You have hit a wall of shapes on your CCAT practice test: five little figures, a question mark, and 18 seconds to figure out what comes next. Abstract reasoning (Criteria Corp calls it spatial reasoning) is the CCAT section most people freeze on, because unlike a word analogy or a math problem, there is no obvious place to start. Here is the good news up front: these questions are not random. Every single one is built on a small, fixed set of pattern rules, and once you can name the rule fast, the answer usually falls out in a few seconds. This guide gives you the three question sub-types, the six rules that generate almost all of them, worked examples for each, and a pacing plan so this section stops eating your clock. Quick note before we start: this is the Criteria Cognitive Aptitude Test for hiring, not the Canadian Cognitive Abilities Test for schoolchildren that shares the acronym.
Quick takeaways
- Abstract reasoning and spatial reasoning are the same CCAT section. Criteria Corp describes it as rotating or flipping images in your head, recognizing patterns, and identifying outliers.
- There are three sub-types: next in series, odd one out, and matrices.
- Almost every question runs on one of six pattern rules: rotation, reflection, movement, adding or removing elements, shading or color, and size.
- Prep sites estimate about 11 of the 50 CCAT questions are spatial. Criteria Corp does not publish an official count, so treat 11 as an estimate.
- You can skip these. There is no penalty for wrong answers, so if a shape question is stalling you, guess and move on.
- The fastest solvers check one attribute at a time instead of trying to see the whole pattern at once.
Abstract reasoning vs spatial reasoning: same thing
First, a naming knot worth untying. You will see this CCAT section called “abstract reasoning,” “spatial reasoning,” “inductive reasoning,” and sometimes “logical reasoning.” On the CCAT, these all point to the same set of shape-and-pattern questions. Criteria Corp’s own materials use “spatial reasoning” and describe it as your ability to “rotate or flip images in your head, recognize patterns, and identify outliers.” The wider aptitude-test world often labels the identical skill “abstract reasoning.” Do not let the different labels make you think you are missing a section. It is one thing.
Why do employers include it? Because pattern recognition with no verbal or numerical crutch is a fairly clean read on raw problem-solving and how quickly you pick up new rules. That is exactly what a company screening for fast learners wants to see.
How many abstract reasoning questions are on the CCAT?
Criteria Corp does not publish the exact count of each question type, so anyone stating a hard number is estimating. That said, the widely used prep-site estimate is about 11 spatial questions out of the 50, or a bit over 20 percent, often split roughly as four “next in series,” four “odd one out,” and three matrices. Treat that as a working estimate, not gospel.
The practical takeaway is not the exact count. It is that spatial questions are a meaningful chunk of the test (roughly one in five), they are mixed in with verbal and math rather than grouped, and they tend to be where people lose the most time. So having a fast, repeatable method for them pays off more than for almost any other section.
The three question sub-types
Every CCAT abstract reasoning question is one of these three shapes. Learn to recognize which one you are looking at instantly, because each has its own attack.
| Sub-type | What you see | What you do |
|---|---|---|
| Next in series | Five figures forming a sequence, with a missing slot | Find the rule driving the sequence, then pick the figure that continues it |
| Odd one out | Five figures, four of which share a rule | Find the rule the four share, then pick the one that breaks it |
| Matrices | A 3x3 grid with one cell missing | Find the rule across each row and down each column, then pick the missing cell |
The trap in all three is trying to see the whole pattern at once. Do not. Check one attribute at a time (shape, then rotation, then shading, then count) and the pattern reveals itself far faster.
The six pattern rules that generate almost everything
This is the part competitors dance around but never lay out cleanly. Nearly every abstract reasoning question is built by applying one or two of these six rules across the figures. Memorize them, and your job on each question becomes “which of these six is happening here?”
- Rotation. A shape or element turns a fixed number of degrees each step (for example, an arrow rotating 90 degrees clockwise per figure).
- Reflection (flip/mirror). A figure flips across a horizontal or vertical axis, producing its mirror image.
- Movement (translation). An element slides position step by step, for example a dot moving one corner clockwise around a square each figure.
- Adding or removing elements. The count of items goes up or down by a fixed amount, for example one extra line added per figure.
- Shading or color. Fills alternate or progress: empty, half-filled, full, or a color rotating through the figures.
- Size. An element grows or shrinks by a consistent step.
Most questions combine exactly two of these (say, rotation plus shading). Hard ones combine two subtle ones. Easy ones use just one. When you scan a question, run down this list mentally and the driver almost always jumps out.
Worked example 1: next in series
Imagine a row of five boxes. In each box there is a single straight arrow, and it points: up, then up-right (rotated 45 degrees clockwise), then right, then down-right. The fifth box has a question mark. Your answer options show the arrow pointing in various directions.
Run the one-attribute-at-a-time method. Shape? Same arrow throughout, so that is not the rule. Rotation? Yes: each step turns the arrow 45 degrees clockwise. That is the single rule driving this. Continue it one more step from down-right and you get an arrow pointing straight down. Pick the option with the arrow pointing down. Done in under 10 seconds once you spotted “rotation.”
The reason people miss this one is they stare at the whole arrow trying to intuit the answer instead of testing “is it rotating?” against rule number one on the list.
Worked example 2: odd one out
Now imagine five figures, each a circle containing some smaller shapes. Four of the circles contain an even number of small shapes (two, four, four, six). The fifth contains three. Everything else (circle size, shape type, shading) looks similar and is designed to distract you.
Attack it as “find the shared rule, then find the breaker.” Check attributes one at a time. Circle size? All similar, not the rule. Shape type inside? All the same, not the rule. Count of inner shapes? Four figures are even, one is odd. That is the rule the four share (even count), and the odd-numbered one breaks it. Answer: the figure with three shapes.
Odd-one-out questions almost always hide the real rule behind two or three decoy attributes that all match. Your job is to keep checking attributes until one splits the group four-versus-one.



