How to Pass a Cognitive or Aptitude Test

You apply to a graduate programme. A few days later, you get an email with a link and a deadline: complete the online assessment within 5 days. Maybe 30 minutes. Maybe 90. The email does not always tell you what kind of test it is.

Most NL graduate programmes (banking, Big 4, ORMIT-tier traineeships, large corporates, consulting) include a stage like this. The test you face was almost certainly built by one of four big providers: SHL, cut-e (now part of Aon), Cubiks (now Talogy), or Pymetrics (now part of Harver). Different providers test the same things in different ways.

This guide does three things. It explains what "online assessment" actually means, names the providers and matches them to common NL employers, and gives you a practice plan that fits what the research actually shows about how much practice helps.

TL;DR

Four big providers cover most NL graduate assessments: SHL (the most common), cut-e/Aon, Cubiks/Talogy, and Pymetrics/Harver.

The test usually measures one or more of: numerical reasoning, verbal reasoning, logical or inductive reasoning, situational judgement, personality, or game-based behavioural traits.

Practice helps, but mainly on the first one or two attempts. After that, returns drop sharply. Spending 50 hours on practice tests is not a good use of your time.

Updated validity research (Sackett et al. 2022, Sackett et al. 2024) finds cognitive ability tests still predict job performance, but less strongly than the famous 1998 estimate. They are one input among several, not a single number that decides your career.

“Online assessment” is a category, not a specific test. The link in your email could open any of these:

Cognitive ability tests (also called aptitude tests)

Time-pressured tests that measure how you reason with numbers, words, or patterns. Numerical, verbal, and logical or inductive reasoning are the three main flavours. Most graduate-level cognitive tests run 12 to 30 minutes per section, often with multiple sections in one sitting. These are the most common test type at NL graduate programmes.

Personality questionnaires

Untimed (or loosely timed) questionnaires that ask how you usually behave at work. Common formats include forced-choice items (“which of these is most like me”) and Likert scales (1 to 5 agreement). There are no right answers in the cognitive-test sense, but there are answers that match what an employer is screening for.

Situational judgement tests (SJTs)

You read a short workplace scenario and choose the best response (or rank several responses). They sit between cognitive and personality tests. They are typically untimed but have a recommended duration of 20 to 40 minutes. Common at consulting firms and graduate programmes.

Game-based assessments

Short interactive games, usually 2 to 3 minutes each, designed to measure cognitive and behavioural traits without traditional test questions. Pymetrics is the best-known provider. Used heavily by banking, consumer goods, and consulting firms hiring at scale.

How to tell what you are facing

The invite email usually contains clues. Look for the platform name in the URL or the sender domain (shl.com, aon.com or cut-e.com, talogy.com or cubiksonline.com, harver.com or pymetrics.com). Look for the time estimate (12 minutes signals a Cubiks Logiks test; 25-35 minutes with games signals Pymetrics; 30 to 60 minutes signals SHL Verify).

If you cannot tell, reply and ask. “Could you confirm which assessment provider this is, and roughly how long the test takes?” That is a normal, professional question. Recruiters expect it.

The three test families at a glance

Cognitive ability tests measure reasoning under time pressure: numerical, verbal, or logical. A typical question shows a chart or passage and asks you to answer in 30-90 seconds, with sections capped at 12-25 minutes. Speed and accuracy both count, and yes, there are right answers. Practice genuinely helps, especially on the first one or two attempts. In the Netherlands, these are standard at Big 4 firms, banking, ORMIT-tier programmes, large corporates, and consulting.

Personality tests and Situational Judgement Tests (SJTs) measure work-style traits and your judgement in workplace scenarios. A typical question asks you to rate a statement or rank possible workplace responses. Time pressure is low, usually untimed or loosely timed. Personality tests have no right answers; SJTs do, though the right answer is not always obvious. Practice on personality is pointless. Practice on SJTs helps a little, mainly through familiarity. Most NL graduate programmes pair one of these alongside a cognitive test.

