Psychometrics · 16 min read

Assessments in Hiring: The Complete Guide

Everything about pre-employment assessments: the main types, what they predict, which combination fits each role and how to use them fairly and lawfully.

Door Ingmar van Maurik · Founder & CEO, Making Moves

Short answer

A pre-employment assessment is a standardized way to measure whether a candidate fits what a role requires: reasoning, behavior, skills, personality or integrity. Used well, assessments predict job performance better than a CV or an open conversation. The key is a combination of methods built on a clear job analysis, using demonstrably valid and reliable instruments in a fair, transparent process.

In this guide you will learn which types of assessments exist, what they predict, which combination works for each role, where they belong in your process and how to use them fairly and lawfully. Each topic links to an in-depth article.

What is an assessment in hiring?

An assessment is any structured measurement instrument that you administer and score the same way for every candidate. That can be an online cognitive ability test, a personality questionnaire, a situational judgment test, a work sample or a structured interview with a fixed scoring method.

The difference from a regular job interview lies in standardization: the same tasks, the same evaluation criteria and a comparison with a norm group. That lets you compare candidates fairly and check afterwards whether your predictions were right.

This guide covers selection assessments: instruments used to decide whom to hire. Development assessments for current employees follow the same principles but serve a different purpose. If you first want to understand the basics of measurement itself, read what psychometric testing is.

Why use assessments? Benefits and pitfalls

The benefits

  • Better predictions. Structured methods predict job performance demonstrably better than an unstructured conversation or reading a CV.
  • Fairer comparison. Every candidate gets the same tasks and is judged on the same criteria. That leaves less room for gut feeling and unconscious preferences.
  • Faster at volume. A short online pre-assessment can objectively pre-sort hundreds of candidates, so recruiters spend their time on the most promising people. See also [hiring at high volume](/en/articles/high-volume-recruitment).
  • Data to learn from. Standardized scores can later be linked to performance and turnover. That makes your selection better every year.

    • The pitfalls

  • False precision. A score looks exact, but every test contains measurement error. Small differences between candidates often mean nothing.
  • Drop-off. Long or poorly explained assessments cost you candidates, especially in tight labor markets.
  • Unequal outcomes. Some tests produce lower average scores for certain groups. Without monitoring you unintentionally exclude talent.
  • The wrong instrument. A generic test that does not match the role mostly measures noise. Read why [generic assessments often don't work](/en/articles/generic-assessments-dont-work).

    • The types of assessments at a glance

    TypeWhat it measuresAvg. validity (r)Duration (indicative)

    |---|---|---|---|

    Structured interviewBehavior and experience via fixed questions and scoring0.4245 to 60 min Job knowledge testKnowledge the role directly requires0.4020 to 45 min Work sampleA real task from the job0.3330 to 120 min Cognitive ability testVerbal, numerical and abstract reasoning0.3120 to 45 min Integrity testAttitude toward rules and dependability0.3115 to 30 min Situational judgment testJudgment in realistic work situations0.2620 to 40 min Personality questionnaireStable traits and work style0.19 (conscientiousness)15 to 30 min Assessment centerMix of simulations, interviews and testsDepends on designHalf to full day

    The validity figures come from the large meta-analysis by Sackett, Zhang, Berry and Lievens (2022) and are averages across many studies. How to read such figures and when a test is good enough is covered in what makes an assessment valid and reliable.

    Cognitive ability test

    Measures how quickly and accurately someone processes new information and solves problems. Strong for roles with high complexity and a lot of learning. Watch for possible group differences and always combine with other information. More in cognitive ability tests in hiring.

    Personality questionnaire

    Maps traits such as conscientiousness, stress tolerance and cooperativeness, usually based on the Big Five model. On its own a personality test predicts modestly; as a complement to a cognitive test or interview it does add value. Never use it as a standalone knock-out.

    Situational judgment test (SJT)

    Presents candidates with realistic situations from the role and asks which response is best. Easy to tailor, pleasant for candidates because they see what the work is really about, and well suited as a first filter at volume. More in situational judgment tests in hiring.

