Logic Tests and Problem Solving

Work through eleven reasoning tests, each designed to measure a different part of how you solve problems.

Logic is the one area on this site where the test can be wrong rather than just imprecise. A reaction time is a stopwatch reading. A logic item is somebody's decision about what counts as the right answer, and that decision is the product.

So every page here shows its working. The common sense test prints the reason under each answer, and no item rests on opinion. The pattern recognition test names the rules a strip was running once you have answered.

The critical thinking test says what settles each argument. The abstract reasoning test lists all four rules a set can be built on, before you start. The matrix reasoning test puts no clock on a grid and times you anyway.

The numerical reasoning test writes out the sum afterwards. The verbal reasoning test names the line that settles a statement. The mechanical reasoning test counts the meshes for you once you have answered.

The number sequence test grows the row of numbers until you can say what comes next, and charges you for every term it added.

The last two are the odd ones out. The cognitive reflection test is three published puzzles, and it is the only page here with a curve under it. The decision making test covers its facts up and charges you time to look at them.

Ten of the eleven rank you against nobody, because no published distribution exists for those item sets. The pages say so plainly rather than borrow a curve from a test somebody else wrote.

That leaves eleven honest numbers and eleven different questions. What you knew, what you spotted, what you accepted, and what you were willing to decide without looking.

The 11 tests in this domain

Items answered correctlyCommon Sense TestAnswer everyday questions with one checkable solution, no clock or ranking, and a clear reason after every choice.%Highest level clearedPattern Recognition TestFind the hidden rule across six tiles and choose the seventh as each new level adds another rule to follow.of 6 Pull of what sounds rightCritical Thinking TestRead two statements and decide whether the conclusion follows, even when it sounds obviously wrong.points Figures placed correctlyAbstract Reasoning TestFind the figure that breaks the shared rule across twenty sets while one clock runs through the whole test.of 20 Items answered correctlyNumerical Reasoning TestRead a table of figures and answer one question at a time, with accuracy counting more than calculation speed.%Missed what the passage settlesVerbal Reasoning TestRead a passage, decide what it supports, and see how often you avoid conclusions that the text cannot settle.%Time to solve a matrixMatrix Reasoning TestComplete a nine-tile grid by following the rules that run across its rows and down its columns at the same time.s Puzzles you got rightCognitive Reflection TestTry three short puzzles designed to produce a tempting first answer, then check whether a second look changes your mind.of 3 Longest chain you followedMechanical Reasoning TestWork out which way a drive train will turn as each machine you clear adds another wheel.parts Net decisions per minuteDecision Making TestReveal what you need from a covered board, choose among four options, and stop the clock by committing to an answer.decisions/min Terms you needed to seeNumber Sequence TestStudy three numbers and predict the next one, knowing that every extra term you reveal will reduce your score.terms

Numbers and reading in this domain

ArticleDo Brain Training Games Work?See why practice improves the trained game, why broader benefits fade, and what regulators say about stronger claims.

Eleven tests, eleven numbers, no overlap

Two tests in one hub reporting the same number in the same unit are one test wearing two names. That is the rule these pages were built against, and it is the reason each of them measures something the others cannot.

Take several of them and the results will not agree, which is the point. A person who reads a hidden pattern quickly can still accept a broken argument because the conclusion sounded sensible.

Common sense: the share of everyday items you answered correctly.
Pattern recognition: the hardest rung of a six rung ladder you cleared.
Critical thinking: the gap belief opens in your accuracy, in points.
Abstract reasoning: how many of twenty figure sets you placed.
Numerical reasoning: the share of table questions you read correctly.
Verbal reasoning: how often you claimed a passage settled something it did not.
Matrix reasoning: how long one grid takes you, with no clock imposed.
Cognitive reflection: how many of three trick puzzles you got right.
Mechanical reasoning: the longest drive train you followed correctly.
Decision making: sound calls a minute, with guessing subtracted out.
Number sequence: how much of a row of numbers you needed to see.

Why almost none of these pages ranks you

A percentile needs a distribution, and a distribution needs a large sample who sat the same items. For ten of these eleven, no such sample exists.

Published research does exist near each of them. Evans and colleagues measured how belief moves the acceptance of an argument. Simon and Kotovsky measured what makes a sequence hard. Neither hands over a curve for a different item set, and borrowing one would be an invented number wearing a citation.

Numerical reasoning is the awkward one, because millions sit it every year. Those distributions belong to the publishers who own the item sets, and none of them is public.

Cognitive reflection is the exception that proves the rest. Frederick published the same three items in 2005, the wording has not moved since, and the spread of that score is in the open literature.

So the other ten report your figure and offer the arithmetic instead. Four options mean a guess is right one time in four. Five figures mean one time in five. That is a benchmark we can actually stand behind.

The thing these tests share

Every one of them can be failed by going too fast, and none of them can be passed by going faster.

The trap on the critical thinking page is a conclusion you already agree with. The trap on the pattern page is a tile that matches one you have seen. The trap on the numerical page is a figure sitting in the wrong row. The trap on the decision page is the fact you never opened.

In each case the quick answer is the one your reading produces before the checking starts. That is worth knowing about yourself, and it is roughly the only thing these pages agree on.

What a score in this hub does not mean

None of these is an intelligence test. Nothing on any of these pages has been checked against any measure of intelligence, so a claim in that direction would be made up.

Each score stays specific to its task. Repeat a test under the same conditions to compare runs.

Practice moves every one of these numbers, and what improves is the task you practised. The item pools are small enough that a second run repeats most of what you have already seen.

The first run is the one worth keeping on all eleven pages.
A wrong answer here is usually a fact or a rule you had not met.
Your own earlier runs are the only fair comparison we can offer.