SHL Error Checking & Clerical Speed Tests: Complete Guide

Error checking tests — also called "checking tests" or "clerical speed tests" — assess a candidate's ability to rapidly compare two sets of data and identify discrepancies, or confirm matches. They measure accuracy and speed simultaneously, placing candidates under time pressure to reveal how they perform when pace is critical and concentration must be sustained.

These assessments are most common in roles where data entry, verification, compliance, back-office processing, or quality control is a regular part of the job. In these environments, a single error can have significant downstream consequences — so employers want objective evidence that candidates can work quickly without sacrificing accuracy. Error checking tests provide exactly that evidence in a standardised, comparable format.

The SHL Checking Test (formerly part of the SHL Clerical Test series) is the market-leading version used by banks, government bodies, and large corporate employers worldwide. It is delivered online via SHL TalentCentral and is typically administered early in the recruitment process alongside, or as a standalone alternative to, the numerical and verbal reasoning assessments used for more analytical roles.

SHL's Verify Checking Test presents two columns of information — typically alphanumeric codes, product codes, names, or reference numbers — and asks you to identify whether the entries in the two columns match or differ. For each row, you make a binary decision: match or no match. The challenge is doing this accurately across 30–40 items within a 5–7 minute window.

The format is deceptively straightforward. Most candidates can complete the task correctly when given unlimited time — the test's discriminatory power comes from the combination of speed and accuracy under pressure. The scoring accounts for both dimensions simultaneously, and wrong answers penalise your score through negative marking. This means a candidate who rushes and makes multiple errors can score lower than one who works more slowly but maintains near-perfect accuracy.

The test is administered online via SHL TalentCentral, typically within a 24–72 hour window from the invitation email. Some employers administer a supervised re-sit at an assessment centre to verify online scores.

Broader clerical assessments go beyond error checking to include a wider range of perceptual and administrative accuracy tasks. These may be combined into a single battery or administered as standalone tests depending on the role and employer. Common clerical subtests include: filing and alphabetical ordering, name and address checking, coding (matching codes to categories from a reference key), and data entry accuracy under timed conditions.

Some employers use the full SHL Clerical Assessment Battery (CAB), which combines multiple clerical subtests into a single extended assessment session. This provides a more comprehensive picture of a candidate's clerical aptitude profile across different task types. Other assessment publishers — notably Saville Assessment and Korn Ferry Talent Q — offer their own versions of clerical speed assessments with comparable formats and difficulty levels.

Although the specific presentation varies by publisher, the core skill being assessed is consistent across all versions: perceptual speed, accuracy under time pressure, and the ability to sustain focused attention over a rapid sequence of comparison tasks. Preparation strategies that work for the SHL Checking Test transfer directly to these alternative formats.

Error checking tests are most commonly used by employers in industries where accuracy in data handling is operationally critical. The primary sectors include: banking and financial services (processing, operations, compliance, and trading support), civil service (data verification and administrative grades), insurance (claims processing and underwriting support), logistics and supply chain, healthcare administration, and accounting support roles. In all of these environments, an employee who makes frequent data errors causes real costs — financial, regulatory, or operational — so employers invest in pre-hire assessment to screen for the trait.

Employers Known to Use Error Checking Tests

• Banking & financial services: Barclays (operations and support roles), HSBC (back-office, transaction banking), Lloyds Banking Group (operations track graduate programmes), NatWest, Standard Chartered.
• Civil service: HMRC (data processing and administrative grades), DWP (benefits processing roles), Cabinet Office administrative grades, NHS Business Services Authority.
• Insurance: Aviva, Legal & General, and Zurich for claims processing and underwriting support roles.
• Healthcare administration: NHS trusts with large administrative teams; NHS Shared Business Services.
• Logistics & supply chain: Major logistics operators and retailers with extensive stock control or order processing functions.

Error checking tests use three main format variants. Each requires a slightly different visual scanning approach, but all are scored on the same match/no-match binary decision. Understanding each format before your test means you won't waste time adjusting to an unfamiliar layout on the day.

Format 1 — Simple Alphanumeric Comparison

Two codes are displayed side by side. Your task is to determine whether they are exactly identical — including capitalisation, spacing, and the distinction between visually similar characters like the letter O and the digit 0, or the digit 1 and the letter l.