Game-based assessments measure cognitive and behavioural traits like risk tolerance, attention, and learning. A typical task is to inflate balloons, sort shapes, or react to colours. Each game has a 2-3 minute timer, but your pace through the game is the actual data. There is no fixed score, the output is a behavioural pattern. Familiarity helps a little, but you cannot game the score. Common NL employers include ABN AMRO, Rabobank (often), JPMorgan, Goldman, Unilever, and consumer brands.

Most NL graduate programmes use more than one type. A typical assessment day looks like 30 minutes of cognitive tests, plus a 10-minute SJT or personality module, plus a short pymetrics-style game battery if the employer uses one.

Vendor names have shifted in the last few years through acquisitions. We use both old and new names where the rebrand is recent enough that you might see either in the wild.

SHL

The dominant cognitive-test provider in the Netherlands and across Europe. Their flagship product line is SHL Verify (and Verify Interactive, the more modern UI). SHL tests typically cover verbal, numerical, inductive (pattern recognition), and deductive reasoning, often with a general ability test that combines all four.

Format: 16 to 30 questions per section, 12 to 25 minutes per section. Multiple-choice or drag-and-drop. SHL also offers situational judgement tests, personality questionnaires (the OPQ32), and a video interview module.

Where you will see them: Big 4 (audit and consulting graduate intakes), large NL corporates, banking, public sector, and many international graduate programmes. Free SHL practice tests are available on shl.com/shldirect.

Cut-e (now Aon Assessment Solutions)

Cut-e was acquired by Aon and rebranded as Aon's Assessment Solutions. You may still see “cut-e” in URLs and reports. Their test family is called Scales, with around 40 different tests across numerical, verbal, logical, mechanical, spatial, and language ability.

Format: Cut-e tests are short. Most run 5 to 20 minutes, but you usually take several in a row. The verbal and numerical tests use a tabbed case-study format (read the company brief, find the answer, decide if a statement is true, false, or insufficient information). The logical tests (cls, clx) use grid-pattern puzzles a bit like sudoku.

Where you will see them: large multinationals (DHL, Lufthansa, Telefonica), some banking and consulting firms, and an increasing share of European corporates.

Cubiks (now Talogy)

Cubiks rebranded as Talogy in 2022 after PSI Services acquired it. Their main ability test is the Logiks family. Logiks General (Intermediate) is a 12-minute, 50-question general ability test. Logiks Advanced is broken into separate verbal, numerical, and abstract reasoning sections.

Format: Highly time-pressured. The intermediate Logiks gives you about 14 seconds per question. They also publish PAPI personality questionnaires and bespoke situational judgement tests.

Where you will see them: smaller- to mid-size NL employers, some financial services firms, and a long tail of European corporates. Less common than SHL but you do encounter it.

Pymetrics (now part of Harver)

The big game-based provider, acquired by Harver in August 2022. The standard Pymetrics test is 12 short games (2 to 3 minutes each), totalling 25 to 35 minutes. Some employers also add 4 numerical and logical reasoning games on top.

Format: You play games. One game asks you to inflate balloons before they pop (testing risk tolerance and learning). Another asks you to remember and react to colours (testing attention and processing). Another asks you to identify emotions in faces (testing emotional recognition). Each game collects multiple data points. There is no single “score” in the traditional sense; the output is a behavioural profile.

Where you will see them: investment banking (JPMorgan, Goldman Sachs), consulting (Bain, BCG), large consumer brands (Unilever, HSBC), and increasingly NL banks like ABN AMRO. The format surprises a lot of graduates because it does not feel like a test.

This is not exhaustive. It covers the patterns that hold across multiple cohorts and that are publicly documented. If your employer is not listed, the four families above still cover almost every test you might face.

Banking traineeships

• ABN AMRO: Cognitive testing combined with game-based assessment. Pymetrics-style games are common in early stages, with cognitive reasoning later.
• Rabobank: Virtual assessment centre with multiple components. Cognitive and personality testing typical, paired with case exercises and a personal pitch.
• ING: Online assessment battery early in the process, including numerical and personality components. Followed by digital interview and assessment day.
• JPMorgan, Goldman Sachs (NL offices): Pymetrics games early, then SHL or proprietary numerical testing, then HireVue async video. Investment banking is the heaviest user of game-based assessments.

