How Our Tests Are Built and Validated.

Every test is built on a model or instrument with a published research record: the Big Five and HEXACO models of personality, Raven-style progressive matrices and classical reasoning formats for cognitive ability, Holland's RIASEC model for career interests, the Schwartz theory of basic values, the WHO Quality of Life framework for well-being, and validated self-report scales for skills such as emotional intelligence and self-efficacy. Where strong public-domain item pools exist (for example the International Personality Item Pool), we use and credit them; where we author items, we follow the published construct definitions and classical item-writing standards.

Each test page lists its scientific references - the same literature we used to build it.

Your responses are scored with documented psychometric procedures: scale scores from keyed items (with reverse-coding where the instrument requires it), aggregation rules fixed in code, and conversion to normed standard scores. Our detailed reports are generated from your scored profile by strict, documented scoring rules - the same answers always produce the same report. Nothing in the scored results is improvised, and no score is ever adjusted to flatter.

A raw score means little by itself. Wherever a test reports standing (percentiles, IQ-metric scores, trait bands), it is computed against a documented norm table - population reference data appropriate to the instrument. Norms are versioned alongside test content, and as our own response datasets grow, norm tables are re-derived and re-documented rather than silently changed.

We hold instruments to the standard reliability bar for self-report measurement (internal consistency typically at or above .80 for primary scales) and prefer instruments with published validity evidence: factor structures that replicate, convergence with established measures, and demonstrated prediction of real outcomes. Where a short form trades precision for speed, the test page says so explicitly - a 10-item screen is honestly framed as a snapshot, never sold as a diagnostic-grade profile.

Self-report measures what you can observe about yourself; unsupervised online testing cannot certify extreme scores or replace professional assessment. That is why our IQ scores are capped at 160, why well-being check-ins are explicitly labeled screening rather than diagnosis, and why reports describe trade-offs of every trait position instead of declaring winners. If a result raises clinical concerns, the right next step is a qualified professional, and our reports say so.

Honest answers require trust. Results are private to you, shareable only by your explicit choice through expiring links, and response data used for norm development is anonymized and aggregated. We do not sell personal data.

Psychology.me assessments are built and reviewed by Dr. Miloš Kankaraš, PhD in psychometrics (Tilburg University), with 20+ years in international large-scale assessment - including co-leading the OECD Study on Social and Emotional Skills and measurement work for the European Union and UNESCO. His published research covers measurement equivalence, latent variable modeling, and cross-cultural assessment methodology (Google Scholar). He co-authored the OECD's conceptual and assessment frameworks for social and emotional skills and led their field-test validation - the research behind this platform.

Found something we should improve? Write to milos@psychology.me - methodological criticism is welcome here.