We owe you transparency about an instrument that asks for twenty minutes of your honest reflection. This page describes how the EnneagramGenius v9.8 assessment was built, how scoring works, and what we currently claim — and don’t yet claim — about its psychometric properties.
The current instrument (v9.8) consists of 113 forced-choice and Likert items designed to elicit the underlying motivations, fears, and desires that organize an Enneagram type — not the surface behaviors people most readily report about themselves. Each item was authored to satisfy three constraints: it has to discriminate between at least two types, it has to be readable in a non-clinical voice, and it has to be answerable honestly by someone who has never heard of the Enneagram.
The item bank deliberately avoids the most common failure modes of mass-market personality tests: stereotyped trait language (“I am a hard worker”), positively-skewed phrasing that everyone agrees with, and items that conflate type with mental health diagnoses. Items that did not separate types in piloting were rewritten or retired.
Scoring produces four distinct outputs from a single sitting:
Each output carries its own confidence value, and the report tells you which of the four is most and least decisive for you. A clean type with a borderline wing reads differently than a tritype where two centers are nearly tied — and we surface that distinction rather than hide it behind a single number.
The confidence score is the model’s estimate of how clearly your response pattern separates from neighboring types — not how psychologically certain you should feel about the result. A high confidence score means your responses pulled strongly toward one type and away from others; a moderate score means the data are real but the picture is more layered, often because of strong wings, integration/disintegration influence, or counter-typing patterns common at certain levels of development.
We deliberately do not collapse uncertainty into a false single-digit answer. When the top two types are within the tie-breaker margin, the assessment surfaces both and asks a small set of adaptive items designed to discriminate them directly.
The instrument is grounded in five strands of the Enneagram tradition. We do not invent theory; we interpret it carefully and translate it into items and reports that an honest reader can use.
The v9.8 instrument is calibrated against an internal pilot dataset and the iterative feedback of practicing Enneagram coaches who sit with clients across all nine types. We treat this as adequate for a free, public-facing instrument — and inadequate for any clinical, hiring, or selection use, neither of which we recommend or support.
We are currently engaged in broader re-norming work: a larger, more demographically representative sample, retest-reliability analysis, and explicit measurement of the confidence score against expert-rated cases. As that work completes, we will publish the sample size, the retest correlations, and the test–expert agreement rates on this page. We will not publish numbers we cannot defend.
If you are a researcher, coach, or clinician who would like access to the methodology details, contact us at research@enneagramgenius.com.
The Enneagram is most useful as a frame for self-knowledge, relational repair, and growth — not as a label to be assigned to other people without their consent, and not as a tool for selection or evaluation. This assessment is designed for the first uses and against the second.
Two practical implications: we never share results with anyone but you unless you explicitly invite them in, and the platform will never sell, lease, or otherwise expose your response data to advertisers or data brokers. See our privacy policy for the details.
About 20 minutes. No account required to start. Your full profile — type, tritype, wing, instinct stack — is free forever.
Take the Free Test