A dimensional model designed with research-oriented concepts. This implementation itself still requires validation.
Evidence, method, and privacy
See what the model can support—and where it stops.
Personality Synthesis Engine follows the same written scoring rules every time. It is a self-reflection tool, not a validated psychological test. That claim would require a specific, unchanged version to complete a properly declared study.
How an answer becomes a result
- Answer a situation: choose the action you would most likely take, say that none fit, or choose not to answer.
- Score the underlying tendency: each answer can affect no more than two things the test measures directly, using amounts set in advance.
- Track the strength of the information: the model keeps the amount, variety, agreement, situation, and current question-design estimate separate.
- Build familiar views: those scores can create four-letter patterns and other familiar views, but those views do not count as new evidence.
- Look for supported connections: written rules describe only relationships supported by the answers, including what could complicate them and what might help.
- Let you inspect it: you can review the reasoning and decide what fits without changing the result.
The 28 things the test measures directly
GroupCountScope
Four-preference anchors4Energy, information, decision emphasis, and closure orientation
Big Five backbone5Openness, conscientiousness, extraversion, agreeableness, and emotional stability
Social operation6Autonomy, approval, status, hierarchy, influence, and social selectivity
Reflective decision process7Reflection, consequences, revision, uncertainty, ethics, perspective, and proportional intervention
Communication2Directness and emotional explicitness
Learning and knowledge handling4Productive struggle, retrieval/help, experimentation, and error integration
Four-letter codes, social archetypes, combined archetypes, communication descriptions, Active/Receptive capacities, and connection findings are built from those 28 measures. They are not extra things being measured.
One section cannot make another more scientific
A familiar continuous four-preference projection, not the official Myers-Briggs assessment.
A platform-defined dimension whose usefulness and reliability must be tested.
The Alpha/Beta/Gamma/Delta/Sigma/Omega/Zeta lens is not a scientific taxonomy or social rank.
Astrology and Calendar Archetypes are optional motifs and cannot modify psychological scores.
Versioned interaction and composite rules selected from already scored evidence; not an independent confirmation.
What confidence does—and does not—mean
Confidence shows how much usable, varied, and consistent information came from your answers, together with the current design estimate for the questions. For a letter or label, it also considers how close the score is to a boundary. The site shows model maturity separately, so consistent answers cannot make an experimental scale look scientifically validated.
- A score out of 100 shows where you landed in this model. It is not a population percentile.
- When a four-letter position is too close to call, the site shows
xrather than forcing a letter. - Different answers in different situations may be a real pattern, not a contradiction.
- Several labels built from the same underlying answers do not confirm each other.
What a saved link or file contains
A saved link or file contains the saved information itself. It is compressed and checked for changes, but it is not encrypted. Anyone who has it can copy or forward it, and the site cannot disable it later. A hash can check or identify the contents; it cannot recreate a lost profile. If every copy is lost, the site cannot recover it.
Technical save formats
r— private resume- Contains unfinished answer state. Keep private.
a— private archive- Contains raw answer IDs and the completed result. Keep private.
p— shareable profile- Contains minimized derived results and no raw answer IDs.
c— comparison- Embeds two complete shareable profiles and a comparison result.
These formats live in the part of a URL after # or in a downloaded .pegasus file. Encoding and compression make the format compact. A random nonce distinguishes separate saved copies, and a hash helps detect changes. None of them encrypts the saved information.
No third-party advertising, session replay, analytics, remote fonts, or remote assets are included in this static web application. A production privacy claim still requires network, logging, browser, and deployment verification; source review alone is not proof.
Material limitations
- Scenario answers are self-predictions, not observed behavior under real stakes.
- Social desirability can distort ethics, empathy, revision, reflection, and autonomy answers.
- Language, culture, disability access, fatigue, device, and context can change how an item functions.
- The seven internet archetypes are informal cultural labels, not natural or ranked kinds.
- Compatibility does not predict success, safety, attraction, consent, loyalty, or obligation.
- Personality does not imply intelligence, morality, diagnosis, trauma, abuse, or protected status.
- Astrology does not provide psychological evidence and cannot confirm a measured result.
Accessibility target
The product targets WCAG 2.2 Level AA with semantic landmarks, full keyboard operation, visible focus, non-color status labels, reflow at narrow widths, 200% text resize, 400% zoom, reduced-motion behavior, and representative screen-reader testing. This is a target, not a conformance claim. Automated checks cannot replace human evaluation, including with disabled users.