Following instructions tests and exact rule use

The skill

What is a following instructions test?

A following-instructions test supplies a written procedure — conditions, actions, exceptions, priority order — and a series of cases to run through it. You are not asked whether the procedure is wise. You are asked what it produces.

Aon publishes the clearest example in Cognivy's launch scope. It describes its Understanding of Instructions module as measuring the aptitude for understanding written information and evaluating the truthfulness of statements on the basis of the given data. The candidate sees multiple text tabs containing instructions and a series of statements, and decides whether each statement is correct based on the instructional information provided.

Aon's response set is deliberately binary: every statement has two answer options, and Aon instructs candidates to select TRUE if the statement is absolutely true based on the given instructions, and FALSE otherwise. That word absolutely is the whole assessment in one adverb.

Try it before anything else

Free following-instructions practice questions

Six original questions in the formats above: routing lists, an exception with limited reach, a conditional sequence, boundary language and Aon's binary verdict. Commit to an answer before opening the solution — each solution traces the rules in order and names the exact clause every wrong option ignores.

What employers are looking at

What do following instructions tests measure?

Employers use instruction items where a documented procedure has to be applied consistently by many people: claims handling, compliance, customer operations, dispatch and routing, quality control, healthcare administration and regulated processing. The prediction being made is about consistency, not judgement.

Two abilities separate here. Precedence handling: can you tell from the wording which clause outranks which, and then apply the rules in that order rather than in the order they appear. Boundary reading: do you treat at least, no more than, up to and under as the different thresholds they are, rather than as approximate synonyms.

Reading your result

How following-instructions tests are scored

No result on this format can be read against a fixed mark, because no provider publishes one. And the response set changes what the raw number carries: with two options, the informative part of a score is which statements you could genuinely decide.

• No provider publishes a pass mark. Aon publishes timing and task counts for scales verbal instruct but no benchmark, and no provider in Cognivy's launch scope publishes one for instruction-following. Korn Ferry states directly that it sets no pass or fail marks and that benchmarks belong to the employer, which is the general position.
• Two options make a raw count less informative. On a binary response set, guessing scores far more often than on a four-option question, so a raw count says less than usual. Treat an answer you were unsure of as information about which clause you have not modelled properly, not as half a mark.
• You are not expected to answer everything. Aon states that the assessment is timed but candidates are not expected to answer all the questions, and that the aim is to get as many correct as possible in the time provided. A long deliberation on one statement earns the same mark as a quick decision and costs the next one.
• The scored standard is absolutely true. Aon instructs candidates to select TRUE only when a statement is absolutely true on the given instructions, and FALSE otherwise. A statement that is plausible but not guaranteed scores as FALSE, so marks sit in the gap between what the instructions make certain and what they merely allow.
• Early modelling time is repaid in marks. Aon states that the tab content does not change during the test. At roughly 20 seconds a statement there is no room to re-read the rules each time, so the half-minute spent building a compact model of them is recovered on every statement that follows.

Question formats

What do the questions look like?

Instruction assessments recur in a handful of shapes, and the interaction — tabs, panels, a rule sheet that stays on screen — is part of the difficulty.

• Statement verification against instructions. Aon's published format: instructions held in text tabs, and a series of statements each marked true or false according to whether the instructions make them absolutely true.
• Conditional rules. If this and this, then that. Multi-condition rules where every condition has to be met, and the case is written so that most of them are.
• Routing and classification. A record assigned to a queue, a team, a category or an approval path. The answer is the destination, and the difficulty is which rule reaches it first.
• Exceptions and overrides. A clause that carves something out of another rule. Some also outrank the rest, and the wording is what says so: regardless of any other rule, in all cases, notwithstanding the above. An exception without that wording only narrows the rule it is attached to, and leaves any separate priority order intact.
• Priority order. A numbered list where the first matching rule wins, or where a later rule applies only when an earlier one does not. Several rules matching at once is normal, not a mistake.
• Switching between rules and cases. Some interfaces keep the instructions permanently visible; others require moving between a rule panel and a question. Aon notes that in its module the correct tab for each statement is pulled up automatically and that the tab content does not change during the test.

Launch coverage

Which providers use following instructions tests?

One launch family publishes a dedicated instructions module. The same discipline is assessed elsewhere inside verbal and checking products.

