Home

The reference desk

Data, methods and sources.

The references behind each test, and how your data is used.

Data use

Statistics are optional. After you save the Survey, reports opened afterward may submit an opaque attempt ID, test and form version, numeric results—including mental-health screening scores—and the broad Survey categories you choose. Height and weight are converted to broad ranges before sharing; the browser also derives a broad BMI category for the BMI Statistics filter. Exact measurements never leave the browser. OpenExaminations uses those fields for aggregate Statistics, selected-nationality country summaries, and a separate exploratory Insights area. Insights test whether a broad Survey field predicts test results on held-out attempts. Association percentages show the share of result variation between Survey categories (η²); cards are selected by held-out prediction gain. The bar shows η² on a 0–100% scale. The strength label is deliberately more conservative: moderate requires η² of at least 6% and a reduction of at least 10% in held-out squared prediction error; strong requires at least 14% and 25%, respectively. The η² cutoffs follow conventional effect-size benchmarks; the prediction-gain cutoffs are OpenExaminations's descriptive choices, not universal standards or significance tests. These exploratory patterns are not validated personal forecasts or causal, clinical, or population-norm conclusions.

Complete reports, item answers, and the full Survey form stay in this browser; the latest attempt for each test is retained. No name, email, account, detailed report, or item answer is sent with Statistics. Older health-screen reports created under the browser-only policy remain private unless you explicitly enable comparisons on that report; newly completed screens can share only their numeric total after Survey setup. These site samples are not clinical reference values or diagnostic norms. For the active perception screen checks, an opted-in submission contains only the reached task threshold, not trial answers or testing conditions. Numeric Statistics rows are retained for up to 365 days. Written responses on tests with AI review are sent automatically after submission, along with their prompt or task material, to the configured AI review service.

Advertising

The library page loads Google's AdSense script for site review. This can contact Google even before an ad appears. OpenExaminations has not placed manual ad units; if Auto ads are enabled in AdSense, ads may appear here after approval. Third-party vendors including Google may use cookies to serve ads based on visits to this site or other websites. Google's advertising cookies let Google and its partners show ads based on those visits. You can control personalised ads in Google's My Ad Center. Regional consent controls and a review of actual placements are needed before enabling ad serving. Advertising code is not loaded on test, challenge, survey, or personal-report pages.

Exam material and evidence

Every SAT, GRE, ACT, MCAT, LSAT, GMAT, TOEFL, IELTS, AP, CLEP, GED, and Duolingo item has a stable ID and an origin record. The registry links domain-level first-party evidence used for item-writing and scoring review; it does not republish secure exam content or provide an exam license. Local reports show raw scored-item accuracy plus the applicable published-style writing or content result; any display estimate is explicitly local and never an official conversion.

Evidence references include NAEP released questions and scoring data; first-edition Illustrative Mathematics content; Oak National Academy Collection 2; PIAAC and PISA released proficiency and coding material; and permission-gated IEA documentation. Rights are source-specific: check the linked item, PDF, stimulus, and third-party notices before reusing anything.

Exam evidence and reuse terms

Personality references

Cognitive references

Map reference

Personality material

Big Five and Vocational Interests use IPIP items and keys, which IPIP places in the public domain. Dark Tetrad uses the author-distributed SD4 instrument, also listed as public domain. These are self-report profiles for reflection, not diagnoses, hiring tools, or admissions decisions.

Adapted cognitive material

Cognitive Ability adapts five TestMyBrain tasks and reference distributions. The code is credited in cognitive-test.js to The Many Brains Project and McLean Hospital; code is LGPLv3 and stimuli/reference materials are CC BY-SA 4.0 unless a task says otherwise. The form supports ages 18–89. Its N=1,000 Analysis Toolkit files are convenience reference samples, not representative population norms; no quality-control exclusions were applied. The profile is not an IQ score, clinical assessment, or validated full-scale composite.

Retired and named instruments

The Self-Compassion Scale was previously available under researcher permission. Goal Persistence and Purpose Exploration were original OpenExaminations reflections. All three are retired; compatibility metadata remains only so existing reports can be read, and retired routes are not part of the current catalog. Jungian Personality is an original, MBTI-independent questionnaire. The Duckworth Grit Scale and Meaning in Life Questionnaire are not current OpenExaminations tests.

AI writing feedback

Written responses on supported tests are reviewed automatically after submission. Every displayed rubric criterion is numeric and participates in the displayed result under the named scoring rule: sums for point rubrics, means and published weighting where applicable, rounded factor means for GRE and TOEFL practice, and clearly labeled local means for LSAT and DET where no official numeric writing result is published. The configured review service has no web access, does not independently verify facts, and can be wrong. Model and rubric versions are saved with a review; no result is an official grade or a calibrated human score.