Sources & methodology

Our other assessments are built on question sets that researchers have given to large groups of people over many years, so we can tell you how consistently each set measures what it claims to. This one is different. We wrote these 90 questions ourselves, and none of them has been through that testing yet.

We did it this way on purpose. Every well-known Enneagram questionnaire is someone’s private property, questions included, so the choice was between licensing one and writing our own. What we can’t do is borrow another questionnaire’s evidence to describe ours — a reliability figure belongs to a specific set of questions, not to the Enneagram in general.

What that means for you: read your result as something to try on rather than something settled. The description either sounds like you from the inside or it doesn’t, and right now your own read is the better evidence of the two.

• 90 questions in total — 72 across the nine patterns (8 each) and 18 across the three instincts (6 each).
• Each one is answered on a 1 to 5 scale, from "Strongly disagree" to "Strongly agree".
• 24 of them are worded the other way round — 2 on every scale. Those are scored in reverse, so agreeing with everything doesn't push any one pattern to the top. It comes out as a flat profile instead, which is the honest answer for someone who agreed with everything.
• The questions are shuffled through each other rather than grouped: no two questions in a row come from the same pattern, so you never answer a block about one thing and then move on.
• All the wording is original. None of it is taken or adapted from any published Enneagram questionnaire.

Each of the nine patterns gets an average from its 8 questions. Those nine averages are put in order, and the one at the top is reported as your type. That’s the whole calculation — there is no separate step where you get sorted into a category.

This matters for how much weight to put on it. The order is a comparison with yourself: which of the nine sounded most like you, set against how you answered the other eight. It is not a comparison with other people, and nothing here can tell you how you’d look next to anyone else.

The other eight are still in you. This is an order, not a scoreboard — a pattern near the bottom of yours is one you reach for less often, not one you’re missing.

One limit worth knowing, because it’s ours rather than yours: some of these patterns are easier to agree with than others. Until we have enough responses to see how people generally answer, a pattern whose questions are simply easier to say yes to has a slight advantage in everyone’s order, including yours. Collecting the responses that would let us correct for that is what the testing is for.

Sometimes two or more patterns come out exactly level at the top. When that happens we say so rather than picking one, because the only thing that could break the tie is the order we happen to list them in — and that would be presenting a coin flip as a finding.

A close result is treated the same way, just more gently. We compare the top pattern with the one just behind it:

• Under 0.3 apart — close enough that a handful of different answers would have swapped them. We name the type, and say plainly not to lean on it.
• Between 0.3 and 0.8 — ahead, but not by much. Worth reading the runner-up too.
• Over 0.8 apart — clear enough to lead with.

To put those in proportion: one question answered one step differently moves a pattern’s average by about 0.13. So the gap that separates a clear result from a close one is roughly 6 single-step answers.

These boundaries are our judgment, not measurements. They came from the people who designed this assessment rather than from data, and they are among the first things we expect to change once there are enough responses to set them properly.

On the Enneagram the nine patterns sit in a circle, so each has two neighbors. Your wing is whichever neighbor came out higher.

We asked you no questions about your wing at all. It is worked out entirely by comparing your scores on the two patterns either side of yours, which makes it a second-hand result — it inherits whatever is loose in those two scores rather than having any evidence of its own. When the two neighbors are within 0.2 of each other we report a balanced wing rather than picking one.

Some Enneagram questionnaires ask about the wing directly. Ours doesn’t, and that’s a real difference, not a detail. Hold the wing more loosely than the type.

The instincts are scored separately from the nine patterns and never compete with them. Any type can lead with any instinct.

This is the shortest part of the assessment — 6 questions each, against 8 for every pattern — so it is also the part most likely to come out flat. When the top two are within 0.3 of each other we report both rather than one, and when none of the three separates we say so.

• Whether your types are compatible. No assessment here has looked at two people at once, so a pairing verdict would be invented. Type-pair compatibility is the most commonly published Enneagram claim and one of the least supported.
• How either of you behaves under stress or in growth. The Enneagram is usually drawn with arrows between types showing where someone is said to move. Nothing here measures that for either of you.
• Which of you is further along. Some Enneagram material sorts people into levels within a type. We don't, for either of you, and nothing here ranks one of you above the other.
• A score, from putting your two orders side by side. A pattern sitting ninth for one of you and second for the other means it came out further down that person's own list — not that they have less of it than you do.

After the 90questions we asked whether you’d met the Enneagram before, and if so what you already thought your type was. Those answers are not scored and change nothing about your result — you get the same report either way.

They matter because comparing what this works out against what people already believed is the most useful check available on a new assessment, and it can only be collected in the gap between finishing the questions and seeing the result. Asking earlier would tell you what the questions are looking for. Asking afterwards would only measure whether you agree with what you’d just been told.

Ticking the research box is optional and always was. Your answers are only kept at all if you choose to save your result — if you delete it, they go with it.