    Integrity test

    Predicts counterproductive behavior such as rule-breaking, theft and absenteeism. Sometimes sold as an employee reliability test. Especially valuable in roles responsible for money, goods or safety. Be transparent with candidates about its purpose. More in integrity and employee reliability tests.

    Job knowledge test and work sample

    Test directly what someone needs to be able to do: a coding assignment, a customer case, a numerical exercise or a simulated customer call. Highly predictive and credible for candidates, but less suitable for entry-level applicants without experience.

    Structured interview

    A conversation can be an assessment too, provided the questions are fixed, focused on the role and scored with a rubric. It then ranks among the strongest predictors available.

    Assessment center

    A combination of simulations (role play, presentation, in-tray exercise), interviews and tests, rated by trained assessors. Rich in information but expensive and time-consuming; most useful for senior or strategic roles.

    AI-supported and game-based assessments

    Newer formats use AI to analyze answers, or measure skills through short game tasks. They can improve the experience and scale faster, but require extra care: how was the model validated, is it explainable and has it been tested for unequal outcomes? Read more about AI pre-interviews and how to combine AI and psychometrics.

    Which combination for which role?

    No single method predicts everything. Good selection combines two to four components that together cover what drives success in the role. Examples as a starting point:

    RoleLogical combinationWhy

    |---|---|---|

    Customer service (volume)Short SJT, personality, structured interviewService judgment and stress tolerance, quick to screen SalesSJT, personality, role playPersuasiveness and resilience are best seen in action Security and supervisionIntegrity test, attention and concentration, SJTDependability and alertness are critical Software developerWork sample, cognitive test, structured interviewShows real skill and predicts learning ability ManagerCognitive test, personality, structured interview, optional simulationComplex decisions and behavior toward teams Graduate or entry levelCognitive test, personalityLearning ability and motivation outweigh experience

    These are starting points, not recipes. Which components predict best for your roles can only be shown with your own data. How to design a tailored battery is covered in how to design company-specific assessments.

    Where do assessments belong in the process?

  • Early and short at volume. A pre-assessment of fifteen to thirty minutes right after the application pre-sorts objectively. Use components that are quick and cause little drop-off, such as a short SJT.
  • Later and deeper for promising candidates. More extensive tests, work samples or an assessment center make sense once the group is small and the investment is worthwhile.
  • Never as the only decision. An assessment is one source alongside the CV, interview and references. Decide in advance how much each component weighs.
  • Design the candidate experience. Explain upfront what is measured and why, make it work on mobile and give feedback. Measure drop-off per step and cut what adds nothing. See also [a candidate journey that converts](/en/articles/candidate-journey-converts).

    • How this fits into a complete screening process is covered in the best candidate screening process.

    Quality: valid, reliable and fair

    Three questions decide whether an assessment is any good:

  • Is it valid? Does it demonstrably predict performance in comparable roles?
  • Is it reliable? Does it produce consistent scores? For selection decisions, 0.80 or higher is the usual standard.
  • Is it fair? Have unequal outcomes between groups been examined and can they be explained by the job?

    • A checklist of eight questions for your vendor is in what makes an assessment valid and reliable. Compare scores preferably with your own successful employees: read how to build your own norm group.

    Fairness, privacy and regulation

  • GDPR (EU and UK). Collect only data needed for the role, inform candidates in advance and do not keep results longer than necessary.
  • No fully automated rejection. Under Article 22 GDPR, candidates may not be rejected solely on the basis of automated processing where this significantly affects them. Keep a human in the decision.
  • EU AI Act. AI systems for recruitment and selection fall into the high-risk category of the EU AI Act, with requirements on data quality, documentation, transparency and human oversight. The obligations are being phased in, so check the current timeline.
  • United States. Under the Uniform Guidelines, adverse impact is commonly checked with the four-fifths rule: a selection rate for any group below 80% of the highest group's rate warrants scrutiny. Provide reasonable accommodations under the ADA.
  • Equal opportunity. Monitor outcomes per group and offer accommodations where needed, such as extra time for dyslexia. More in [how AI can reduce bias in hiring](/en/articles/ai-reduce-hiring-bias).

    • For large-scale or AI-driven assessment of candidates in the EU, carry out a data protection impact assessment (DPIA) before you start.

