Built on real science

The science behind Cogniself.

Validated assessments provide the scores. Our AI connects them with your context and relevant research to help you understand your patterns and what they mean in everyday life.¹⁶

Explore the evidenceMethodology review · August 2026

The interpretation gap

Psychometric instruments produce dimensional scores, facet patterns, and overlapping signals; the instrument result is the beginning of interpretation. Relating those variables to one another, to a person’s context, and to the scientific literature has traditionally required specialist time. Popular type-based frameworks achieve accessibility by compressing continuous variation into a small set of categories. Cogniself preserves the dimensional structure. Its AI synthesis layer evaluates patterns across traits, facets, screeners, and context, then links the interpretation to relevant scientific findings. This makes nuanced behavioral-science interpretation practical at scale while keeping measurement and scoring anchored to published instruments.⁹

Foundation at a glance

Five published instruments form the measurement foundation described in this note.

953,398combined observations in the five studies shown below

IPIP-NEO-120
CATI
RAADS-14
CAT-Q
ASRS v1.1

Combined study observations · source-level evidence

How your results come together

Three evidence-bearing inputs remain separately represented until synthesis. Deterministic transformations preserve the provenance and interpretive role of each input.

Interpretive functionI = f_θ(x, c, E | G)

x scored constructs · c structured context · E retrieved evidence · G report specification · I draft interpretation · R verified report

Measurement input

Validated psychometric battery

Published item content, response scales, scoring keys, facet structure, and screening thresholds.

Deterministic scoring

Domains · facets · dimensions · thresholds

Context input

Respondent context

Goals, environment, lived experience, current questions, and self-described situational constraints.

Context structuring

Goals · environment · history · constraints

Evidence input

Scientific evidence layer

Construct-matched peer-reviewed findings, source metadata, study populations, and DOI provenance.

Construct-matched retrieval

Claims · populations · estimates · DOI metadata

Model-mediated interpretation and verification

Evidence-conditioned synthesis

A specialized model evaluates cross-trait, cross-facet, screening, and person-by-context patterns against the retrieved evidence set.

Verification gate

Score fidelity
Construct consistency
Citation provenance
Schema completeness

Grounded interpretive report

Dimensional profile, contextual implications, screening interpretation, evidence links, and calibrated uncertainty.

Figure 1. Cogniself inference architecture. Measurement variables (x), structured context (c), and retrieved evidence (E) retain distinct provenance through synthesis. Verification is applied to the complete report artifact before release.

Selected foundational validation samples

Scale of the primary instrument-development and validation samples referenced in Cogniself’s methodology.

10²10³10⁴10⁵10⁶

IPIP-NEO-120⁰¹

948,692

CATI⁰⁵

2,487

RAADS-14⁰³

1,233

CAT-Q⁰⁴

832

ASRS v1.1⁰²

154

Figure 2. Logarithmic scale. The selected studies contribute 953,398 observations or participants in aggregate. This is a study-level evidence map: sample designs, populations, and units of observation vary across publications.

Methods

A six-stage, measurement-first pipeline

Each report follows the same ordered sequence: instrument administration, score derivation, structured profile assembly, literature retrieval, report synthesis, and output verification. Resolving scores and published thresholds before synthesis preserves the source instruments’ scoring logic while enabling cross-construct interpretation.

01
Instrument administration
Published item wording and response scales are presented with battery-integrity checks.
02
Score derivation
Fixed keys calculate domain, facet, dimensional, and screening variables.
03
Structured profile
Calculated variables are assembled with their construct and instrument provenance.
04
Literature retrieval
Construct-specific claims are matched to peer-reviewed records and DOI metadata.
05
Report synthesis
A specialized model integrates the scored profile, verified context, and retrieved evidence.
06
Output verification
Schema, citation provenance, score consistency, completeness, and safety checks determine release.

What the validation studies report

Selected metrics retain their definitions from the source publications; interpretation is study-, population-, and construct-specific.

WHO ASRS v1.1n = 154 · Kessler et al.²

sensitivity

68.7%

specificity

99.5%

classification accuracy

97.9%

RAADS-14n = 1,233 · Eriksson et al.³

sensitivity

97%

RAADS-14 specificity by comparator

ADHD

46%

Psychiatric

64%

Non-psychiatric

95%

CATIn = 2,487 · English et al.⁵

internal consistency · α

95%

convergent validity · r ≥

79%

42items · six subscales

Figure 3. Published results from foundational studies. Sensitivity and specificity are conditional on study design, population, comparator group, and threshold; each value characterizes performance within its reported validation setting.

Validity model

Evidence is evaluated at three levels

Cogniself assigns evidence to the layer where it is generated. This preserves the distinction between instrument-level psychometrics, implementation-level reproducibility, and interpretation-level quality assurance.

Measurement

Instrument-level evidence

Source instruments contribute peer-reviewed evidence for reliability, construct validity, and screening performance in the populations studied.

