Psychometric testing is now a normal part of hiring at most large employers and a growing number of small ones. The question for anyone considering it is simple: does it work? This article sets out what the research says, where it is strong, where it is contested and what you have to do for the evidence to apply to your own hiring.
Should you use psychometric testing to hire?
Psychometric tests are standardised measures of ability, skill or working style. Ability tests (numerical, verbal, logical and other reasoning) have right and wrong answers; personality and interest measures describe preferences. Employers use them because a CV and an unstructured interview are poor at telling candidates apart, and because a short test gives every applicant the same chance to show what they can do.
Whether you should use one depends on the role and the volume of applicants, which the last section covers. Whether they predict performance at all is a research question with a long answer.
Can psychometric testing predict candidate performance?
Yes, with care about what is being predicted and how strongly.
The history of psychometric testing and prediction
The first systematic attempts to measure individual differences date from the late nineteenth century, with Francis Galton's work on measurement in the 1880s. Ability testing was developed at scale in the two world wars for selecting and placing personnel, and spread into civil service and industry from there. Since then there have been thousands of studies relating test scores to later job performance.
What the meta-analyses found
The most cited summary is Frank Schmidt and John Hunter's 1998 review of 85 years of research. It found that general mental ability tests were among the strongest single predictors of job performance across occupations, alongside work sample tests and structured interviews, and that combining an ability test with a second measure such as an integrity test or a structured interview predicted better than either alone. Unstructured interviews, years of experience and education predicted much less.
In 2022 Paul Sackett and colleagues re-analysed the same body of evidence with different statistical corrections and arrived at lower estimates for most methods, including ability tests. Structured interviews and job knowledge tests moved to the top of their ranking, with ability tests still well above experience, education and unstructured interviews. The debate about the exact numbers continues; the ordering of what works has been stable for decades.
What this means in practice
Three things follow from the research.
- A relevant test plus a structured interview is better than either alone, and far better than an unstructured interview.
- Prediction is strongest when the test matches the work: numerical reasoning for roles that interpret figures, a software test for roles that live in that software, a situational judgement test for roles built on everyday decisions.
- No test predicts perfectly. A test narrows the field and tells you where to look in the interview; it does not tell you who to hire.
How to shortlist candidates quickly
For a vacancy with many applicants, the research-backed process is short:
- Sift CVs only for essential requirements such as right to work or a licence.
- Invite everyone who passes to a short assessment of two to four tests chosen for the role, before the interview.
- Interview the strongest group with a structured set of questions, using the test results to decide what to probe.
- Add a work sample for the final two or three.
On Test Candidates each candidate receives the same tests, results are scored as soon as they finish and you see them side by side. Our cognitive ability, numerical, verbal and logical reasoning tests cover the ability measures in the research; situational judgement, skills and software tests cover the job-related ones.
Is there a proven link between psychometric tests and hiring the best candidate?
There is a well-established link between relevant, well-constructed tests and later job performance, strong enough that most selection researchers recommend them as part of a structured process. There is no evidence that any test on its own picks the best candidate, and the strength of the link depends on choosing tests that match the work and using them alongside a structured interview and a work sample. Used that way, testing is one of the few hiring practices with decades of evidence behind it.