Format 2 — Name & Address Checking

Two address blocks or contact records are displayed side by side. You must identify any discrepancy — a missing character in a name, a transposed digit in a postcode, or a punctuation difference. These items test your ability to compare longer strings without losing focus at the character level.

Format 3 — Code-to-Category Matching

A reference key maps codes to categories. You are shown a code and a category label; your task is to verify whether the label matches the key entry for that code. This format adds one additional cognitive step — looking up the key — which increases the working memory demand compared to simple side-by-side comparison.

For example: given the key A4 = Finance, B7 = Operations, C2 = Compliance, you might be asked to verify whether the entry “C2 — Operations” is correct. The answer is No (C2 should be Compliance). The speed challenge comes from needing to scan the reference key quickly and accurately for each item.

Your raw score is the number of correct answers adjusted downward for incorrect answers using a correction-for-guessing formula. The exact deduction varies by publisher, but a typical structure deducts a fraction of a mark for each wrong answer — so random guessing across all items is expected to produce a near-zero adjusted score. This is why confident, considered decisions are critical: it is better to skip an uncertain item than to guess randomly, but better to answer when you have a genuine basis for your decision even if not fully certain.

The adjusted raw score is then converted to a normed percentile score compared against the relevant comparison population — typically administrative or finance sector professionals at a graduate level. The report employers receive typically shows three key metrics: overall accuracy percentage, processing speed (items completed per minute), and the normed percentile score relative to the comparison group.

Employers see a single overall percentile score on the summary report, along with a traffic-light or band indicator (typically: Below Expectation / Meeting Expectation / Exceeding Expectation). Unlike reasoning tests, where there is wide variance in scores, most candidates answer the majority of items correctly — the key differentiator on error checking tests is speed without accuracy loss. Candidates who rush and make errors score below slower-but-accurate candidates, even if the faster candidate completed more items overall.

Cut scores vary by employer and role. For administrative and data-processing roles, a typical threshold is around the 50th–60th percentile. For more competitive positions such as civil service fast stream or operations roles at major banks, the bar is typically the 65th–75th percentile. Unlike numerical reasoning, very high percentile scores (90th+) are less common because the test is narrow in scope and most candidates cluster in a tighter band once adequately prepared.

Error checking performance is highly trainable. Unlike abstract reasoning, where gains from practice plateau quickly, perceptual speed and accuracy continue to improve with targeted practice over a 2–3 week window. The following eight strategies are based on established cognitive psychology of perceptual learning and feedback from candidates who have significantly improved their scores.

• Don't rush the first 5 items. The opening items are your calibration window. Establish your rhythm and scanning pattern before you accelerate. Candidates who rush immediately from item one typically make early errors that cost disproportionately under negative marking.
• Use a fixed scan pattern — character by character, left to right. Consistent scanning prevents your eye from skipping sections. Do not attempt to take in the whole code at a glance; this increases error rate on items with subtle single-character differences.
• Chunk long codes into 3–4 character segments. When comparing alphanumeric codes of 8–12 characters, mentally divide each code into chunks before comparing. For example: “GH29 | 47-A | X” rather than trying to hold the entire string in working memory simultaneously.
• Pay extra attention to the END of codes. Errors cluster in the final 2–3 characters of codes and addresses. After the first few characters match, attention naturally relaxes — test designers know this and embed errors accordingly. Make a conscious habit of double-checking the last segment of every item.
• Watch for visually similar character pairs. The most common single-character traps are: O (letter) vs 0 (digit), I (capital i) vs 1 (one) vs l (lowercase L), rn (two letters) vs m (one letter), cl vs d, and vv vs w. Train pattern recognition for these pairs specifically.
• Skip rather than guess randomly. Under negative marking, a blank item scores 0. A randomly guessed wrong answer reduces your score. Skip items where you have no basis for a decision. However, if you can rule out one option, the expected value of answering is positive — so don't skip indiscriminately.
• Target 6–8 seconds per item maximum. This is the pace required to complete a 40-item test in 5 minutes with time to review. Practise with a timer from day one. Most candidates who score below their target have a pacing problem, not an accuracy problem — they slow down mid-test and run out of time.
• Do 20 minutes of focused practice daily for 2 weeks before your test. This duration is sufficient to build the neural pattern-matching pathways that underpin perceptual speed. Longer sessions produce diminishing returns and fatigue. Consistency across days matters more than session length.