘psychology beats engineering’ shape travels

The features are almost comically consistent. A single round number — 80 percent, 22 times — gives every retelling a hook that needs no decimal. A one-step causal chain: walk farther, complaints fall; etch a fly, spillage drops.

No interaction effects, no seasonality. And the conclusion flatters the listener’s intuition — people are irrational, a cheap psychological fix beats expensive engineering.

That’s satisfying enough that nobody presses for the dataset.

Then the corporate-training deck does the rest. Once a story enters a slide, it acquires authority without acquiring a source. Nobody cites in a keynote, and nobody asks in one either.

A null result — “the change made no measurable difference” — has no hook, so the satisfying version survives. University press offices may amplify this flattening, but that’s an open mechanism, not a demonstrated cause.

What this means for the next airport claim you see

The practical takeaway is defensive. Before you plan a connection around a walk-further tip, or believe an airline that borrows behavioral-science language, run the claim through three checks.

• Can you name the airport, terminal, and year? If not, the story has not been pinned down.
• Is there a measured baseline — a before number and an after number from a source that isn’t a keynote deck?
• Does anyone on the record, preferably an airport authority, claim the result?

Two of the three should pass before you act. That checklist is the practical version of this desk’s sourcing standard: publish only traceable claims, and when a good story fails the test, say plainly that we could not verify it. That’s an editorial position, not a research finding — but it’s the one thing in this piece you can actually use at the gate.

Key terms

Behavioral design

Behavioral design is the practice of shaping a physical space so that people behave differently without being told to. It works through small cues — a sightline, a sign, a target etched into porcelain — and it turns up in airports, hospitals and public buildings. Both parables in this article are behavioral-design stories, which is part of why they travel: the fix sounds cheap, the result sounds large, and nobody asks for the trial data.

Null result

A null result is a study outcome that finds no measurable difference between the conditions being compared. Null results are rarely published or cited, so the public record fills up with the studies that found something. That gap is why this article keeps “unverified” and “false” apart: a missing measurement is not the same as a disproving one.

Replication

Replication is the practice of running a study again, usually with new participants, to see whether the original result holds. It is the standard test of whether a finding is real or a fluke of one sample. It is also the line this article’s comparison table draws: some claims never had a study to replicate, while others had one whose meaning was stretched as it was retold.

Bystander effect

The bystander effect is the finding that people are less likely to step in during an emergency when others are present. It is usually traced to the 1964 Kitty Genovese murder, the case that made the effect a standard item in psychology teaching. That case is one of the few in this article where the underlying event definitely happened — the distortion sits in the retelling, not in the murder itself.

Questions? Answers.