Big 4 and consulting

• Deloitte, EY, KPMG, PwC: SHL is most common for cognitive testing. Personality and situational judgement modules are usually paired alongside. Final stage is a live assessment day.
• Bain, BCG, McKinsey: Each has its own ecosystem. McKinsey uses the Solve assessment (a gamified problem-solving test). BCG used to use Pymetrics-style games (now updated). Bain runs its own SOVA online test plus case interviews. None of these is a standard SHL or cut-e test.

ORMIT-tier and graduate trainee programmes

• ORMIT, Cqore, Yacht, JBR: Cognitive testing typical, often SHL or cut-e. Personality and motivation assessment included. Heavy emphasis on the live assessment day later in the process.

Large NL corporates

• Shell, Philips, ASML, Heineken, Unilever NL: Mix of providers. Unilever is well-known for using Pymetrics games. ASML uses cognitive and personality testing. Shell and Philips use varying combinations of SHL and cut-e depending on the role.

Public sector and Rijksoverheid

• Rijkstrainee programmes and government departments: Tests are usually administered by an external assessment partner (often Aon/cut-e or a Dutch-specific provider like Ixly). Cognitive plus personality plus a written case is typical.

What the research shows about practice effects

The most-cited meta-analysis on this is Hausknecht et al. (2007), which looked at 50 studies of cognitive ability test retesting. The headline finding: practice does improve scores, and the improvement is meaningful (around 0.26 standard deviations on the second attempt, which translates to several percentile points).

But: the effect drops sharply after the first retest. Going from one practice attempt to two is much smaller than going from zero to one. Going from ten to twenty is essentially nothing.

In plain English: the first 2 to 4 hours of practice get you most of the available improvement. After that, you are getting tired without getting better.

A practice plan that fits the research

• Step 1, identify the provider. Use Section 3 of this guide plus a quick search of recent candidate reports on Glassdoor or prep sites.
• Step 2, do one full-length practice test. Take it under realistic conditions: the right time limit, no calculator if the real test does not allow one, no breaks. The provider websites (shl.com, aon.com, talogy.com) all have free samples.
• Step 3, review your wrong answers carefully. This is where most of the actual learning happens. Understand the question type, not just the answer. Numerical questions usually fall into a small number of patterns (percentage change, ratios, currency conversion, reading from layered tables); knowing the patterns matters more than calculation speed.
• Step 4, do a second full practice test. Same conditions. You will see most of your improvement here. Compare your time per question, not just your score.
• Step 5, stop. Two practice tests for each provider you face is enough. Beyond that, the time is better spent on other parts of your application: your CV, your motivation letter, or preparing for the interview that follows the test.

Free vs paid practice resources

Free resources are usually enough for the first practice round. Provider websites publish samples, and university career services often have access to subscription practice databases (ask). Paid platforms like JobTestPrep, GraduatesFirst, AssessmentDay, and TestHQ offer larger question banks and detailed explanations.

Paid resources are worth it in two cases. One: you are applying to a single specific firm and want to drill its exact format. Two: you have failed the test once before and want to be sure of the second attempt. For a general first try, free is fine.

What does not help

• Memorising sample questions. The actual test never reuses the exact items.
• Reading 50 articles about “how to pass aptitude tests” without taking any practice tests. Tips are no substitute for time on the format.
• Practising the wrong provider. Cubiks logic puzzles look very different from SHL inductive reasoning.
• Practising at 10 PM the night before. Test performance is sensitive to fatigue. Take the real test in the morning if the platform lets you choose.

Setup

• A reliable internet connection. Wired ethernet if you have a port. Avoid coffee shop Wi-Fi.
• A second device on hand. If your laptop crashes, some platforms let you continue from a different device. Most do not, but it is worth knowing.
• Paper and pen for working out numerical questions. Most platforms allow this.
• A calculator if the platform allows one. Cubiks and SHL numerical tests usually do; cut-e Scales numerical tests usually do not. Check the test instructions before you start.
• A quiet room. Tests are time-sensitive, and an interruption mid-question can cost you several percentage points.

Timing strategy

Almost every cognitive test is designed so that finishing every question is unlikely. The pacing strategy that works for most graduates: never spend more than the average time per question. If you do not see the answer in 60 seconds (or whatever the average is for that test), guess and move on. Returning to skipped questions is usually worth less than getting through more questions on the first pass.

