Free Error Checking Test

Error checking: the ability that no amount of intelligence replaces

Most ability tests ask you to work something out. An error checking test asks something narrower and, in daily working life, often more valuable: can you look at two pieces of data that are supposed to be the same and reliably notice when they are not? The reasoning load is almost zero. There is no rule to infer, no vocabulary to know, no arithmetic to perform. What the test loads instead is attention — the capacity to hold a boring comparison at full resolution for eight minutes without your eyes starting to see what they expect rather than what is on the screen. That is a genuinely separable skill, and people who are very strong on abstract reasoning are frequently unremarkable at it.

How the comparisons are built

Every item presents a reference record and a copy. A record has three fields: a short alphanumeric code, a surname in capitals, and a numeric reference. They are joined into one line so that the two records are always exactly the same length, in exactly the same order, with no shifting or reordering to complicate the comparison. That constraint is deliberate. If characters could move, counting differences would become a judgement call, and the test would stop having a single defensible answer. As it stands, a difference means one thing only: a character position where the top line and the bottom line do not agree.

Half of the twenty-four items are perfect copies. Of the remainder, half carry exactly one altered character and half carry two, placed in different fields where possible so that a reader who checks only the code and skims the rest will be caught. The substitutions favour the pairs that genuinely fool the eye — zero against the letter O, one against capital I, five against S, eight against B, and adjacent digits that share a silhouette. Nothing is hidden in punctuation and nothing depends on spacing; the traps are the ones a real clerk meets when transcribing a reference from a form.

Why three answer choices instead of two

A plain same-or-different question is easy to game. Guess at random across twenty-four items and you land near twelve correct, which makes the bottom half of the score range almost meaningless. Asking how many characters differ — none, one, or two or more — cuts random success to roughly a third and, more importantly, changes what the item demands of you. You cannot stop scanning at the first mismatch you find. Having spotted one difference, you still have to finish the line to know whether there is a second. That is exactly the habit that distinguishes a careful checker from a fast one: the discipline to complete the sweep after the interesting thing has already happened.

Reading your score

Your raw total is converted to the familiar hundred-point scale where 100 is average and fifteen points is one standard deviation. Because attention tests have a hard ceiling — a determined careful reader can get every item right — high scores compress and the interesting information sits in the middle and lower bands. Treat the table below as descriptive rather than diagnostic.

The two ways people lose points

Almost every low score on this format comes from one of two failure modes, and they need opposite corrections. The first is the miss: you read the copy as correct because your brain reconstructed the reference from memory instead of reading it. This is what makes proofreading your own writing so hard, and the fix is mechanical — force your eye down the line in fixed chunks of three or four characters rather than reading the record as a word. The second is the stall: you check every item three times, run out of clock, and leave four comparisons blank. Blank items score nothing at all, so a rushed guess on the last four is always worth more than a fourth pass over the first four. Look at your result and ask which of the two happened to you; the answer tells you what to practise.

Getting a fair result from yourself

• Use a real screen if you can. On a phone the records wrap and scroll, which adds an eye-movement cost that has nothing to do with your accuracy.
• Set a rhythm before you start. Fifteen seconds a pair leaves you slack; twenty-five does not.
• Compare field by field, not line by line — code against code, name against name, number against number. Chunking beats sweeping.
• Answer everything. There is no penalty for a wrong answer, so an unanswered item is a guaranteed loss where a guess is a one-in-three win.
• Do not copy the records out or run them through a text tool. That measures your tooling, not your attention, and the number you get back will be worthless to you.

How it relates to the other tests here

Error checking sits at the opposite end of the range from a matrix test. The culture fair IQ test gives you few items and asks a great deal of each one; this page gives you many items and asks almost nothing of each, then measures whether you can keep asking it. A large gap between the two is informative rather than contradictory. High reasoning with mediocre checking is a very common profile and describes someone who should build verification into their process rather than trust their eye. The reverse profile — excellent checking, ordinary reasoning — is exactly what a great deal of operational work actually needs, and it is chronically undervalued by people who only look at reasoning scores. Neither result is a verdict on you: like every test on this site, it is a practice estimate taken on a website, not a supervised assessment.