• scales verbal instruct. Aon lists its Understanding of Basic Instructions module at 8 minutes and 24 tasks. The candidate is shown multiple text tabs containing instructions and a series of statements, and decides whether each is correct based on those instructions. Aon states that the tab content does not change during the test, that the correct tab is pulled up automatically for each statement, and that the response set is TRUE or FALSE with TRUE reserved for statements that are absolutely true on the given instructions.

Provider names identify the assessment format. Cognivy is independent and is not affiliated with or endorsed by these providers.

What varies, and what does not

Timing, delivery and scoring

Only one launch provider publishes figures for a dedicated instructions assessment, so the numbers below are specific rather than a range.

• Aon scales verbal instruct: 8 minutes, 24 tasks. That is roughly 20 seconds a statement, which does not leave room to re-read the rule tabs for each one, so the rules have to be modelled once and then applied.
• You are not expected to answer everything. Aon states that the assessment is timed but candidates are not expected to answer all the questions, and that the aim is to get as many correct as possible in the time provided.
• The instructions do not change during the test. Aon states this explicitly, and it has a practical consequence: time spent understanding the rules on the first few statements is reused on every later one rather than being spent again.
• Working in the given order is recommended. Aon states that candidates can navigate between statements but that it is recommended to work on the tasks in the predefined order, and its practice material notes that becoming familiar with the different sheets is part of the task.

A repeatable approach

A method that survives the clock

The rules are the same for every question, so the work is front-loaded. Build the model once, then run cases through it.

1. 1

   Rewrite the rules as conditions and actions. One line each: if X and Y then Z. A compact model built on the first statement is reused on all 24, and it is far faster to consult than the original prose.

2. 2

   Mark the exception words, then ask what each one outranks. Unless, except, only, regardless, other than, provided that. Each one restricts or overrides something, but not always everything: regardless of any other rule beats the whole list, while unless the receipt is missing only narrows the rule it is attached to. Write down the reach of each clause, not just its presence.

3. 3

   Number the priority order explicitly. Write the sequence down even when the text already numbers it, and place any clause whose wording outranks the list above it. Several rules matching one case is the normal design, and the order is what decides the answer.

4. 4

   Check the boundary of every threshold. £250 or less includes £250. Under £250 does not. Up to five means five is allowed. Boundary cases look like arithmetic and are actually reading, which is how a correct calculation still produces the wrong action.

5. 5

   Answer from the model, then sanity-check against it, not against experience. If your answer feels wrong, the useful question is which clause you applied, not what a reasonable organisation would do. Aon's instruction is that a statement is TRUE only when the given instructions make it absolutely true.

What slows progress

Common mistakes, and the fix for each

• Replacing the written rule with common sense. Fix: answer from the model you wrote, and treat disagreement with your instinct as expected. This format deliberately uses procedures that a sensible person would improve, because that is where the measurement is.
• Missing only, unless and except. Fix: mark them on the first read. Each one inverts or restricts what surrounds it, and none of them is decoration. Aon's own instruction hangs on the word absolutely for the same reason.
• Assuming an exception always wins. Fix: read what the clause claims for itself. Regardless of any other rule, in all cases and notwithstanding the above put a clause above everything, and it belongs at the top of your model however far down the page it was printed. A clause that only says unless or except narrows the rule it is attached to and changes nothing else, so a separate priority order still decides the case.
• Reading a threshold as approximate. Fix: for every limit, decide explicitly whether the boundary value is included. Or less includes it, under does not, and the case will usually sit exactly on the boundary.
• Re-reading the full instructions every time. Fix: build the compact model once. Aon states that the tab content does not change during the test, so the rules you modelled on statement one are still the rules on statement twenty-four.

In practice

How do you hold four rules in mind at once?

You do not hold them — you compress them. Prose is expensive to keep in working memory and cheap to re-read badly, so the first job is rewriting each rule as one line of condition and action: heavier than 20 kg, bay 1. Fragile and 10 kg or less, bay 2. The compressed list is short enough to hold, and short enough to consult in a glance when it is not held.

Then write the order and the reach down rather than trusting memory with either. A numbered list where the first match wins needs its numbers kept; a clause that claims precedence — regardless of any other rule — belongs at the top of the model however far down the page it was printed. Most cases in this format match several rules at once by design, so what you are really holding is not four rules but one ordered list with its exceptions already placed.

