How the shadow test measures and where the limits are
What the test measures
24 statements on four sides we rarely talk about: the mask (showing whatever is useful at the moment), the cold calculator (own advantage against the rules), the hunger (the pull of status, money and influence) and the stage (standing taller than the others). In research these four facets are called sincerity, fairness, greed avoidance and modesty; together they form the honesty-humility dimension of the HEXACO model.
The link to the well-known “dark triad” (narcissism, Machiavellianism, psychopathy): low honesty-humility is considered in the literature to be the common core of these three traits. So we measure the axis on which they converge, rather than imitating protected questionnaires.
Where the questions come from
All 24 statements come from the IPIP (International Personality Item Pool), a public-domain scientific item pool that is explicitly free to use, including commercially. Well-known triad questionnaires such as the SD3, Dirty Dozen or Mach-IV are protected by copyright and are not used here.
We deliberately removed two original items even though they are, measurement-wise, the sharpest of the fairness facet: statements about actual crimes. About yourself, anyone may tick such a box; in the outside rating the same statement would be a criminal allegation about an identifiable person. This sacrifice measurably costs precision (the facet's reliability drops from 0.84 to 0.72), and we pay that price on purpose.
The reliability values, recomputed rather than borrowed
Because we shorten the original scales from ten to six statements per side, their published values no longer apply to our version. We therefore recomputed them on a public dataset of about 22,800 people that contains the very same items. Reliability (Cronbach's alpha) of our short form: the mask 0.73, the cold calculator 0.72, the hunger 0.77, the stage 0.74, overall value across all 24 statements 0.88.
This yields the hierarchy you also see in the result: the overall value is measured most precisely, while the four individual sides, with six statements each, are markedly blurrier. We show the measurement error on every axis instead of hiding it.
Why there are no percentile ranks
Your result sits on a scale from 1 to 7, without a comparison against “the population”. That is deliberate: an honest comparison would require a representative norm sample, and we do not (yet) have one. Passing off a borrowed or self-selected sample as a “norm” would be exactly the fake precision this site stands against.
The test's main statement needs no norm at all: the comparison between self-view and outside view is measured within your pair. A difference is only claimed when it is larger than the combined measurement error of both measurements; anything below that is explicitly called “in agreement”.
The second person: what is stored and what is not
You can invite one person to answer the same 24 statements about you. The invitation link works exactly once and expires after 14 days; its language is locked to yours, so that you both answer the same statements rather than two translations being compared. The invited person confirms before the first question that you will see their rating.
We explicitly do not claim that this rating is anonymous towards you: from the displayed values one can work back to roughly how the other person answered. Both sides know this from the start. If you delete your result, the rating about you is deleted with it.
What the test is not
Not a diagnosis and not a verdict on your character. What is measured are normal psychological traits, not disorders; high values do not mean something is wrong with someone. The test is not suitable for personnel selection and not meant for judging third parties without their knowledge. And a difference between self-view and outside view never means one side is lying: whoever rates themselves knows their intentions; whoever rates someone else sees only what comes across.
Price and data
The result and the pair comparison are entirely free. No account, no e-mail address, no tracking; your result hangs on an anonymous link only you know, and you can delete it at any time. Details are in the privacy policy.