Abstract Reasoning Trainability: Modest but Real
Abstract reasoning tests measure fluid intelligence (Gf), the ability to identify patterns and relationships without prior domain knowledge. Unlike verbal or numerical tests, where practice produces measurable score gains, abstract reasoning improvements are typically smaller and ceiling-bound. In a longitudinal study by Cattell and Horn (1966), fluid intelligence showed minimal improvement with training compared to crystallized intelligence, which responds sharply to instruction.
That said, the gains you can expect are real within a narrow band. Raven's Progressive Matrices (the gold standard abstract reasoning measure) shows 5–15 point score improvements with focused 4-week preparation, primarily through format familiarity and speed optimization rather than raw pattern-detection ability. The constraint is neurological: your ability to hold spatial relationships in working memory, rotate mental images, and identify abstract rules is partly stable. But test-specific fluency, knowing what to look for, reducing hesitation, and managing time, is entirely trainable.
Week 1: Matrix Logic Fundamentals
Abstract reasoning tests are almost always matrix completion formats. You're given a 3×3 grid with one cell missing, and you must identify which of 6–8 answer options completes the pattern. The rules governing the pattern are visual, not verbal.
Start by reading systematically. Examine the matrix left-to-right on each row, then top-to-bottom on each column. Most patterns obey both row and column logic simultaneously. Common transformation rules include: rotation (shapes turn 45°, 90°, 180° increments), color or shading changes (white-to-black gradations), addition or removal of elements (line count increases, shapes accumulate), reflection (mirror images across axes), and size scaling (elements grow or shrink). Some matrices combine two or three rules at once.
Spend 15 minutes daily on untimed practice. For each matrix, write down the rule you observe before looking at answers. This forces deliberate pattern recognition over guessing. Use the first 5 examples in the official Raven's manual or equivalent practice sets (SHL's Inductive Reasoning Practice Guide includes detailed solutions showing the logic). The goal is to build a visual vocabulary, when you see a 3×3 grid in the actual test, you immediately recognize which 3–4 transformation patterns are at play.
Week 2: Pattern Type Recognition
Abstract reasoning patterns cluster into families. Learning to recognize each family reduces cognitive load during the test because you know where to direct attention.
Arithmetic progressions are rare in visual matrices but common in number-series items. A sequence increases by a fixed amount each step (e.g., +2, +3, +4). Train by covering the last 1–2 items in a series and predicting the next.
Geometric progressions scale by a constant factor (e.g., double, halve, rotate 90°). These are extremely common in Raven's. Look for patterns where each cell is a rotated or scaled version of the previous.
Set operations treat visual elements as members of logical sets. A cell might show "all shapes that are both triangles AND filled," while another shows "shapes that are triangles XOR filled." Saville Wave and Cubiks Logiks tests use this extensively. The key is learning to parse "AND," "OR," "NOT," and "XOR" visually without words.
Two-dimensional rules apply independently to rows and columns. For example, the top row might rotate 90° left, while columns increment the number of sides. Kenexa Logical Reasoning tests lean heavily on two-dimensional decomposition. Train by covering each row and column separately, identifying the rule in each, then synthesizing.
Spend 10 minutes daily categorizing practice items by family before solving. Say aloud: "This is a set operation," or "This is a scaling pattern." This metacognitive step embeds the taxonomy and reduces hesitation during the timed test.
Week 3: Timed Practice
Most abstract reasoning tests run 15–25 items in 15–25 minutes, yielding 60–90 seconds per item. Speed is half the battle. You can be a pattern genius, but if you solve 8 items in 20 minutes, you've lost the test before the last two.
This week, introduce time pressure. For days 1–3, use a 2-minute limit per item. Set a timer, solve, check your answer, then review the explanation. The goal is not accuracy, it's training your visual scan to be faster. You'll miss difficult items; that's intended. Difficult items often aren't worth the time anyway.
For days 4–7, shift to block timing: 5 items in 8 minutes (96 seconds per item average). This is closer to test tempo. You should skip or guess on items 2–3 minutes in if you're not converging. Don't be married to the difficult one.
