Validating in the open.
what-world-way is a working framework. As a non-profit, rather than rush or claim an accuracy we haven’t earned, we will validate it gradually and in the open as people take it — publishing the plan first, so the predictions are on the record before the data arrives.
Where we actually stand.
The three pillars draw from established research traditions — the Big Five for personality, the Gravesian developmental levels for the worlds, the motivational-type lineage for the archetypes — and from years of our own discussion and thought. The specific instruments we built on top of them are early. They haven’t been through independent validation, and we don’t say they have.
So we launch as a stated pre-validation prototype and work toward validation in the open. Results are published stage by stage, with honest caveats, and updated as each part fills in. The weights and cut-offs in the test today are expert first guesses, labelled as such — not settled values.
What we can prove, and when.
We don’t fund one large study up front. We validate what we can, for free, as traffic accumulates — and pre-register each stage before we run it.
- Stage 0Now — as people take the free test
How: The accumulating Glimpse takes, plus the funnel that gives What, World, and Way in one sitting.
What it shows: Internal structure, reliability, the real rate of the Open pattern, and a coarse first check on whether the three pillars pull apart.
Already running, growing with traffic.
- Stage 1At the deeper Insight tier
How: Insight What takes, as they accumulate.
What it shows: The same checks, at Insight resolution.
Underway, as Insight takes come in.
- Stage 2The decisive test
How: A recruited sub-sample doing Insight What alongside an established Big Five measure and a values measure.
What it shows: The definitive independence test and the prototype structure analysis — measured against instruments built by other people, not our own.
Rolling, as the sub-sample fills.
One caveat carries a lot of weight, so we state it plainly. Stages 0 and 1 compare the What test against our own World and Way tests — same wording, same sitting — so they can catch gross overlap but cannot prove independence. Stage 2, measured against a Big Five and a values instrument built by other people, is what earns the claim that the three pillars are separate. We never present the early stages as that proof.
What we committed to, before the data.
These are the falsifiable predictions — written down in advance, so they can come back wrong. Each is pre-registered before its stage is run.
- Independence
Once you subtract what personality traits and values already explain, each part of the What test should still carry its own motivational signal. The part most at risk — stability and change — may turn out to belong with World or Way instead.
- Structure
When we let the data group itself, do roughly eleven distinct archetypes appear, or do some merge? We expect some to collapse together — the honest prediction is six to eight clear groups, with the eleven sitting as recognisable positions within them.
- The Open pattern
The Open route should pick out people with a real fluidity across contexts, told apart from people who weren't engaging with the questions.
- Reliability
The items meant to measure one thing should agree with each other.
- Added value
The real prize: does an archetype predict something — what work energises or drains you, what you find meaningful — that traits and values do not? If it adds nothing of its own, it has not earned its place.
Three ways the framework can be wrong.
A framework that can’t lose isn’t doing science. So we named, in advance, the three things the data could do to an archetype — and committed to following them rather than defending the original design.
- Relocate
The distinction is real but filed in the wrong place — it belongs with World or Way rather than What. The role stays; the part that leaked moves.
The stability-and-change region is the live candidate.
- Merge
Two recognisable roles turn out to be one underlying thing — people can't reliably tell them apart, and nothing measurable separates them. One name goes; the role survives under a single name.
Stag and Wolf are the pair we watch first.
- Drop
An archetype fails on its own terms — strangers can't recognise it any better than a flattering generic description, and it adds nothing beyond a blend of the others. It is removed entirely.
The Beaver is named in advance as the most likely to go this way.
The keep-or-cut decision turns on recognition. An archetype stays if people reliably see themselves and others in it — and that is itself a test, not a free pass: we measure it by whether people place themselves consistently, whether someone who knows them agrees, and whether strangers can sort a person to the right archetype better than to a flattering generic description. A role that fails that test loses its reason to stay.
What this is, and isn't.
The claim we lead with is modest and, we think, the interesting one: people can recognise meaningful distinctions that existing models don’t explicitly name. Not that we have discovered new dimensions of personality — recognisable roles, tilted by temperament, sitting on semi-independent levels.
A useful comparison from outside the framework: an archetype is to motivation roughly what “social introvert” is to personality — a recognisable, useful name for a real pattern, not a claim to a brand-new dimension. “Social introvert” is both reducible to known traits and genuinely useful, and those two facts aren’t in tension. That is the bar we hold ourselves to: keep the role, and never oversell it as an independent dimension it isn’t.
Pinned at every step, but never frozen.
Clare Graves, whose developmental work sits behind the six worlds, called his life’s study a never-ending quest. This is the same. Validation here is rolling and perpetual, not a one-time gate — archetypes can merge, relocate, dissolve, or be vindicated at any stage, as evidence accumulates.
Never-ending is not never-pinned. Each stage still commits its goalposts before the data is seen, registered publicly on an open-science record, with the exact version of the instrument fingerprinted so anyone can check that the build which produced a result wasn’t quietly changed afterwards. The quest is endless; every step of it still calls its shot.
An open invitation.
We offer this; we don’t impose it. If you’d like to help validate or sharpen the work, there’s room for you. For researchers: the pre-registration is open, and the cross-test data building up here — many frameworks measured on the same people — is an unusual asset to work with. For students: it’s a live project to cut your teeth on. For everyone else: taking the test, and telling us whether your result rings true, is itself part of the evidence.
Before we lock the first stage’s goalposts, there’s a short window to help us sharpen them. When that stage is registered, we’ll link the timestamped record from this page — the called-the-shot receipt, in public.