TERM

What is reliability in testing?

Reliability is how consistent a test score is: would the same person score about the same again? See why it matters in hiring and how it differs from validity.

The short answer

Reliability is how consistent a test's results are. A reliable test gives a similar score when the same person takes it again under similar conditions or takes another version of it, and its items agree about what they measure. Reliability is necessary for a useful test, but it does not show that the test measures what the job needs.

Reliability vs validity

A bathroom scale that always reads two kilograms heavy is reliable but wrong. Tests work the same way. Reliability asks whether the score is consistent; validity asks whether the score supports the decision you are making. A test can be highly consistent and still measure something irrelevant to the job, so you need evidence for both.

Common forms of reliability

Test developers usually look at three kinds of consistency. Test-retest reliability compares scores from the same people on two occasions. Parallel-forms reliability compares different versions of a test, which matters when questions come from an item pool. Internal consistency checks whether the items in a test agree with each other. Interviews have their own version, inter-rater reliability: whether two interviewers give the same answer the same score.

A hypothetical example

Hypothetical example: a hiring manager and a team lead both interview the same five candidates without a scoring guide, and their ratings disagree on three of the five. Before the next round they write a structured interview guide with rating anchors for each question. Their ratings now mostly match, so the interview score is more consistent and can carry more weight in the decision.

How employers protect reliability

You rarely calculate reliability yourself, but you can protect it. Give every candidate the same conditions, time and instructions, use tests with comparable versions, score interviews against written anchors, and treat small score differences as ties. When choosing a test provider, ask what consistency evidence exists for the test and for which group it was checked.

Common mistakes

  • Assuming reliable means fair or relevant. Consistency is only the first question.
  • Changing conditions between candidates. Different time limits or settings reduce consistency.
  • Overreading tiny differences. Every score carries some measurement error.

See also raw score, percentile and norm group, or browse the full hiring glossary.

About the editorial team

Written and fact-checked by the HireValid editorial team. Sources checked . Read the editorial policy.

LESS GUESSWORK. MORE GOOD PEOPLE.

AI can polish an answer.
Hire for the ability behind it.

Explore advanced hiring assessments and integrity tools built for the AI era.

Start freeOr try a sample test No credit card. No annual commitment.