Numerical reasoning (scales numerical)
37 statements · 12 minutes (about 19.4 seconds per question)
What it tests. Selective visual filtering, high-frequency data retrieval and rapid quantitative deduction, not advanced maths.
Worked example. Tab 1 shows EMEA Acquisition Financing revenue of £450m (FY24) and £520m (FY25); Tab 2 shows headcount of 110 then 130. The statement 'average revenue per employee rose more than 5%' is False because per-employee revenue fell from about £4.09m to £4.00m.
Common traps. The 'perfect calculation' pitfall (typing exact decimals into a calculator and completing only 15 of 37), and misusing 'Cannot Say' just because you feel rushed.
How to handle it. Spend the first 10-15 seconds clicking through all 6 tabs to map their titles, then estimate aggressively rather than computing exact figures.
Inductive logic (scales ix)
12 tasks · 12 minutes (60 seconds per task)
What it tests. Fluid intelligence, non-verbal pattern recognition and conceptual flexibility under no linguistic guidance.
Worked example. Group Alpha grids have two vowels adjacent to an even number and digits summing to an odd total; Group Beta has vowels separated by consonants and digits summing to an even total. You classify a new matrix within the 60-second window.
Common traps. Fixating on a single variable (only the letters) and sunk-cost time management, spending three minutes convinced you are 'close'.
How to handle it. Run a fixed checklist (count, then maths, then geometry, then case/type). If you cannot crack it in 45 seconds, log your best guess and advance.
Situational judgement (scales sjt)
20-25 scenarios · Untimed or softly capped around 20-25 minutes
What it tests. Alignment with Nomura's values, professional judgement, risk awareness and escalation protocols.
Worked example. At 8:15pm you find a DCF formula error that moves enterprise value by 12%, with the deck due to the MD by 9:00pm and the Associate gone. The strongest action is to message the MD transparently, explain the discrepancy and request a short extension to ensure accuracy.
Common traps. The 'hero' fallacy (fixing everything in a silo with no notification), extreme passivity, and applying start-up 'rules are made to be broken' logic to a regulated bank.
How to handle it. Filter every option through Nomura's pillars (Entrepreneurial Leadership, Teamwork, Integrity): accuracy beats deadlines, escalate responsibly and protect the team without covering up failures.
Behavioural / personality questionnaire
55 statements · Untimed, typically 30-35 minutes
What it tests. A psychometric map of personality, risk appetite, resilience and working style, benchmarked against high-performing Nomura analysts.
Worked example. Choosing between 'I am highly competitive and measure success by outperforming peers', 'I prioritise consensus before executing', and 'I keep exceptional attention to detail under extreme time constraints'.
Common traps. Creating an inconsistent profile that the algorithm's consistency metrics flag as unreliable, and defaulting to neutral middle answers that produce a weak, low-definition profile.
How to handle it. Hold a consistent professional persona: highly analytical, detail-oriented, collaborative, receptive to feedback and resilient, without presenting yourself as a rule-ignoring risk-taker.