In-depth guides on Arctic Shores, SHL, Pymetrics, Cappfinity, and other assessments, plus interview preparation to help you stand out across every stage of graduate hiring.
The most expensive preparation mistake is practising the wrong test. How to work out which provider your invitation came from, using the product names in the email and the format itself.
Two tests can run for the same twenty minutes and work completely differently underneath. What computer-adaptive testing changes about pacing, review and reading your own performance.
Most providers publish no validity period, because the employer decides. What is actually known about reuse windows, score banking and who your result travels to.
Some assessments will not open on a phone at all and some are built for one. What the publishers actually say about devices, and a recommendation per test type.
A rejection after a test does not tell you the test caused it. How to find out which stage ended the application, what you can ask for, and whether practice will change anything.
Neither firm publishes what its test contains, though Bain names its two vendors, and both vary the stage by office. What the sources agree on about Casey and Sova, where they contradict each other, and why preparing for "the MBB online test" as one thing does not work.
Most Fast Stream guides describe a process that is no longer published. What gov.uk actually lists for 2026 entry, where the Judgement Test and Work Strengths tests really belong, and why the assessment centre marks 4D rather than the nine behaviours.
A German Einstellungstest is not an aptitude battery, French law gives you rights before the test rather than after it, and EPSO lets you choose which of 24 languages you reason in. Three systems with one legal floor underneath them.
Five publishers use three words for two abilities, so the word on your invitation tells you less than it looks like it does. What each one measures, who calls it what, and why the two need opposite habits at the desk.
Banking and finance run more selection stages than almost any other sector, and the first two are decided without a person. The sequence, the vendors, and why the numerical bar is genuinely higher.
Ireland has the British test types without the British timetable, and a majority of its graduate recruiters do not use aptitude tests at all. Canada runs a slower, competency-driven process where the Second Language Evaluation ends more applications than any reasoning test.
The arithmetic is the small half. The marks go missing in the wrong row, the units line nobody read and the second table nobody opened, and every one of those has a matching wrong answer waiting in the option list.
A corporate assessment predicts performance. A police or fire service also screens for public risk, which is why judgement comes first, physical tests are pass or fail, and disclosure is itself an assessed task.
The two formats are the same exercise with different plumbing. Which scoring stream you are in at any moment decides where your minutes should go, and doing the work yourself is the most expensive habit you can bring.
Assessors do not form an impression of you, they run a procedure: observe, record, classify, rate. Knowing the procedure explains why dominating the discussion costs marks and which four contributions reliably earn them.
A one-way interview is judged as a file, not a conversation. What survives into the score, why expression management is wasted effort, and the settings to read before question one.
One evening is enough if you spend it on the right things. What to research, the handful of stories worth preparing, and why saying them out loud matters more than reading.
If your last interview was a chat over coffee, the current process can feel like a different sport. What has changed, what has not, and where to put your preparation.
Assessors are collecting written evidence against four or five competencies, not judging your stage presence. What that changes about structure, missing data and the questions afterwards.
True, False and Cannot Say look like reading comprehension and are not. What each verdict claims, where implied stops being allowed, and how to read at 50 seconds a question.
There are no right answers, and that is exactly why candidates lose ground on them. What the scales measure, what the validity checks catch, and how to answer.
You got 32 out of 40. Is that good? The honest answer is that nobody can tell you without knowing the norm group. Here is how assessment scoring really works.
Practice effects on cognitive tests are real and well documented. They are also front-loaded, modest and easy to waste. Here is where the gain actually comes from.
Extra time on a speeded test is the difference between measuring your ability and measuring your reading speed. How to ask for adjustments, and when.
The retake question has two halves: what the vendor allows technically, and what the employer allows in practice. They are almost never the same answer.
Assessment anxiety is not a character flaw, it is a predictable response to an unfamiliar interface and a visible clock. Most of it can be designed out in advance.
Most SJT advice ignores the response format entirely. Ranking all four options and picking the single best answer are not the same task, and they are not scored the same way.
