Abstract Reasoning as the Purest Measure of Cognitive Flexibility
Abstract reasoning is the cognitive function that recognises patterns in novel material, infers the rule that generates a sequence, and applies that rule to new situations. Because the test items use shapes, sequences, and matrices with no semantic content, the score is largely independent of the candidate's education, language, or domain background. This is exactly why employers use it. A candidate's verbal reasoning score might reflect their schooling, their numerical reasoning score might reflect their mathematics teacher, but their abstract reasoning score reflects what Charles Spearman in 1904 called g, the general factor of intelligence.
The Cattell-Horn-Carroll model of cognitive abilities, the dominant framework in contemporary psychometrics, identifies fluid intelligence (Gf) as the broad ability most closely measured by abstract reasoning tests. John Raven first published the Progressive Matrices in 1938 specifically to measure Gf with minimal cultural bias, and the test remains the most widely used abstract reasoning instrument in the world a century later. MENSA, the high-IQ society, uses Raven-equivalent items for admission. Many of the corporate batteries described below are descended directly from Raven's original design.
The Test Instruments Behind the Interview
Raven's Standard Progressive Matrices (SPM) and Advanced Progressive Matrices (APM) remain the gold standard. Each item presents a 3 by 3 matrix of geometric figures with one cell missing, and the candidate selects the option that completes the pattern. The SPM is calibrated for the general population, the APM for high-ability adults. Selection programmes that want a defensible measure of fluid intelligence reach for Raven first.
SHL Inductive Reasoning is the most widely deployed corporate descendant. The test presents sequences of abstract shapes and asks the candidate to identify the next item in the series. SHL's items are tighter and more time-pressured than Raven's, designed to fit within graduate selection batteries where the candidate has 25 minutes for 24 items. The cognitive demand is the same: extract the generating rule under pressure.
Cattell Culture Fair Intelligence Test, developed by Raymond Cattell in 1949 specifically to minimise cultural and linguistic bias, is used in international hiring and research contexts where the candidate pool includes multiple language and education backgrounds. The Culture Fair items are visually distinctive from Raven's but measure the same underlying construct.
Saville Wave Aptitude and Talent Q Elements Logical include abstract reasoning components calibrated for adult professional populations. Saville's items in particular are designed for senior selection, where a chief executive or board candidate may face an abstract reasoning component as part of a leadership assessment.
McKinsey Solve and BCG's Online Case load abstract reasoning indirectly through their game-based scenarios. The ecosystem and case scenarios require the candidate to extract the underlying rule structure of the simulated environment and apply it to novel situations, which is abstract reasoning rendered in a contextualised form.
How Abstract Reasoning Shows Up in the Live Interview
Strategy interviews at McKinsey, BCG, and Bain include cases that probe abstract reasoning under the surface. The candidate is presented with a novel industry, often unfamiliar (let us say, regional propane distribution, or speciality industrial gases), and must reason about the industry's structure without prior knowledge. The case is fundamentally an abstract reasoning task: extract the structure of the system, identify the key relationships, predict how a change to one variable would propagate.
Technology firm system design interviews require abstract reasoning fluency. The candidate is asked to design a system the firm has never built (a chat application for one billion concurrent users, a real-time fraud detection pipeline, a global content delivery network). No design pattern from the candidate's prior experience maps directly. The candidate must reason from first principles about the structural requirements and the system that satisfies them, which is applied abstract reasoning.
Quantitative finance interviews at firms like Jane Street, Citadel, and Two Sigma include probability puzzles and pattern problems that are pure abstract reasoning. "Given the following sequence of moves, what is the probability that the next move is X?" Candidates who can extract the generating process from a few observations outperform candidates with stronger formal mathematics but weaker pattern fluency.
Research and development hiring at pharmaceutical and biotech firms uses abstract reasoning probes through novel experimental design questions. The candidate is asked how they would design an experiment to test a hypothesis they have never encountered before, with constraints designed to invalidate the standard textbook approaches.
Industries Where Abstract Reasoning Sets the Ceiling
- Strategy consulting: McKinsey, BCG, Bain. The cases that distinguish the top of the candidate pool are the ones in unfamiliar industries, which depend on abstract reasoning rather than recall.
- Quantitative trading and research: Jane Street, Citadel, Two Sigma, DE Shaw, Renaissance Technologies, Optiver, IMC. The interview rounds are explicitly designed to find candidates with strong abstract pattern recognition.
