How our assessments work
Every High Performance Culture assessment is built by the same pipeline, scored by the same engine, and held to the same rule: structured self-reflection, honestly labelled — never a diagnosis dressed up as science.
What these instruments are
Each assessment is a scenario-based self-assessment of 20 to 25 items that reads one construct — habit strength, sleep recovery, deep-work capacity, crisis readiness — across five dimensions drawn from the peer-reviewed literature on that construct. The output is a profile: an overall reading, a per-dimension breakdown, your strongest dimension, and a short list of first structural moves.
They are educational instruments, not clinical ones. No High Performance Culture assessment is a validated diagnostic, and none substitutes for professional evaluation. Result bands are editorial guidance, not clinical thresholds.
How items are authored
Every instrument passes through a staged pipeline with hard gates between stages:
- Research. A research pass gathers the peer-reviewed sources for the construct. Every citation is registered in a sitewide registry and its DOI verified against Crossref before it can be used. Sources that fail verification never enter an instrument.
- Item authoring. Questions are written as concrete scenarios against those sources — each item is anchored to a specific finding, and the evidence reveal you see after answering cites it.
- Science gate. An automated gate checks every claim-to-citation link, every scoring structure, and every band definition. A failed check blocks the build — the instrument cannot ship.
- Editorial QA. A final audit pass reads the assembled instrument end to end before deployment.
Why forced choice, four options, no midpoint
Each item offers four graded responses and deliberately omits a neutral middle option. Midpoints attract non-answers — the “sometimes, it depends” response that produces flattering, uninformative profiles. Removing the midpoint forces a lean, and the scenario framing (“what did you actually do this week”) pulls answers toward recalled behaviour rather than self-image. Response order is shuffled on every run so position habits cannot pattern your answers.
How scoring works
Each response carries a weight from 1 to 4. The engine sums an overall reading and computes a per-dimension profile. Two honesty rules constrain what the engine may claim: compound-risk flags (patterns that span dimensions) fire only when every involved dimension actually shows the pattern — a strong dimension suppresses the flag rather than letting arithmetic invent a crisis; and the “strongest dimension” card relabels itself when no dimension is genuinely strong, rather than congratulating the least-bad reading.
How the bands were derived
Score bands (the verdict tiers) were set editorially, anchored to the content of the items: the boundaries mark where the pattern of responses changes character, not where a validated clinical cut-off sits — no such cut-offs exist for instruments like these. We do not show percentiles or “you versus other readers” comparisons, because your scores never reach us and we will not invent a population.
Privacy
Answers and results are stored only in your browser's local storage. Nothing is transmitted — no account, no server-side record, no analytics on your responses. An unfinished run expires from your browser after 30 days. A completed profile stays on your device until you retake the assessment or clear your browser data. Retaking preserves your previous reading so the instrument can show you what changed — the Second Reading.
Sharing a result. The “Copy result link” button encodes your dimension scores and the instrument version into the link itself — nothing else. No item-level answers, no timestamps, no identifiers travel with it, and nothing is stored anywhere: the recipient's browser decodes the scores from the link and renders a read-only reading. Tampered or malformed links are ignored. If a shared result was scored under an older instrument version, the page says so.
Known limitations
We would rather name these than have you discover them:
- Self-report. Every item measures what you say you do. Recall is imperfect and self-image leaks into answers; the scenario framing reduces this but does not eliminate it.
- Acquiescence and social desirability. In v1.0, all items were keyed in the same direction — the more flattering answer was generally recognisable. Instrument v1.1 (rolling out per instrument; each assessment's appendix shows its own version) adds 4–6 reverse-keyed items per instrument, where the attractive-sounding response — grit over redesign, flexibility over consistency, willpower over environment — is the structurally weak one. Option shuffling continues to blunt position effects. Self-presentation bias is reduced, not eliminated.
- No population norms. Bands are editorial, and no normative sample stands behind them. Treat the profile as a structured mirror, not a measurement against other people.
- Educational scope. These instruments inform reflection and habit design. They do not detect, diagnose, or rule out any condition.
Version history
- v1.1 · August 2026 — Reverse-keyed items: 4–6 items per instrument rewritten so the flattering-sounding response is the structurally weak one, with keying metadata in every payload (counter to acquiescence and social-desirability bias). Shareable result links (dimension scores + instrument version only — never answers). Cross-links from each dimension to its protocol, deep dive, and glossary entries. Because item rewording changes what a score means, the Second Reading marks any comparison across instrument versions as approximate. Rolling out per instrument — each assessment's appendix shows the version that scored it.
- v1.0 · July 2026 — Initial release of the assessment estate: 20–25 items across 5 dimensions per instrument, Crossref-verified evidence files, in-place evidence reveals, Second Reading, honest-encoding scoring rules (compound-risk floors, strongest-dimension relabelling).
Compiled and maintained by the High Performance Culture editorial team. Questions about the method? Every assessment's appendix lists its full evidence file with DOIs.