If you are applying in the United States, the vendor list is different, the tests are shorter, and the legal framework around them is unlike anywhere else. Here is the map.
Australian graduate recruitment runs on a vendor list that barely overlaps with the UK one. If you are preparing from British advice, some of it does not apply.
A game-based assessment records how you behave, not what you answer. That single difference explains almost everything candidates find strange about them.
A calculator does not save a numerical reasoning test. Locating the right figure and estimating faster than you calculate is what saves it.
You sat the test, the employer got a report, and you got a rejection email. In several jurisdictions you have a right to more than that.
Most candidates prepare in the wrong order: content first, format last. With seven days, reversing that is worth more than doubling the hours.
Checking tests are the most improvable section in any assessment battery and the one candidates spend the least time on. Both facts are worth acting on.
Abstract reasoning looks infinite and is not. A small set of rule families covers most items, and a fixed inspection order finds them faster than staring does.
HireVue sells four different things and sends them all from the same platform. Most candidates prepare for the video interview and then meet a set of games. How to read the invitation.
Four numbered versions, six kinds of exercise and one pass mark that is only the floor. Which VEA you will sit, what each section asks, and what practice is actually for when USPS says you do not need to study.
Most TSA prep still sells an X-ray and English test that applicants stopped meeting in early 2023. What the TSO Assessment Battery actually asks, section by section, and where the marks go.
Discover proven strategies to prepare for Arctic Shores psychometric assessments. Learn what to expect, how to practice effectively, and tips to perform at your best.
SHL is the most used aptitude test in the world, which is why most advice about it is generic. Timings, traps and tactics for each test in the Verify range, plus the OPQ32.
The FAA says no preparation is necessary for its air traffic controller test. Candidates who have sat it say knowing the mechanics was the difference. What each of the seven sections asks, and where marks go.
Every police constable applicant in England and Wales now sits the same two-part sift. What the rating scales really mean, why the questionnaire counts, and where most preparation material is years out of date.
McKinsey publishes no module names, no durations and no scoring model. The one number you do have is the length in your invitation email, and it tells you which modules you are getting.
Every EU institution job opens with the same three reasoning tests, but the pass rules are not what most candidates assume, and in the AD5 graduate competition only one of the three ever ranks you. The arithmetic, the weights and the pace for each test.
Aon is the only major provider that punishes wrong answers. That single rule changes when you should answer and when you should skip. The maths, and what it means per test.
A TMA invitation is two different things: a drives questionnaire with no right answers and a set of short, fast, separately scored capacity subtests. The pace differs fourfold between them, and a 5 is average, not a fail.
Four names, one publisher. If your invitation says eelloo, Meurs HRM, Q1000 or Matrigma you are dealing with Assessio, and the score you get back will not be a percentile.
An invitation from invitation@alvalabs.io means two short adaptive tests and a result you see the moment you finish. What the tasks look like, how the score is worked out, and why Matrigma habits will pace you wrongly.
LTP’s G-test mixes six non-verbal item types through fifty questions, and it is only one of five LTP capacity tests. What each contains, how the score is judged, and what to practise.
Law firms screen with the Watson Glaser because it catches something general intelligence does not. Each section uses a different standard of proof, and mixing them up is where the marks go.
Five names, one test library, and a platform that shut down at the end of 2021. How to work out which of them your invitation is actually pointing at, and why seeing BrassRing tells you nothing about who is testing you.
The aptitude test for almost every RAF, Royal Navy and Royal Marines role is six sections sat as eight separate clocks, and its results decide which roles you can apply for rather than whether you pass. Part by part, what it asks and how to prepare.
The average CCAT taker gets 24 of 50 right and fewer than 1% finish. The test is built that way, and the candidates who lose most are the ones trying to complete it.
An EEI invitation means one of several batteries, picked by the job rather than the utility. How to tell CAST from POSS, MASS, TECH and SO/PD II, what each subtest asks, and why there is no universal pass mark.
Wonderlic weights each question by difficulty and its items get harder as you go, so the marks are concentrated in the part of the paper most people never reach.