During timed runs, track which patterns trip you up. If you're slow on set operations, spend 2 minutes daily on set-operation drills. If you hesitate on 3D rotation, use a physical cube or folding-net exercises (widely available online) to speed up spatial reasoning. The bottleneck is usually one pattern family, not all of them.
Accuracy targets: aim for 70% on untimed work, 60% on timed work in week 3. The timed score will feel lower, but that's expected. Your actual test performance will likely be 65–75% if you're hitting these benchmarks in practice.
Week 4: Mock Tests and Test-Day Strategy
Run one full mock test under exam conditions: 20 items, 20 minutes, in a quiet space, no pauses. Do this on at least two days this week (Monday and Thursday are good to space them). Score both, identify which pattern families caused misses, and spend 15 minutes reviewing each incorrect item with the solution.
Test-day strategy centers on triage. Abstract reasoning tests punish slow deliberation. If you spend 90 seconds on an item and still haven't converged to a hypothesis, mark it and move on. Come back only if you have time at the end. The test is usually designed so that 50–60% of items are "quick recognize and solve" and 30–40% are "deliberate 90-second work" and 5–10% are "guess or skip." Optimal performance means nailing the quick ones, getting 60–70% of the deliberate ones, and gambling strategically on the hard ones.
Anxiety management: abstract reasoning tests trigger anxiety in high-performing people because they feel less controllable than verbal or numerical tests. You can't "reason through" a pattern the way you reason through a logic puzzle. Acknowledge this. Your brain will activate fight-or-flight over a difficult matrix. Expect it. Use a simple grounding technique: when you feel stuck, take a 3-second breath pause, then shift focus to the next item. Come back to the hard one only after you've locked in the easy wins.
The night before, avoid heavy practice. Review the pattern families once with fresh eyes, then sleep well. Morning-of, do a 5-minute warm-up: solve 2–3 easy matrices to calibrate your visual scan, then go into the test.
Employer Tests You'll Encounter
Raven's Advanced Progressive Matrices (APM). The research standard. 48 items, 40 minutes. Matrices are larger and more abstract than beginner versions; many involve multiple transformations simultaneously. No color; purely shape and spatial logic. Used by consulting firms (McKinsey, BCG), investment banks, and academic selection. Your 4-week plan is built around Raven's format. Official UK distributor: Pearson. Practice sets available from their website.
SHL Inductive Reasoning (formerly CEB). 24 items, 20 minutes. Similar to Raven's but faster-paced. Emphasis on geometric progressions and set operations. Clear visual design, more "commercial" feel than Raven's austere format. Heavily used in tech hiring (Google, Amazon) and graduate schemes (Big 4 consulting, Goldman Sachs). SHL publishes free practice items on their site; the paid practice tests (£10–20) are worth it for familiarity.
Saville Wave Inductive Reasoning. 24 items, 25 minutes. Lower time pressure than SHL. Heavy emphasis on set operations ("shapes that are both X and Y") and logical negation ("everything except Z"). Common in UK public sector and civil service entry. The format is more about logical deduction than visual pattern completion. Their practice pack includes detailed explanations.
Cubiks Logiks (abstract). 20 items, 20 minutes. Combines visual matrix completion with number-series logic. More varied than pure matrix tests. Used in engineering recruitment and graduate programs. Fewer practice resources publicly available; many candidates find this one harder to prepare for because the variety is higher.
Kenexa Logical Reasoning. 18 items, 20 minutes. IBM's test. Format is closer to Raven's than to SHL. Emphasis on two-dimensional decomposition (row + column logic applied separately). Less commonly used than SHL or Raven's, but appears in large-corporation hiring. IBM publishes sample items on their assessments page.
Your strongest prep is to master Raven's format and SHL format, which account for 70% of employer tests. If you're applying to a specific employer, check their career page or ask the recruiter which test they use. Five minutes of format-specific familiarity the day before can yield 5–10 points on the day.
For deeper analysis of how these tests is associated with job performance and how to prepare strategically, visit our abstract reasoning test resource center.