Raven’s Progressive Matrices Test: A Recruiter’s Guide 2026

A candidate can score in the 90th percentile on the Raven’s Progressive Matrices test and still wash out in the first 90 days. A 2022 reanalysis of decades of hiring data cut this test’s predictive validity nearly in half.

Most hiring guides treat a single cognitive score as a hire-or-pass line, which is exactly how strong candidates get filtered out and weak ones slip through anyway. The real skill is knowing what this test actually measures and how much weight one number deserves.

This guide breaks down what the test measures, how to score and interpret a result correctly, and where it fits inside a fuller hiring signal instead of standing alone. Get the mechanics right and the score becomes a genuine time-saver, not a liability.

Summarise this post with:

Chatgpt Gemini Claude Grok Perplexity

What is the Raven’s Progressive Matrices test?

Raven’s Progressive Matrices is a multiple-choice test of fluid intelligence, the ability to reason through new problems without relying on learned facts. Each item shows a grid of shapes with one piece missing, and the candidate picks the option that completes the pattern.

The psychologist John C. Raven first published the test in 1938, and it still holds up because the format never depended on any single era’s curriculum. Because the questions use shapes instead of words, the test does not reward vocabulary, reading speed, or a particular school system.

That is why it travels well across regions and first languages, and why it sits in the same family as an abstract reasoning test. It reads one slice of ability well: can this person see structure in something unfamiliar and work out what comes next.

How is the test structured?

The test comes in three versions, each tuned to a different ability level. Standard and Advanced are the ones used in hiring.

The Coloured version is built for children and clinical assessment, so it rarely belongs in a recruitment process. The published item counts and time limits for each version are below.

What does a test item actually look like?

Here’s what a typical Standard-version item looks like, a grid of shapes with the bottom-right piece missing:

How do you score and interpret the test?

Scoring starts with the raw count of correct answers, which is then converted to a percentile against a norm group, a comparison population of people who took the same test. A 75th-percentile result means the candidate scored higher than 75 percent of that group.

The percentile, not the raw number, is what you read. Two more interpretation calls shape how you use it.

What interpretation calls affect the result?

The first is timed versus untimed: a timed run reads reasoning under pressure, while an untimed run reads reasoning depth. Choose the format that matches what the job actually demands.

The second is how wide a gap you treat as meaningful. A candidate at the 80th percentile and one at the 70th are close enough that other evidence, not the 10-point spread, should decide between them.

Benefits of Raven’s Progressive Matrices test

Used well, the test earns its place in a hiring process for four specific reasons. Each one solves a different problem most screening methods run into.

Language-neutral

A verbal reasoning test rewards vocabulary and reading speed in whichever language it is written in, which quietly favors native speakers. Raven’s uses shapes instead of words, so a candidate’s first language never enters the score.

Reads learning speed and adaptability

Most roles do not stay static for long, and how fast a hire picks up what the job throws at them next matters more than what they walk in already knowing. Raven’s reads that adaptability directly, which is why it holds up well for roles built around change rather than repetition.

Quick to administer

A 40-minute session fits inside a single screening step without adding a second scheduling round. That speed matters most at high volume, where even a small add-on per candidate compounds into real recruiter hours across hundreds of applicants.

Hard to game

There is no manual, glossary, or crash course that raises a score overnight, since the test measures reasoning in the moment rather than recalled information. That makes it harder to fake preparation and easier to trust that the result reflects the candidate’s actual thinking.

Limitations of Raven’s Progressive Matrices test

The test measures one thing, and the common mistake is treating it like it measures everything. Here’s what a high score stays silent on.

Says nothing about hands-on skills for the role

A strong reasoning score does not confirm the candidate can run the specific tools, workflows, or systems the job requires right now. Pair it with a role-specific skills test if hands-on competence is what actually decides performance.

Silent on employee motivation

Two candidates can post identical scores while one is genuinely invested in the role and the other is simply a strong test-taker. Only a structured interview or work sample surfaces that difference.