Your agency picks the CritiCall modules and sets the bar for each. The decision rules and the typing rate are what reward practice most.
Korn Ferry Talent Q combines two features nothing else does: adaptive difficulty and a separate clock on every question. Both change the pacing rules you learned on other tests.
Which USA Hire assessments you sit depends on the job’s series and grade, not the agency, and the announcement names them in code. How to decode it, and the grade pattern in OPM’s lookup data.
The AFQT is not a raw score or an average. It is four standard scores, with the verbal one counted twice, turned into a percentile by a published table. Step by step, with the worked examples behind the numbers recruiters quote.
Three Australian emergency services use ACER, and the same abstract reasoning test runs at 45 questions in 20 minutes for one of them and 30 in 35 for the others. What else changes between them, and what does not.
Most Police Scotland practice material still describes the Standard Entrance Test, which was retired in 2020. The current paper is one hour, 49 questions and three elements you must each pass, and its language element looks nothing like the old one.
The national firefighter selection tests are no longer national. Most services now sift online with a vendor, and the old paper tests survive mainly for on-call recruits and in Wales. Which one you face changes everything from the calculator to the answer format.
Most PELLETB practice material still teaches CLOZE and a 2.5-hour paper booklet. Both are gone: POST removed CLOZE from every sitting from 1 January 2025, and in 2026 it began moving the test onto a computer, taken in person at the agency.
Swift Analysis sections are reported at about 45 seconds a question, and the diagrammatic section uses an operator format most candidates meet for the first time on test day.
Most preparation pages still describe a BRAIN test AssessFirst retired in 2020. What the three questionnaires ask today, what the 1 to 10 score means, and why you should never practise on the real site.
Most advice about the GIA tells you to work quickly and accurately, which is what Thomas says and not much help. The five tasks do not reward the same balance, and the published sample reports show exactly where each one tips.
Every section of the ACT hides its material before you answer. What the five sections ask, what the screener and the Verifier are, and which GTI soldiers say each role needs.
The CAT is a platform of modules, not one test, it is adaptive by design, and it has no pass mark. What the Bundeswehr actually says, and what preparation sites add without a source.
Logiks Intermediate strips out the reading and the data tables and leaves raw speed: word puzzles, mental arithmetic and block sequences in four-minute bursts. Advanced puts the reading back.
Almost every Civil Service test is untimed, the Verbal and Numerical tests stop when they know your level, and the judgement tests have a rating scale with no middle. What that changes about how you prepare.
The PSC swapped three of its four occupational tests on one day in July 2025, and much of the prep material online still teaches the old ones. A dated list of what changed, and what the new tests reward.
Prep sites still sell CFAT practice, but the Canadian Armed Forces dropped the test for Regular Force applicants in October 2024. What it measured, what replaced it, and what is worth preparing instead.
Most advice about adaptive tests stops at "it gets harder if you do well". Test Partnership publishes the number it targets, and it is not the one you would expect, which changes what your practice accuracy is telling you.
The word in brackets on a Clevry invitation is the level, and each level is a different test with a different clock. Then comes the page rule, which quietly rules out the two habits most practice teaches.
PI sends a 12-minute cognitive test and an untimed adjective checklist, and confuses candidates by calling both an assessment. What each measures and how to approach them.
Connector Ability is the test PiCompany built and GITP now sends. It gives you far more time per item than prep sites claim, does not score speed, and compares you with the education level your employer picked.
Most timed-test advice tells you to pace yourself. Ixly has defined an unanswered question as a wrong one, which makes silence the only answer that cannot score, and its sten scale means a 5 is average rather than a fail.
The police publish their pass mark, in a ministerial regulation. It is lower than it used to be, it is a percentile rather than a score, and a supervised check decides whether your home result stands at all.
There is no good Pymetrics score, only a profile compared against a role model you cannot see. That changes what preparation means, and makes strategic answering actively risky.
A TestGorilla invitation lists its tests by name. What each of the seventeen cognitive tests is, which roles tend to meet which, and why practising on TestGorilla itself is a trap.
