How Reliable Is Zensitively's Neurodivergence Test? The Psychometric Properties in Numbers

The psychometric properties in numbers: what the ZNI measures, whom it is normed on, how stable its scores are, and what it cannot do. A summary of the technical documentation, version 1.0, September 2026.

What the ZNI is

Zensitively's neurodivergence test, the Zensitively Neurodivergence Inventory (ZNI), is a psychometric self-report instrument: 60 statements, five response options, scores from 0 to 100. It describes how you perceive, focus and react by your own account, and what this currently costs you in everyday life. Seven scales form its backbone: High Sensitivity, Monotropism, Kinetic-Cognitive Learning Style, Alexithymia, Emotional Overwhelm (associated with rejection sensitive dysphoria, RSD), Camouflaging and Self-Criticism. The results report shows values computed from the same 60 answers; the technical documentation reports the psychometric properties of the seven scales and of the total score.

The ZNI does not make a diagnosis. A psychometric instrument measures and describes where you stand. That is its job, and that is what it is judged by.

What "psychometric" means can be checked. An instrument earns the word when its quality criteria are known: how reliably it measures, how stable its scores are, whether it measures what it claims to measure, and whom it compares you with. This article lays those numbers open. The full technical documentation is available as a PDF (link at the end).

Where the numbers come from

By September 2026 the ZNI had been completed 155,609 times, 129,800 of them in German. The figures in this article come from a reference group of 9,270 people: German-language first completions between 16 July and 16 September 2026, with a completion time of at least two minutes and without click-through response patterns.

One caveat belongs at the start. Whoever completes the ZNI went looking for it. The reference group is therefore not a population sample, and the age, gender and country of residence of participants are not collected. When the report compares your score with others, those others are people who sought out a test on neurodivergence of their own accord. That group is the comparison the data can support.

How reliably the scales measure

Reliability here means that the statements of a scale measure the same thing. The usual index is Cronbach's alpha; values from .70 count as adequate, from .80 as good.

Scale	Statements	Alpha	Omega
High Sensitivity	9	.81	.82
Monotropism	10	.85	.85
Kinetic-Cognitive Learning Style	8	.81	.82
Alexithymia	10	.90	.91
Emotional Overwhelm	6	.74	.74
Camouflaging	13	.91	.91
Self-Criticism	2	.82

Six scales lie between .74 and .91. Self-Criticism consists of two statements and is therefore carried as an indicator, not as a scale. Every score comes with a measurement error; the documentation states it for each scale, and it lies between 7 and 11 points. Differences between two of your scores that are smaller than roughly 20 to 30 points, depending on the scale, therefore cannot be interpreted.

How stable the scores are

A trait score is only worth something if it is still roughly the same after a few weeks; strain scores are allowed to change, but they should not fluctuate at random. In 1,795 cases the ZNI was completed a second time under the same user ID after 14 to 90 days. The scores at the two points in time correlate at .67 to .79, depending on the scale. After three to twelve months (2,692 cases) the correlations are .60 to .74.

Some of these repetitions do not come from the same person, because a user ID is sometimes passed on to a partner or a friend. We estimated that share by comparing the response patterns at both points in time: about one pair in fifteen is a different person. Adjusted for that share, stability over two weeks to three months is .72 to .85. We report both figures. Pairs were not filtered by similarity; that would inflate stability artificially.

Whether the scales measure what they claim

For validation, 445 people completed five established instruments in addition to the ZNI: the Highly Sensitive Person Scale, the Monotropism Questionnaire, the Conners' Adult ADHD Rating Scales (CAARS), the Toronto Alexithymia Scale (TAS-20) and the Camouflaging Autistic Traits Questionnaire, with a median interval of 160 days. This makes it possible to test whether the ZNI scales measure the same thing as the established instruments.

Because months lie between the two sets of answers, the question is not whether the correlation is 1, but how close it comes to what the established instrument correlates with itself over the same interval.

ZNI scale	Established instrument	Correlation	Share of the ceiling
High Sensitivity	Highly Sensitive Person Scale	.70	95 %
Monotropism	Monotropism Questionnaire	.67	90 %
Kinetic-Cognitive Learning Style	Conners' Adult ADHD Rating Scales	.69	95 %
Alexithymia	Toronto Alexithymia Scale	.71	94 %
Camouflaging	Camouflaging Autistic Traits Questionnaire	.69	95 %

With the other instruments these scales correlate markedly lower, .30 to .41; Camouflaging also correlates at .57 with the RSD checklist, because both capture withdrawal for fear of rejection. This suggests that each scale captures mainly its own construct; the scales are not independent of one another, as the next section shows. Two exceptions are stated explicitly: for Emotional Overwhelm, the reference is a clinical screening checklist for rejection sensitive dysphoria, not a validated instrument; we carry the agreement of .60, 83 percent of the ceiling, as supporting information. Self-Criticism has no validation instrument of its own.

How the ZNI is structured

A factor analysis of the 60 statements shows a strong common factor; a parallel analysis indicates eight factors in total, seven of which correspond to the scales. In content, the common factor is the factor of strain: its strongest statements are about pretending, suppressing one's own reactions and not knowing what one feels. That is a finding about the people who complete the ZNI, and it carries the interpretation: the costs are connected, and whoever pays in one place usually pays elsewhere too.

Strict structural models, in which every statement may belong to its own scale only, therefore do not quite reach the customary fit values: CFI .82 for the seven scales, .85 for a model with a common factor, at RMSEA .05 and SRMR .05 to .06. The documentation reports these values with all their limitations. In four further languages, Spanish, French, Portuguese and Turkish, the ZNI has a similar structure; for English the agreement is lower, and for Italian the sample is still too small for the seven scales.

The total score

The Neurodivergence Score, the report's total score, correlates at .75 over two weeks to three months. By its content it does not measure how different someone is, but how far a person's own experience is discriminated against by neuronormativity, in other words what someone pays for their way of experiencing. Its computation is not published; its stability and its distribution in the reference group are in the documentation.

What the ZNI cannot do

• It does not make a diagnosis and says nothing about whether you have one. No data on clinical diagnoses were collected, and there are no cut-off scores.
• Its reference group consists of people who sought out the ZNI, not a population sample; its composition by age and gender is unknown.
• All scores are self-reports.
• The check against the established instruments rests on time-lagged answers from the same people, not on concurrent assessment.
• Self-Criticism consists of only two statements.
• About one repetition pair in fifteen comes from a different person under the same user ID.

What will be examined next

A repeat study with explicit consent to a second completion, a concurrent assessment of the ZNI scales with the established instruments, a test of whether the statements work the same way in every language, and the collection of age and gender for fairness analyses.

The technical documentation

The full documentation with all tables, sample definitions, confidence intervals and references: Zensitively Neurodivergence Inventory (ZNI): Technical documentation, version 1.0, September 2026, as an HTML version and as a PDF; the DOI will be added upon registration. Professionals can obtain further figures on request at hallo@zensitively.com.