Computation

Implementation-level evidence

Versioned scoring keys and published thresholds are executed deterministically and verified through software quality controls.

Interpretation

Interpretation-level evidence

Structured synthesis integrates the scored profile with retrieved literature and is evaluated for consistency, provenance, completeness, and safety.

Cogniself extends the practical frontier of psychometrics by making multivariable interpretation scalable. Validated instruments provide measurement, deterministic scoring preserves their logic, AI integrates the resulting patterns with personal context and scientific literature, and expert judgment governs high-stakes use.Cogniself methodology statement · 2026

Evidence register

Availability, provenance, and use

Every instrument has a different publication and data-access history. The register distinguishes public-domain items, open publications, supplementary materials, and data available only by request.

Instrument	Construct	Items	Foundational sample	Evidence availability	Cogniself use
IPIP-NEO-120¹	Big Five + 30 facets	120	948,692	Public-domain item pool; peer-reviewed paper	Primary personality measurement
WHO ASRS v1.1²	Adult ADHD screen	6 / 18	154	Published validation study	Screening anchor; published threshold
RAADS-14³	Adult autism screen	14	1,233	Open-access validation article	Screening anchor; published threshold
CAT-Q⁴	Autistic camouflaging	25	832	Published scale and psychometric study	Contextual dimension
CATI⁵	Autistic trait dimensions	42	2,487	Open article + scoring supplement; data by request	Contextual dimension

Figure 4. Availability is reported for each cited foundational source, distinguishing item access, publication access, supplementary materials, and author-mediated dataset access.

Conditions of interpretation

Scope, transportability, and appropriate use

Cogniself is a technology company. Its reports are educational and designed for structured self-understanding and to support the work of psychologists, psychiatrists, researchers, coaches, and other qualified experts. Cogniself does not provide clinical care, diagnosis, or treatment, and its reports do not replace human expertise. The methodology has received clinical and methodological input through external expert advisory review in the United States; Cogniself does not currently maintain an in-house clinical team.

Cogniself supports self-understanding, reflection, and screening; it does not replace professional assessment, diagnosis, or treatment. The complete product and its AI interpretations have not been independently clinically validated.

Response basis

Interpretation begins with structured self-report and represents patterns in the respondent’s account of their typical behavior and experience.

Population transport

Evidence is strongest for the populations, languages, and administration conditions represented in each source study; new settings are treated as transportability questions.

Generated synthesis

Interpretations are evaluated against the scored profile, retrieved evidence, and output contract. Consequential applications add independent evidence and qualified human judgment.

Screening interpretation

Screening scores estimate signal relative to published thresholds. Professional evaluation integrates those scores with developmental history, differential assessment, and clinical judgment.

References

Primary literature

Instrument foundations are supplemented by a broader literature on dimensional personality structure, categorical typologies, life outcomes, prediction, and replication.

1
Johnson, J. A. (2014). Measuring thirty facets of the Five Factor Model with a 120-item public domain inventory. Journal of Research in Personality, 51, 78–89.
Open source →doi:10.1016/j.jrp.2014.05.003
2
Kessler, R. C. et al. (2005). The World Health Organization Adult ADHD Self-Report Scale: a short screening scale. Psychological Medicine, 35(2), 245–256.
Open source →
3
Eriksson, J. M. et al. (2013). RAADS-14 Screen: validity of a screening tool for autism spectrum disorder. Molecular Autism, 4, 49.
Open source →
4
Hull, L. et al. (2019). Development and Validation of the Camouflaging Autistic Traits Questionnaire. Journal of Autism and Developmental Disorders, 49, 819–833.
Open source →
5
English, M. C. W. et al. (2021). The Comprehensive Autistic Trait Inventory: development and validation. Molecular Autism, 12, 37.
Open source →
6
Ozer, D. J. & Benet-Martínez, V. (2006). Personality and the prediction of consequential outcomes. Annual Review of Psychology, 57, 401–421.
Open source →doi:10.1146/annurev.psych.57.102904.190127
7
Roberts, B. W. et al. (2007). The power of personality: the comparative validity of personality traits, socioeconomic status, and cognitive ability. Perspectives on Psychological Science, 2(4), 313–345.
Open source →doi:10.1111/j.1745-6916.2007.00047.x
8
Soto, C. J. (2019). How replicable are links between personality traits and consequential life outcomes? Psychological Science, 30(5), 711–727.
Open source →
9
McCrae, R. R. & Costa, P. T. (1989). Reinterpreting the Myers-Briggs Type Indicator from the perspective of the five-factor model of personality. Journal of Personality, 57(1), 17–40.
Open source →doi:10.1111/j.1467-6494.1989.tb00759.x

Cogniself · O.C.E.A.N.

Traceable measurement. Evidence-grounded interpretation.

Cogniself makes rigorous behavioral science practical by transforming standardized self-report data, personal context, and peer-reviewed research into nuanced, literature-linked reports for individuals and professionals.

Take the Big Five assessment