Sova shows you no timer and times you anyway. That one design decision changes how the whole sitting should be paced, and it is the thing candidates find out afterwards.
Most guides to the prison officer online test still describe a game that HMPPS retired in April 2026. What stage 1 and stage 2 look like now, what is published and what is not.
The train driver tests did not change in December 2024; the rules around them did. How the three attempt rule and its waits work, what the national database records, and what national and enhanced standards mean when nobody shows you a score.
Cognify looks like a set of puzzle games and is scored like a cognitive test. A game-by-game breakdown of what each one rewards and where candidates lose time.
Most Australian candidates invited to sit "Revelian" have the RCAT, not Cognify. What it actually is, how its three question types rotate, and the triage habit the vendor itself recommends.
NZ Police never names the test on its own page. It is Podium’s GCAT: one mixed, adaptive run of 60 questions with no going back, and a personality profile of 100 statements.
The Hogan Development Survey reads like an accusation if nobody has told you what it is. Every scale on it is a strength taken too far, and knowing that changes how you answer.
Two rules decide more Master ACE results than any arithmetic practice, and both run against the habits other tests teach. Why a skip costs you, why the calculator is expected, and what your STEN means for the role.
A strengths assessment records not just what you answer but how quickly. That makes constructing the ideal response counterproductive, and it changes how you prepare.
Most candidates never see the name Amberjack except in the link. What the immersive assessment actually asks, why there is no clock but every answer is timed, and what the video questions need.
A 5 out of 10 on a Dutch assessment report is not five out of ten. It is a sten, the scale has a midpoint of 5.5, and the publisher says in its own report that a 5 is not a fail.
Eight subtests with three names each, a report that shows Lager, Midden or Hoger instead of a score, and a 21+ admission route that puts hundreds of applicants a year through it. What the NOA test is and how to prepare.
An invitation that says Harver can mean reasoning subtests, a two-sided personality questionnaire, best-and-worst scenarios, a three-customer chat simulation, a typing test or pymetrics games. How to tell which you have, and how each one is scored.
Learn what psychometric tests measure, why employers use them, and how to demonstrate your cognitive abilities, personality traits, and behavioral patterns.
Master cognitive assessments with these expert tips. From managing test anxiety to optimizing your performance, learn the strategies that work.
Deep dive into the Balloon game mechanics. Learn optimal pumping strategies, risk assessment techniques, and how to maximize your score.
Improve your performance on the Arrows game with proven focus techniques. Learn how to filter distractions and maintain consistency under pressure.
Enhance your working memory capacity with evidence-based strategies. Learn chunking, visualization, and spatial memory techniques for the Sequence game.
Complete guide to preparing for Amazon graduate and professional interviews. Learn how to answer Leadership Principles questions using the STAR method.
How to prepare for a KPMG graduate interview. Covers KPMG values, typical competency questions, the assessment centre, and tips for standing out.
A practical guide to the STAR method. Learn how to structure strong answers for graduate and professional competency-based interviews.
Criteria publishes more than most about how its test is scored, and less than the prep sites claim. What the raw score, the percentile and the three sub-scores mean, where the median sits, and how employers decide what is enough.
Test Partnership publishes a full candidate report for a sample candidate. Game by game, what the percentile, sten and band mean, why the norm group moves the number, and how six results become one combined score.
Kenexa sells two families of test that are scored in completely different ways, and IBM does not set the bar for either. What each report contains, what IBM says about cut-offs, and where the commonly quoted percentile figures come from.
Thomas publishes a full PPA sample report. Once you see how the three graphs relate, with Self Image exactly Work Mask minus Behaviour Under Pressure, the numbers are easy to read, and it is clear why there is no good score to aim for.
Finish the HireVue games and you get no number, but the recruiter does. What the employer report contains, how the battery-level score is built and validated, who you are ranked against, and what the Candidate Insight Report does and does not tell you.
The EPP has no right answers, so its report is a profile rather than a score: twelve trait percentiles, a percentage match to a job family, interview questions built from your results, and flags for careless or self-flattering answers.
Put these strategies into action with our comprehensive practice platform
Start Practising Free