Doesn’t read communication or conflict handling

Abstract reasoning happens on a screen with no audience, so it says nothing about how a candidate explains a decision or takes feedback in a room. Roles built around teamwork need a separate, direct read on those behaviors.

Blind to integrity and reliability

A pattern-matching task cannot surface whether a candidate follows through, tells the truth under pressure, or shows up when it counts. That gap is exactly why a single cognitive score should never be the final gate.

Why the validity number matters

Cognitive ability still predicts performance, but by less than recruiters have assumed for years. A 2022 reanalysis by Sackett and colleagues corrected the long-cited validity figure from 0.51 down to roughly 0.31, and ranked structured interviews among the strongest standalone predictors.

The stakes for getting this wrong are real. Gallup puts U.S. voluntary turnover costs at roughly a trillion dollars a year, with replacement running one-half to two times the departing employee’s salary, and the U.S. Department of Labor estimates a bad hire can cost up to 30 percent of first-year salary on top of that.

Strong test-takers who freeze under a timer can underperform their real ability, and a candidate who has practiced dozens of pattern sets can beat one who has not, even with similar reasoning. For hands-on or relationship-heavy roles, abstract reasoning is a weak proxy on top of everything above, so weigh it accordingly or skip it.

How do you combine it with other tests?

Treat the matrices score as one input, not the answer. Pair it with two or three other reads of the candidate so no single signal carries the hiring decision alone. Three additions cover what the matrices test cannot:

• Add a role-specific skills test to confirm the candidate can do the actual work, not just reason abstractly.
• Run structured interviews to read judgment, communication, and how they explain a decision.
• Use a work sample, where the role allows it, to see real output under job-like conditions.

Within Testlify’s cognitive ability tests and broader psychometric tests, the matrices result becomes context for a decision rather than the decision itself. That framing keeps the score useful without letting it override everything else you know about the candidate.

What does this look like in a hiring funnel?

Picture a 300-person software company hiring data analysts and drowning in 400 applications per role.

1. Run the Standard matrices test on every applicant to sort for the abstract reasoning the job leans on, trimming the pool to a workable shortlist fast.
2. Send shortlisted candidates a role-specific SQL and analytics test.
3. Interview the final few with a structured interview scored against a fixed rubric.
4. Let the matrices result open the funnel while the skills test and interview close it, so a fast screen never turns into a careless one.

Best practices for creating fair assessments

Fair, useful results come down to five practices, and skipping any one of them reopens the door to bias or a bad read. Here’s what each one looks like in practice.

Set expectations before the assessment

A one-line note that the test is timed, nonverbal, and not about prior knowledge cuts candidate anxiety before it starts. It also heads off the unfairness that comes from candidates guessing at the rules mid-test.

Pick the version that reflects job complexity

Standard suits most positions, while Advanced is built for senior or analytical roles where a ceiling effect would otherwise erase differences between strong candidates. Matching version to seniority keeps the score meaningful instead of flattened at the top.

Use time limits only when they reflect the job

A timed run reads speed under pressure, and an untimed run reads pure reasoning depth. Choose whichever mirrors what the actual job demands, not whichever setting happens to be the default.

Interpret every score against a matching norm group

A percentile only means something relative to the comparison population behind it, so a generic norm group can flatter or punish a candidate for no real reason. Pick the norm that reflects the role and the applicant pool, not a one-size-fits-all default.

Never auto-reject on the score alone

Use the result to shape what you probe for in the interview, not to eliminate a candidate before anyone talks to them. The final call belongs to the fuller picture, not to one number.

How do you put this into practice?

Testlify combines the Raven’s Progressive Matrices test with role-specific assessments, conversational AI interviews, and AI insights to give you a complete view of every candidate

Start a free trial to build your first assessment, or book a demo to see how Testlify brings every hiring signal together in one place.

Frequently asked questions (FAQ)

Yashika Khandelwal

Linkedin profile

Content Writer

Yashika Khandelwal is a Content Writer with 3+ years of experience creating research-backed content on hiring, talent assessment, and HR technology. She is a registered Organizational Psychologist and subject matter expert who combines behavioral science with practical recruitment insights to produce accurate, evidence-based content.