- Research and development: Pharmaceutical and biotech firms hiring discovery scientists. Industrial research labs at Google, DeepMind, Anthropic, OpenAI, Meta FAIR, Microsoft Research.
- Software architecture and senior engineering: System design and senior staff engineering rounds that require structural reasoning about systems the candidate has never built.
- Investment management: Public equity research and macro hedge fund hiring weights abstract reasoning heavily, because the work demands extracting patterns from data without the candidate's prior experience providing the template.
- Top-tier graduate programmes: Civil service, central banking, supranational organisations. The UK Fast Stream and the European Commission AD5 entry use abstract reasoning batteries explicitly.
- Elite university admissions: The LSAT, GMAT, and many institutional aptitude tests load on abstract reasoning even where the surface content is verbal or quantitative.
What Effective Preparation for Abstract Reasoning Actually Looks Like
The first thing to internalise is that abstract reasoning is less trainable than verbal or numerical reasoning, but it is not untrainable. Research on practice effects in fluid intelligence (Jaeggi, Buschkuehl, Jonides, and Perrig's 2008 work on working memory training and the substantial follow-on literature) suggests that targeted training can produce modest gains on transfer measures, though the size and durability of those gains remains debated. For test-specific preparation, the gains are clearer: candidates who practise on Raven-style items see meaningful score improvements over four to six weeks of consistent work.
The effective routine is mechanical. Twenty to forty timed sets of 24 SHL-equivalent items, with detailed review of every wrong answer. The review identifies the rules the candidate is missing: 90-degree rotation, reflection, intersection of two operations, addition of two patterns, subtraction of common elements, sequential transformation. Candidates with strong baseline abstract reasoning typically recognise these rules in a handful of items; weaker candidates need explicit exposure to each rule family.
The Jaeggi-style working memory training, particularly dual N-back, has shown some near-transfer effects to fluid intelligence measures in some studies, though replication is mixed. For candidates with time, including 10 to 15 minutes of N-back practice in the daily preparation routine is low-cost.
Reading widely in unfamiliar fields, particularly mathematics and theoretical physics, builds the cognitive habit of extracting structure from new material. The candidate who has worked through introductory linear algebra, group theory, or topology has practised the same cognitive operation that abstract reasoning tests measure, and the transfer is real if the engagement was active rather than passive.
The Errors That Cost Candidates the Role
Locking in too fast is the single most common abstract reasoning failure. The matrix has eight cells, the candidate sees a plausible rule from the first row, and selects the answer that fits that rule without checking whether the same rule holds in the columns. The discipline of checking both axes before committing is what separates strong from middling candidates.
Missing composite rules is the second. Many items combine two or more operations: a rotation plus a colour swap, an addition plus a reduction in size, an intersection plus a reflection. Candidates who find one rule and stop typically miss the second, which is exactly the discriminator the test designers built in.
Anchoring on the most visually salient feature is the third. The candidate's eye is drawn to the largest, brightest, or most distinctive element, and they reason about its transformation while ignoring a subtler but more important rule operating on a different element. Strong candidates make a habit of cataloguing all the features in the matrix before committing to a hypothesis about the rule.
Treating the items as cultural puzzles rather than rule-extraction tasks is the fourth. Candidates from non-Western educational backgrounds sometimes assume the items have culturally specific content and search for meaning that is not there. The items are intentionally semantically empty. The task is to find the pure structural rule.
What Strong Abstract Reasoning Means for the Career That Follows
Abstract reasoning is the cognitive trait that most clearly predicts where a candidate ceilings out. Strong domain experience plus weak abstract reasoning produces a senior practitioner who can do familiar work well but struggles with novel problems. Strong abstract reasoning plus modest domain experience produces a junior who can grow rapidly into senior roles because they can extract pattern from new material faster than their peers. The leaders who shape new industries, who write the influential papers, who build the systems that define a generation of engineering practice, are reliably found in the top end of the abstract reasoning distribution.
If you want a baseline measure of your abstract reasoning before facing a Raven-equivalent screen, an SHL Inductive Reasoning battery, or the structured reasoning components of a McKinsey, BCG, or Jane Street interview, take the Abstract Reasoning test to see where you sit on the same kind of items employers use, with breakdown by pattern type so you know which rule families to focus on if you decide to prepare further.