    What does an assessment cost, and what does it return?

    Vendor online tests are usually licensed per candidate or per year. An assessment center with assessors costs considerably more per candidate. With your own system you shift costs from a per-candidate fee to a fixed investment, and the data stays yours.

    The return comes mainly from avoided mis-hires. A wrong hire easily costs a multiple of an annual salary in recruitment, onboarding, lost productivity and replacement; see what a bad hire really costs. How assessments also lower cost per hire is covered in reducing cost per hire with automation and assessments.

    Buy, tailor or build?

    OptionSuitable ifWatch out for

    |---|---|---|

    Off-the-shelf vendor testsYou want to start quickly with common rolesGeneric norm group, per-candidate fees, data with the vendor Tailored assessmentYour roles are specific or the standard does not predictRequires a job analysis and validation research Your own assessment systemYou hire a lot and want to keep learning from your dataUpfront investment, but full ownership and ongoing calibration

    A detailed comparison is in build vs buy: should you create your own hiring platform.

    Step-by-step plan: introducing assessments in 7 steps

    1. Define success per role

    Do a job analysis: which tasks, situations and behaviors make someone successful? Also agree how you will measure success later, for example reviews after 6 and 12 months and turnover.

    2. Choose a suitable combination

    Select two to four methods that together cover what the analysis shows. Start small: a well-founded SJT and a structured interview beat five unrelated tests.

    3. Check the quality

    Request validation research, reliability per scale and information about the norm group. If in doubt, pilot with current employees first.

    4. Design the candidate experience

    Keep the total duration limited, especially early in the process. Explain why you measure, make sure it works on mobile and give candidates feedback.

    5. Set decision rules

    Define the weighting per component in advance and work with score bands instead of hard single-point cut-offs. Agree who decides and how a recruiter can deviate, with a rationale.

    6. Arrange privacy and fairness

    Inform candidates, set retention periods, carry out a DPIA where needed and set up monitoring for unequal outcomes.

    7. Measure and improve

    Link scores to performance and turnover, see which components predict and adjust the weighting. That way your selection keeps getting smarter. How to measure potential more precisely is covered in measuring candidate potential accurately.

    Frequently asked questions

    What is a pre-employment assessment?

    A standardized test or task an employer uses to measure whether you fit what the role requires, such as a cognitive ability test, personality questionnaire, situational judgment test or work sample. All candidates take the same component and are judged on the same criteria.

    How long does an assessment take?

    An online pre-assessment usually takes fifteen minutes to an hour. A full assessment center can take half a day to a full day. The earlier in the process, the shorter the assessment should be.

    Can an employer require an assessment?

    An employer can make an assessment part of the process, provided it is job-related, candidates are properly informed in advance and data protection law is respected. Candidates have a right to know what happens to their data.

    What is the difference between an assessment and a cognitive ability test?

    A cognitive ability test is one type of assessment: it measures reasoning. "Assessment" is the umbrella term for all structured measurement instruments, and often for a combination of them.

    How much does an assessment cost?

    Online tests are usually licensed per candidate or per year; an assessment center with assessors is more expensive per candidate. With your own system you invest upfront and per-candidate fees disappear.

    Can you use AI in assessments?

    Yes, for example to structure open answers or run a pre-interview. AI in hiring falls under the high-risk rules of the EU AI Act, and a human must make the decision. Read how AI candidate scoring works.

    How do you keep assessments from discriminating?

    Choose job-related instruments, check outcomes per group, combine methods so no single test decides, offer accommodations where needed and keep a human in the decision.

    Summary

  • An assessment is a standardized measurement instrument that you administer and score equally for all candidates.
  • Combine two to four methods that fit a clear job analysis; no single test predicts everything.
  • Choose instruments that are demonstrably valid, reliable and fair.
  • Use assessments early and short at volume, and in more depth later in the process.
  • Arrange data protection, human oversight and monitoring of unequal outcomes.
  • Measure and improve: link scores to performance so your selection predicts better every year.

    • Want assessments developed for your roles, or a system that learns from your own data? See how we develop custom assessments, see how our AI hiring system builds in assessments, or get in touch.

    Book intake call · View our AI Hiring System