The Dark Mirror Index
Edition No. 196 · daily
Updated
Sample 11,450 takes+40 since yesterdaySince 4 April 2026 · 6 axes · 20 items49 regions with 30+ takes

The Dark Mirror Index

Release calendar

The Index recomputes every morning. On the first of each month the day's edition becomes the edition of record: archived, citable, and the one deposited for a permanent identifier. Every figure describes people who took the Dark Mirror quiz; the methodology and disclosure page states exactly what that means and what it does not.

Next release

November 1, 2026

Writing about it? Embargoed advance figures are available to press: Kanika@kanikarose.com with your outlet, and the next edition arrives before it publishes.

How the Index has grown

2,9005,7008,60011,500AugSepOct
People who took the Dark Mirror, by edition: 2,202 on 16 Aug 2026 to 11,450 on 16 Oct 2026. Takers, not takes: retakes and sub-60-second completions are excluded.

Editions of record

Today's live edition: 196, n = 11,450. Aggregates for any edition: open data, CC BY 4.0. Numbered citable facts: statistics. Permanent identifier for the dataset: 10.5281/zenodo.22228970.

Your numberTake the Dark Mirror, seven minutes, and see where you land on this curveThe long readDark psychology, the beginner's guideThe bookThe Sociopathic Dating Bible, the instrument this quiz grew out of

Method

The Dark Mirror is a 20-item scenario quiz. Each answer is attributed to one of six axes; a score is the share of your answers that fell to that axis, so scores move in steps of five. Your type is simply your highest axis.

Baseline
0 to 15
Present, not a feature.
Traits
20 to 35
Recognisable and common. Not the person.
Pronounced
40 to 55
The axis is shaping choices.
High
60 to 100
Dominant. The only band to read as the type.

Every axis is read in four bands: baseline (0 to 15), traits (20 to 35), pronounced (40 to 55) and high (60 to 100). Bands are fixed cut points chosen so that high means at least twelve of twenty answers went to one axis. They are the Index's reading aid, not clinical thresholds.

Percentiles are computed against everyone who has taken the quiz and recomputed daily at 09:41 UTC; each recomputation is an edition. Ties are split down the middle: higher than 91% means 91% of takers scored strictly lower, plus half of those who tied.

Regions are browser timezones, so New York means Eastern-time takers rather than the five boroughs; regions with fewer than thirty takes are held back. Nothing here is a population estimate and nothing here is a diagnosis.

Written and read by Kanika Batra, diagnosed ASPD, clinically assessed as Factor 1 psychopathy. An audience portrait, not a census.

The full method, exclusions, reliability and how to cite