After that, each case costs almost nothing. Read it once for the trigger facts the rules actually reference — weight, marking, destination in sample question 3 — and run them down the model, stopping at the first rule that governs. Aon states that the tab content does not change during the test, so a model built on the first statement is still the rules on the twenty-fourth, and every statement it survives repays the time it took to write.

In practice

What makes exception clauses the score-killers?

Because one clause does two different jobs, and only the wording says which. An exception can narrow the rule it is attached to — unless it is heavier than 10 kg removes some parcels from one rule and changes nothing else — or it can outrank the entire set, which it announces in phrases like regardless of any other rule, in all cases or notwithstanding the above. On the page the two look alike. In the model they sit in different places, and every case that matches several rules is decided by which place that is.

Two of the free questions above are built as a pair on exactly this. In sample question 1 the security clause says regardless of any other rule, so it beats three numbered rules the ticket genuinely satisfies. In sample question 3 the unless-clause only removes its parcel from rule 2, and the case then falls through to the next rule in the ordinary order. Assume exceptions always win and you get the first right and the second wrong. Assume they never do and you make the opposite pair of errors.

The defence is clerical rather than clever. When you build the model, write down what each exception claims for itself, then position it: above the list when its wording outranks the list, inside its own rule when it only narrows one. Settled once, precedence never has to be argued under the clock — which matters at roughly 20 seconds a statement.

In practice

Why does common sense lose marks on a following-instructions test?

Because the measurement is consistency, not judgement. Employers use these questions where a documented procedure has to be applied the same way by many people — claims handling, compliance, dispatch and routing, regulated processing — and what they are predicting is whether you will do what the procedure says under pressure to do something else. A candidate who quietly improves the procedure mid-case is, for this purpose, wrong.

The conflict is engineered. Procedures in these questions are often ones a sensible person would redesign — a security clause printed after the rules it overrides, a refund path that guarantees a referral and nothing beyond it — because the gap between what is sensible and what is written is where the measurement sits. Experience makes the pull stronger, not weaker: the more familiar the scenario, the more confidently you answer from how it works in your world instead of from the rules on the screen.

Aon compresses the required discipline into one adverb: TRUE only when a statement is absolutely true on the given instructions. Not reasonable, not likely, not what a competent team would do. Answer from the model you wrote, and treat disagreement with your own instinct as the format working as designed — sanity-check against the model, never against experience.

A clear route

How to prepare in the days you have

This format rewards a habit rather than knowledge, so preparation is rehearsal of a procedure rather than revision of material.

1. 1

   First, practise modelling rule sets. Take any policy — a refund policy, a parking sign, a benefits page — and rewrite it as conditions and actions, ordered by what the wording says outranks what rather than by where each clause appeared. Do it untimed until the shape comes out the same every time.

2. 2

   Next, drill boundary language. Write out what at least, no more than, up to, under, over and or less each include and exclude. The list is short and finite, so it can be settled once rather than decided again under time pressure.

3. 3

   Then work cases against your model. Deliberately choose cases that match several rules at once, because that is what real items look like. The skill is choosing which rule governs, not spotting the one that applies.

4. 4

   Add the response set. A two-option true or false verdict rewards different behaviour from a four-option routing choice. If your invitation names Aon's instruction module, rehearse the binary decision and the standard of absolutely true.

5. 5

   Finish with the clock and the interface. Around 20 seconds a statement for a 24-task, 8-minute module. Rehearse moving between a rule panel and a question, because the switching cost is real until it is familiar.

Cognivy uses your assessment date to choose the route rather than asking you to predict a study schedule. When you sit down to practise, you choose the session length that fits that day.

Pace

Getting faster without losing accuracy

Almost all the speed in this format comes from work done before the first answer. Four habits produce it.

• Front-load the model. Spend the first thirty seconds building the conditions-and-actions list. It feels like lost time and it is repaid on every subsequent statement, because the rules do not change.
• Test the overriding clause first. Where a clause claims precedence over the rest, checking it first resolves any case it covers in one step and leaves the numbered rules unread. Establish that reach once, when you build the model, rather than re-deciding it on every case.
• Read the case for its trigger words. You need the facts the rules actually reference. Everything else in the case description is there to make it read like a real situation.
• Decide and move on a binary response. With two options and a per-paper clock, a long deliberation earns the same mark as a quick one and costs the next statement. Aon tells candidates they are not expected to answer everything.

Keep exploring

Related test types and how they differ

Following instructions borders three test types that share its reading demand.

Direct answers

Following instructions test FAQs