Ninety-two percent said they would refuse. One percent did.
A century of research says people are poor predictors of their own behaviour.
Organisations are no different.
In 1934, LaPiere drove 251 establishments across America. The couple was refused service once.
In 1934, a sociologist named Richard LaPiere drove across the United States with a young Chinese couple. This was a period of open, widespread anti-Chinese sentiment, so he expected trouble. Over two years they visited 251 hotels, restaurants and roadside establishments. They were refused service exactly once.
Six months later, LaPiere wrote to the same establishments and asked a single question: would you accept members of the Chinese race as guests? Of the 128 who replied, around 92% said no.
Ninety-two percent said they would refuse. One percent actually did.
That gap, between what people said they would do and what they actually did, is one of the most durable findings in the history of social science. It wasn't a fluke of one eccentric study. In 1969, a review of 42 separate experiments found that stated attitudes rarely explained more than about 10% of the variation in real behaviour. The thing people tell you, it turns out, is a remarkably poor predictor of the thing they do.
The decades since have explained why. Festinger showed that when our actions contradict our stated values, we don't change the action. We invent a comfortable story that lets us keep both. Ajzen showed that intention is only a small input to behaviour, easily overridden by environment and by what we think others expect. And Kahneman gave us the cleanest model of the machinery: a fast, automatic system that drives most of what we actually do, and a slow, deliberate system that produces what we say. The slow system doesn't have full access to what the fast one is up to. So when you ask someone what they'll do, they're not lying. They're sincerely guessing, and they're usually wrong.
I didn't come to this through psychology. I came to it through work.
A discipline built on the gap
I started as a UX designer. Then UX researcher, then research lead, then product manager, and now Head of R&D and Innovation. On paper that looks like a series of different jobs. Looking back, it was one job at five sizes.
Because user experience research, at its core, is a discipline built entirely around the say-do gap. It exists because you cannot trust what users tell you in isolation. Ask someone if a checkout flow is easy and they'll say "yes, fine." Then you watch them click the wrong button three times and stall for two minutes, and you learn what's actually true. The whole craft is the discipline of measuring both, what people say and what they do, and paying attention precisely to the distance between them.
I have a clear memory of when this stopped being a technique and became a way of seeing.
We were redesigning the navigation for a large product suite. We ran a card sort: we handed people labelled cards and asked them to organise the structure the way it made sense to them. They did, confidently. Then we ran the reverse test. We showed them the structure and asked them to find things. And they couldn't. The very people who had just told us where everything should live now failed to locate the same items when they had to behave rather than describe.
Said one thing. Did another. And the gap between the two was exactly where the product was failing its users. That's a small story about a navigation menu. But it rewired how I look at everything.
The same gap, at organisational scale
Here is the leap that took me years to actually see, even though I'd been standing on it the whole time.
The say-do gap doesn't only live in users. It lives in organisations, and it is exactly as wide, as invisible, and as consequential.
A strategy is a stated attitude. It's the organisation telling you, sincerely, what it values and what it intends. The organisation's actual behaviour, what gets resourced, what gets shipped, what gets rewarded, what quietly gets avoided, is the action. And the distance between those two is the value-action gap operating at the scale of a whole company.
You have seen this. The organisation that declares innovation a priority, then measures and pays its people entirely on delivery, so no rational person spends a minute innovating. The company whose mission is inclusion while its own product excludes the very people it claims to serve. The strategy that everyone nods along to in the town hall and nobody can act on the next morning, because it was carefully written around the one problem the organisation has tacitly agreed not to name.
None of this is dishonesty. That's the part most people miss. Just as the survey respondent in 1934 wasn't consciously lying, the leadership team isn't either. They're operating on the stated version, the System 2 story, without full access to what the organisation is actually doing underneath. The gap is a blind spot, not a lie. Which is precisely why it persists: you cannot close a gap you cannot see, and the people best positioned to see it, the ones on the ground watching the behaviour rather than reading the strategy, are usually the furthest from the room where it could be fixed.
What I actually do
So here's the through-line I spent a career not noticing. I started by measuring the gap between what users say and what they do. I just kept applying the same instrument to bigger and bigger units: the interface, the user, the team, the product, the whole organisation. The unit of analysis grew at every step. The move never changed.
I find the gap between what an organisation believes it's doing and what it's actually doing, and I make it visible enough that the people inside can finally act on it.
Not by telling them their blind spot. You can't, by definition. But by building the conditions, the questions, the structures, where they see it for themselves. The same way a usability test doesn't argue with a user; it just lets them discover, in their own hands, that the thing they called easy isn't.
The science has known about this gap for ninety years. We measure it obsessively in users and barely look for it in the organisations those users work inside.
That's the gap I'm interested in now. Where's the widest one in yours, and who in your organisation can already see it, while the people who could close it cannot?
Sources
Richard T. LaPiere, "Attitudes vs. Actions," Social Forces 13 (1934): 230-237. Across roughly 251 establishments visited over two years, the couple was refused service only once.
Of the 128 establishments that responded to LaPiere's follow-up questionnaire, approximately 92% of restaurants and 91% of hotels stated they would not accept Chinese guests. Note the two different denominators: 251 visited, 128 responded to the survey.
Allan W. Wicker, "Attitudes versus Actions," Journal of Social Issues 25 (1969): 41-78. A review of 42 studies finding attitudes rarely accounted for more than ~10% of variance in behaviour.
Leon Festinger, A Theory of Cognitive Dissonance (1957).
Icek Ajzen, "The Theory of Planned Behavior," Organizational Behavior and Human Decision Processes 50 (1991): 179-211.
Daniel Kahneman, Thinking, Fast and Slow (2011).
