Editorial policy
How we decide what is true enough to publish.
This page is the standard every guide and calculator on Statohub is held to: where the maths is checked, what a review has to catch, how errors are corrected, and where AI tools are and are not used.
Last updated: September 20, 2026
Who is accountable
Everything on Statohub is published by the Statohub Editorial Team, the group that commissions, writes, reviews and maintains the site. We publish under a team byline rather than individual names because every page passes through more than one pair of hands before it goes live, and the team as a whole answers for it. Who we are and why the site exists is on the About page. If something on a page is wrong, the address to write to is on the contact page.
What we publish, and for whom
Statohub explains statistical methods and gives you a calculator to run them. It is written for students, self-learners and working analysts who need to understand a method well enough to use it and to know when it will mislead them. Every guide is expected to cover three things: what the method is, one example worked all the way through in real units, and the situations in which the method breaks down.
We do not publish gambling or betting-odds tools, invented evidence, or anything that stands in for medical, legal or financial advice. The About page explains each of those rules.
Where the maths is checked
Formulas, definitions and worked examples are checked against primary references rather than against other websites. The references we return to most often, and cite from the guides themselves, are:
- the NIST/SEMATECH e-Handbook of Statistical Methods for definitions, formulas and reference values;
- Penn State's online course notes, in particular STAT 501 (Regression Methods), for regression and correlation;
- OpenStax Introductory Statistics for the introductory material.
When a guide gives a number, that number is computed, not quoted. Worked examples are recalculated by hand and, wherever the page has a calculator, cross-checked against the calculator's own engine so that the prose and the tool cannot disagree. Where a guide uses data to demonstrate a method, it is presented as a worked scenario and never dressed up as a real-world finding.
How the calculators are verified
Each calculator is a small, self-contained function maintained in this site's own codebase, not a third-party widget. Every function is covered by an automated test suite that has to pass before the site can be deployed. The tests compare results with values worked out independently and deliberately include the awkward inputs that produce quietly wrong answers elsewhere: empty lists, a single value, ties, negative numbers, and the sample-versus-population distinction.
The calculators run entirely in your browser and nothing you type is sent to us. Results are shown with the intermediate steps so you can follow the arithmetic rather than take the final number on trust. A calculator is still a calculator: check that the method fits your data before you rely on its output.
What a review has to catch
Every draft is reviewed against a fixed written checklist before it is published. A draft that fails comes back with a specific list of fixes, not a general note. The checklist covers, among other things:
- the arithmetic in every worked example, recomputed;
- whether the page answers its own question and stays on one topic, so it does not duplicate another page on the site;
- whether the failure modes and edge cases a reader will actually hit are covered, not just the happy path;
- that no study is cited that has not been read, and no statistic is quoted that cannot be sourced;
- that every internal link resolves to a live page (the site will not build otherwise);
- that the frequently-asked-questions block has one entry per genuinely distinct question, with answers that add something rather than restate it.
How we use AI
We use AI tools for research and for proofreading only: to find and summarise source material, and to catch typos, unclear sentences and inconsistencies in a finished draft. The guides are written by the editorial team, every worked example is recomputed by the team, and nothing is published without a team review. AI output is never used as a source of facts or figures on its own.
Commissioned guides
Some guides in the Applied section are commissioned from outside subject-matter writers. They are not published as they arrive: each one is edited for house style, checked for accuracy and for overlap with existing coverage, and only then published under the same byline and to the same standard as the guides written in-house.
Advertising
Statohub is funded by advertising served by Google AdSense, which is what keeps every guide and calculator free. Ads are selected by Google, not by us, and advertisers do not see, commission or influence any editorial content. How advertising cookies work here is set out in the privacy and cookie policy.
Dates, updates and corrections
Every guide and calculator page shows two dates: when it was first published, and when it was last reviewed. The review date moves when the page is edited or when it has been checked again without needing changes. A page is revisited when we find an error, when an explanation is not landing, or when a better example replaces a generic one.
If you find a mistake, tell us. Corrections to the maths are treated as urgent. When something substantive is corrected, the page is fixed, its review date changes, and we would rather hear about it from you than leave it up. The contact page says what to include so we can find and fix the problem quickly.