Methodology

How a test is scored

1. Each answer becomes a number from 1 to 5.
2. A reverse-keyed item is mirrored, so 5 becomes 1 and 4 becomes 2.
3. The answers of each facet are added up.
4. The facets of each domain are added up.

The site shows a raw score with its minimum and its maximum, for example 16 of 20. The scoring code is a pure function, and the tests check it against the published scoring key.

Why there are no percentiles

A percentile needs a reference sample. The International Personality Item Pool publishes no norms, and it advises against canned norms, because no internet sample is representative of a given visitor.

This site therefore describes a score against the scale itself: above the midpoint, below it, or on it. It never ranks you against other people. The source statement is at ipip.ori.org/newNorms.htm.

How a test enters this site

• The licence must be public domain, Creative Commons, or free with attribution. Research-only and unconfirmed licences are excluded.
• The licence text, the source link, and the citation live in the data file beside the items.
• The items, the scoring direction, and the reliability values are imported from the source instead of being typed by hand.
• The import fails when the source does not produce the expected item count, facet count, or reliability values.

Instruments published here: Anxiety screen (GAD-7), Big Five personality test (IPIP-NEO-120), Big Five short test (IPIP-50), Depression screen (PHQ-9).

What the site does not claim

• A result is not a diagnosis and not clinical advice.
• Self-report answers change with mood, context, and the order of the items.
• A reliability value describes a sample, not the accuracy of a single result.

The full plan, including the licence review of every candidate instrument, is in docs/PLAN.md.