Pearson

Raven's Progressive Matrices Practice: The Non-Verbal Test That Refuses to Play Favorites

Raven's is the test people underestimate because it looks like a puzzle book. There is no math, no vocabulary, and no passage to read. Just a 3x3 grid of shapes with one square missing. That simplicity hides what makes it hard: the rules that govern each matrix are invented on the spot, and you have roughly 60 seconds to reverse-engineer them.

Questions
36
Time Limit
40 min
Difficulty
Medium-High
Sections
2
Start Free Raven's Practice

What Raven's Progressive Matrices actually measures

John C. Raven built the first matrices in 1936 specifically to strip language and culture out of intelligence testing. Pearson now publishes three main versions: the Standard Progressive Matrices (SPM) with 60 items, the Advanced Progressive Matrices (APM) with 36 items aimed at the top 20 percent of adults, and the Colored Progressive Matrices used mostly with children and clinical populations. Employers almost always use the APM for graduate and professional hiring.

Each question shows a pattern with a missing piece and eight possible answers. The patterns grow harder as you move through the test. Early items test simple continuation. Middle items mix two rules at once, like rotation plus color inversion. Late items layer three or four rules across rows and columns simultaneously, which is where most candidates stall.

What Raven's measures, in psychometric terms, is fluid intelligence: the ability to spot patterns in material you have never seen before. Because the format is non-verbal, it is prized in multinational hiring where candidates do not share a first language, in military assessment centers, and in research settings that care about cross-cultural validity.

The rule families that appear on every Raven's test

There are only five or six underlying rule types. Learn them and most matrices stop feeling random.

Constant in a row

The same shape repeats across a row but changes between rows. Simplest family. Always appears in the first 10 items. Solve these in under 20 seconds each.

Quantitative progression

A feature grows or shrinks across rows or columns: dots multiply, lines lengthen, shapes duplicate. The trap is mixing up the direction of progression. Look at rows first, then verify with columns.

Distribution of three values

Three different shapes or colors each appear once per row and once per column, like a tiny Sudoku. Appears heavily in mid-test items. Always check what is missing from the incomplete row.

Figure addition and subtraction

Shapes in columns A and B combine or cancel to produce column C. Overlap, XOR, and superimposition are the three sub-variants. Train your eye to check this family whenever constant-in-row fails.

Logical operators

Late-test items use AND, OR, and XOR on shapes, colors, or orientations. The APM leans on these heavily in items 25 through 36. This is the hardest family and where top scorers separate themselves.

Two-rule combinations

A rotation plus a color swap, or a distribution-of-three plus a size change. Whenever a single rule does not explain the pattern, assume two rules are stacked and isolate each separately.

How Raven's is scored and what 'good' looks like

Your raw score is the number of correct answers out of 36 for the APM or 60 for the SPM. There is no wrong-answer penalty, so you should always guess. Pearson then converts the raw to a percentile against an age-matched norm group. Employers care about the percentile, not the raw number.

For graduate and professional roles, most hiring teams look for a 75th percentile or better. Consulting firms that use Raven's as a backup screen, along with a handful of defense contractors, push the bar to the 90th percentile. McKinsey in certain geographies has used a 95th percentile cutoff on the APM for final-round candidates, though the practice is uneven across offices.

Time limits vary by deployment. The classic APM is 40 minutes for 36 items, which is about 66 seconds per question. Some employers run an abbreviated 23-item version at 40 minutes, giving you over 100 seconds each. Pace accordingly. If you are unsure which version you are sitting, assume the strict timing.

Who uses the Raven's?

Raven's shows up in global hiring pipelines where candidates do not share a language. It is also standard in military selection and academic research programs that need a culture-neutral cognitive measure.

McKinsey (some geos)Various militaryResearch institutions

A 10-day Raven's prep plan built around pattern recognition

Day 1: Diagnostic on a full APM set

Do 36 items untimed. Do not score yourself on time, score yourself on rule identification. For every item you got wrong, write one sentence describing what the rule actually was. The list you end up with is your weakness map.

Days 2-3: Rule family drills

Work through 20 items from each of the six rule families. Constant-in-row and progression are easy warmups. Spend extra minutes on distribution-of-three and figure-addition. These are the two families where self-taught candidates stall.

Day 4: Timing calibration

Do 18 items in 20 minutes, which mirrors the APM per-item budget. Note the items where you spent more than 90 seconds. Those items are either guesses or they used a rule family you still do not recognize quickly.

Day 5: Logical operator deep-dive

AND, OR, and XOR applied to shapes and colors trip up most candidates who have not trained them explicitly. Spend 60 minutes on operator-only question sets. Draw out the logic by hand if you need to. By the end of the day you should see XOR without thinking.

Day 6: Full timed mock

36 items, 40 minutes, no pauses. Compare to Day 1. If you did not gain at least 20 percentage points, your issue is almost always rule-switching speed rather than rule-recognition accuracy.

Day 7: Review and bank examples

Re-solve every item you missed on the mock. Build a one-page cheat sheet with a canonical example for each rule family. Read it before bed.

Days 8-9: Mixed timed sets

Alternate between 12-item sprints at 12 minutes and 18-item sets at 20 minutes. The goal is to practice mode-switching under clock pressure. Keep your error rate under 20 percent.

Day 10: Rest and cheat-sheet review

No practice items. Skim your rule cheat sheet twice. Sleep 8 hours. Fluid reasoning tests punish fatigue harder than almost any other cognitive assessment.

Four Raven's mistakes that drop you two deciles

Staring at the grid before naming a rule

Most candidates try to feel out the pattern rather than actively naming which rule family is at play. If you do not form a hypothesis in the first 10 seconds, switch into rule-family checklist mode: constant, progression, distribution, addition, operator. Run the list.

Checking rows but not columns

Many matrices encode rules along both axes. Candidates who only verify rows get the top 20 items right and stall on the rest. Always confirm a candidate rule works down columns before committing.

Burning time on one hard item

APM item 28 or 31 can eat three minutes if you let it. Use a hard 90-second cap per item, then guess on the answer that follows the rule you are most confident about and move on.

Practicing on SPM when the employer uses APM

The SPM is dramatically easier. Candidates who drill on SPM items and then sit the APM are blindsided by items 20 onward. Always confirm which version you are taking and drill the correct one.

Raven's FAQs

Raven's favors the trained eye. Train yours.

Timed APM practice, rule-family drills, and percentile feedback to benchmark where you sit.

Start Raven's Practice