Psychometric Test

The two main types of psychometric test

Aptitude and ability tests have right and wrong answers and measure maximum performance under time limits. Common formats include numerical reasoning (reading tables and charts), verbal reasoning (drawing conclusions from a passage), abstract or logical reasoning (pattern sequences) and mechanical reasoning. Together these are often grouped as a cognitive ability test.

Personality and behavioral questionnaires have no right answers and measure typical behavior. Candidates rate statements like "I prefer to finish one task before starting another." Most workplace instruments are built around the Big Five traits: conscientiousness, agreeableness, extraversion, emotional stability and openness. A personality test of this kind describes tendencies, not abilities.

A third, smaller group covers motivation, values and interests, which are more often used in development and career planning than in hiring decisions.

How psychometric scoring works

A raw score means little on its own. Psychometric tests compare each candidate to a norm group, such as graduate applicants or customer service employees, and report a percentile or a standardized score.

Example: a candidate answers 22 of 30 numerical reasoning questions correctly. Against a norm group of general working adults that might be the 70th percentile. Against a norm group of finance graduates, the same 22 might be the 40th percentile. Choosing the wrong norm group can make a test look harder or easier than it is for your applicant pool.

Two properties decide whether a test is worth using. Reliability is consistency: the same person should get a similar score next month. Validity is relevance: scores should relate to how people actually perform in the job. Ask any provider for technical documentation covering both, including results for roles similar to yours.

When psychometric testing helps and when it does not

Psychometric tests earn their place when you have more qualified applicants than you can interview and need a consistent way to compare them, or when the job depends on reasoning and learning speed that a resume cannot show. Graduate programs, management trainee schemes, analyst roles and technical support are typical fits.

They add less in very high-volume hourly hiring. A 30-minute reasoning battery before a warehouse interview usually costs more applicants than it saves in bad hires. A short job-specific skills or simulation test is often the better choice there.

Personality questionnaires are best used to shape interview questions, not to make cuts on their own. If a candidate scores low on conscientiousness, probe for examples of meeting deadlines rather than rejecting them on the number.

A worked example

A regional insurer recruits 30 claims trainees a year from about 1,800 applicants. Hiring manager Laura Kim used to read every resume, and results were uneven.

The team adds a 25-minute online test: verbal reasoning and numerical reasoning, plus a short work-style questionnaire. Candidates above the 40th percentile on the combined reasoning score against a norm group of entry-level professional applicants go forward to a structured interview. The personality profile is shared with interviewers only as prompts for follow-up questions.

About 600 applicants move on, which the recruiters can actually screen properly. After two cohorts, trainees selected this way pass the claims licensing exam at a higher rate than earlier groups. The team also checks pass rates by group each cycle under the four-fifths rule and lowers the threshold from the 50th to the 40th percentile after the first review shows an adverse impact ratio close to the line, with no measurable drop in exam results.

Legal and fairness considerations

In the US, psychometric tests fall under the same rules as any selection procedure. Title VII requires that tests with adverse impact be job related and consistent with business necessity, and the Uniform Guidelines on Employee Selection Procedures (1978) set out how to monitor it. Cognitive tests in particular tend to show average score differences between racial and ethnic groups, so validate them for the role and monitor pass rates closely.

The ADA bars medical examinations before a conditional job offer. Tests designed to identify mental health conditions can count as medical exams. In Karraker v. Rent-A-Center (2005), a federal appeals court found that using the MMPI, a clinical personality inventory, in promotion decisions violated the ADA. Use instruments built for workplace selection, not clinical diagnosis.

Outside the US, rules differ. In the UK, for example, test use is shaped by the Equality Act 2010, and the EU AI Act treats AI systems used to evaluate candidates as high-risk. Check local requirements before rolling tests out across countries.

Common mistakes

• Using personality scores as a hard cut. Faking is easy and validity for single traits is modest. Use them to guide interviews.
• Wrong norm group. Comparing frontline applicants to a graduate norm group screens out people who would do the job well.
• Too long for the role. Batteries over 30 minutes early in the funnel drive off candidates with other options.
• No accommodation route. Timed tests disadvantage some candidates with disabilities. Tell everyone how to request extra time or an alternate format.
• Clinical tools in hiring. Never use instruments designed to diagnose psychological conditions.
• Ignoring practice effects. Candidates who retake the same test often score higher. Use parallel forms or a retest waiting period.