Most platforms penalise wrong answers the same as blank ones, so a guess is better than a blank. The exception is cut-e logical tests, which sometimes use formula scoring (wrong answers cost more than blanks). Read the test instructions on this point specifically.

If your internet drops

Do not panic. Most platforms autosave progress every few questions. Reopen the test link from the same device. If the platform says you cannot continue, email the recruiter the same day, briefly explain what happened, and ask whether the attempt can be reset. Stay factual. Do not over-apologise.

The honesty disclaimer

Almost every online assessment opens with a screen that says something like “I confirm I am taking this test alone, without help.” Tick that and then have a friend help you, and two things can happen. First, you score much higher than the cohort average for someone with your CV. That can flag your test for review. Second, many employers re-run a shorter, supervised version at the assessment-day stage. If your scores diverge sharply, the offer is at risk.

The practical answer: take the test alone, even if no one will check. The supervised retest is real.

The famous Schmidt and Hunter (1998) finding

For 25 years, almost every guide to graduate assessments cited Schmidt and Hunter's 1998 meta-analysis. The headline number: cognitive ability tests had a validity coefficient of 0.51 for predicting overall job performance. That made cognitive ability the single strongest predictor available, beating interviews, work samples, education, and reference checks.

That finding shaped how employers thought about hiring for two decades. It is also why so many graduate programmes lean heavily on cognitive testing as an early-stage filter.

The 2022 recalibration

In 2022, Sackett, Zhang, Berry, and Lievens published a major recalculation. They examined the statistical methods Schmidt and Hunter used (specifically the corrections for range restriction) and concluded that prior validity estimates were systematically overstated.

Their corrected estimate for cognitive ability dropped from 0.51 to 0.31. That is a meaningful drop. They also found that structured interviews now appear to be the strongest single predictor of job performance, with a validity of 0.42, higher than cognitive ability.

A follow-up analysis by Griebie et al. (2022) looked only at 21st-century data and found the cognitive-ability validity even lower: around 0.23. That is still a real, useful signal, but it is far from the deciding-factor framing the 1998 number invited.

What this means for you

Cognitive ability tests still predict performance and still get used everywhere, so doing well on them still matters. But the picture is now: they are one input among several, not a single deciding score. Imagine two graduates. One scores in the top 30% on the cognitive test, performs strongly in the structured interview, and shows up well on the work-sample exercise. The other scores in the top 10% on the test alone. The first usually wins.

Practically: do not panic if you do not feel like a test wizard. The interview round is now believed to carry more weight than the test in the overall hire decision. If you can demonstrate strong reasoning during the actual conversation, that compensates for a non-spectacular test result more than candidates assume.

In Dutch graduate programmes, the cognitive or aptitude test is one of several stages, not the deciding step. A typical sequence:

• CV and motivation-letter screen
• Online assessment (this guide)
• Recruiter or HR phone or video screen
• Hiring manager interview, sometimes with a behavioural-questions focus
• Assessment day or final round, often in person, with cases and panel interviews
• Offer

The relative weight of the test depends on the firm and the stage. Earlier in the process, it acts as a filter (pass a threshold and continue). Later, it becomes one input among several when the panel decides between final candidates.

Related posts in this library

• [LINK PLACEHOLDER: post #10 How to Prepare for a Job Interview in the Netherlands]: the orientation guide for the whole NL graduate-interview process. Worth reading first if this is your first NL application.
• [LINK PLACEHOLDER: post #11 Succeed at a Video Interview]: covers async AI video rounds (HireVue and similar) and live video interviews. These often follow the cognitive test.
• [LINK PLACEHOLDER: post #12 STAR for Behavioural Questions]: the framework for handling “tell me about a time when” questions in the interview round.

Identify which provider you are facing. Run two timed practice tests in that format. That is the single most useful thing you can do before a graduate-programme assessment, and most candidates do neither. They walk in cold.

Aurora, GradGuide's free AI career coach, can help you run that pre-test routine. You can paste in the invite email and have her identify the likely provider, walk through worked examples in the matching format, and time your responses against the real test cap. After two laps with her, the real test stops being a surprise.