Skip to main content

Compliance

Bias auditing: our current position

Last updated 23 August 2026

This page replaces an earlier "Bias Audit Report". That report described an audit that had not been carried out, and published selection ratios calculated from candidate demographic data that we do not collect and have never held. We withdrew it on 23 August 2026. If you relied on it, please read what follows and contact us.

No independent bias audit has been carried out

We have not commissioned or completed an independent bias audit of our assessments. No auditor has examined them, and we hold no audit report, selection-rate table or impact ratio for any protected class.

We are stating this plainly because the alternative — leaving a favourable-looking report in place — is worse for the employers relying on us than having nothing at all.

What this means if you hire in New York City

NYC Local Law 144 requires an employer using an automated employment decision tool to have an independent bias audit performed and to publish a summary of its results. That obligation sits with you as the employer, and our page cannot discharge it. If you have been pointing at this URL to meet it, you do not currently have an audit behind that reference.

We would rather you knew that now than found out during a complaint. If it would help to talk it through, email [email protected].

What candidate data we actually hold

We do not collect, and have never collected, any of the following from candidates:

  • Race or ethnicity
  • Gender
  • Age or age band
  • Disability status
  • Any other protected characteristic

No form on this site asks for any of it and no field in our database can store it. What we hold about a candidate is their name, their email address, the answers and edits they submitted, the time they took, and the score those produced.

One consequence follows directly: because we hold no demographic data, we cannot compute selection rates by protected class, and neither can anyone else from our records. Any adverse-impact analysis of your hiring has to be done by you, against demographic data you hold, using the pass/fail outcomes we give you.

What we can tell you about our assessments

These statements are true and we can support them:

  • Our tests measure editorial skills — grammar, editing, proofreading, writing, industry vocabulary and word processing — against a marked answer key or a master-corrected text.
  • Scoring does not use any candidate attribute other than the work they submitted. There is no demographic input to a score, because there is no demographic data.
  • The pass mark is a setting, not a judgement about a person. You choose where to set it, and you should choose it on the basis of the job.
  • Open-response work — writing, editing, proofreading and MS Word document tasks — is reviewed by a human before the result is released to you.

What we intend to do

Whether to commission a real bias audit is a decision we have not yet taken. A genuine audit needs demographic data that we do not currently collect, and collecting it is not a small step: race and disability are special-category data under GDPR Article 9 and need explicit consent, a lawful basis and a data protection impact assessment before a single field appears on a form. We will not publish anything on this page again until there is an audit, performed independently, that we can produce on request.

Questions about this page

Email [email protected]. If you are a candidate who wants to know how your test was scored, our AI notice sets out the scoring process and your rights.

Editors from these organisations have used our services since 1998

Reuters BBC Oxford University Press Penguin Random House Springer Microsoft Suncor Energy United Nations Fisher Investments IBM The Home Depot KODAK CHEVRON