The legal position
The Equality Act framework includes direct and indirect discrimination, harassment and victimisation, with distinct legal tests. Protected characteristics include age, disability, gender reassignment, marriage and civil partnership, pregnancy and maternity, race, religion or belief, sex and sexual orientation. Indirect discrimination can arise from an apparently neutral rule that disadvantages a protected group and an individual, unless justified under the applicable test. Disability also brings specific protections and reasonable-adjustment duties.
Source and attribution [1]Using it in practice
Review how a decision is made and who it affects. Ask whether criteria are necessary for the actual work, whether alternatives exist and whether an adjustment is needed. Keep decision evidence and provide a way to challenge errors. Use appropriately protected equality information to examine patterns, while avoiding unsupported conclusions from very small groups. Seek current specialist guidance for sensitive or contested applications of the law.
An example, not a reported case
Worked example · illustrative
A promotion process requires attendance at late-evening networking sessions. HR asks whether attendance predicts the capability needed for the role and who may be excluded. It tests alternatives such as documented project contributions and structured assessment. The review does not assume that a rule is fair merely because the same words apply to everyone.
What to watch for
Protected characteristics do not all operate identically across every provision. Northern Ireland has a separate equality-law framework; this overview is not a substitute for case-specific advice.
Examine the decision and its effect
Identify the specific decision, rule or behaviour under review: recruitment shortlisting, access to overtime, promotion criteria or treatment after a complaint. Then ask who experiences disadvantage and how. Intention is only part of the picture. A seemingly uniform requirement can affect groups differently, while a single person's account can reveal a problem hidden by aggregate data. Use the appropriate legal test for the situation; fairness in everyday language and unlawful discrimination are related but not interchangeable concepts.
Audit a real decision pathway
An original practical exercise is to trace one vacancy from advertising through offer. Identify where criteria were introduced, whether assessors used the same evidence and whether candidates could access the process. Review rejected as well as successful applications. A criterion such as 'available at any time' deserves scrutiny against the actual job requirement. Record questions for specialist review rather than declaring a criterion lawful because it appears commercially convenient. Use anonymised examples when discussing improvements with managers.
Combine numbers with individual evidence
Suppose one group progresses less often through an assessment. Check sample size, missing data, role mix and the stage where the difference appears. A gap is a prompt to investigate, not on its own a complete explanation or legal finding. Conversely, an organisation-wide average does not disprove an individual's experience. Examine decision records and listen to affected people. Make sure any equality monitoring has an appropriate purpose, access arrangements and explanation; collecting more sensitive information is not automatically the best answer.
Improve the system and verify the result
Possible responses include clearer job-related criteria, accessible assessment options, calibrated decision making and a route to challenge mistakes. Assign ownership and review actual decisions after the change, rather than counting training attendance alone. Explain what changed without identifying complainants. Watch for retaliation or exclusion after someone raises a concern. This guide concerns the Equality Act framework in Great Britain; Northern Ireland uses a distinct framework and should be checked through its own official institutions.
Investigate the decision rule and its effects
A rule that appears consistent may have unequal consequences. Describe the actual requirement, who it affects and what disadvantage is alleged. Then examine its purpose, supporting evidence and possible alternatives, using the relevant legal test rather than a general assertion that everyone is treated the same. For example, a fixed training time may exclude some staff even when the invitation is universal. Keep the investigation focused on the work and the specific facts, not assumptions about a protected group.
Separate a monitoring signal from a legal conclusion
A difference in selection or promotion rates may justify investigation but does not, by itself, prove or disprove unlawful discrimination. Check the process, criteria, opportunity to apply, evidence used and the explanation for decisions. Small samples and incomplete data require care. Record what can be concluded and what remains uncertain, and seek appropriate advice for an actual complaint. Good practice includes correcting barriers before they become disputes; statistical reassurance should not be used to dismiss an individual’s evidence.
Take it into your next conversation
Three useful questions.
- What does this criterion actually measure?
- Who may be disadvantaged?
- Is a less exclusionary approach available?
Related terms
Go to the evidence
Sources & attribution
[1] Acas. Discrimination and the Equality Act 2010. ↗
Official guidance supports the legal overview. Application notes and examples are original editorial material, not individual legal advice. Check the linked source for changes and the full qualifications.
Application guidance expanded on 26 September 2026. These practical suggestions and invented examples are editorial explanations; further-reading sources are not the original authors of the historical models unless expressly identified. Original overview review dates are retained. Older books inform practice discussion only; use current official sources for legal requirements and commencement dates.
Published 2026-09-20 · Reviewed 2026-09-22. Editorial approach