Fault Diagnosis: Finding the Cause Without GuessingFault Diagnosis. The question a hiring manager actually has is whether this person can find a fault they have never seen before.
Fault diagnosis is the skill of narrowing an unknown cause: forming a hypothesis before testing, buying the information that splits the possibilities, knowing when you have enough, and handing the problem on properly. It is measured here on four fictional faults that need no technical knowledge at all.
Two readings that come apart, and the way they come apart is the finding
Four faults, four rungs each. A shared printer that returns blank pages, but only some of them and only since yesterday. A tracking screen showing parcels as delivered that never arrived, on one route. Hot water that fails on the top floor each morning and returns by mid-morning. A file that rejects about one row in twenty with no message saying why. At every rung you are told what is now known and you choose the next move, and every move costs something and returns something.
Your correctness and your efficiency are reported as two separate numbers and never averaged, because they come apart in four different ways and each one is a different person to work with. Right and cheap. Right by the long route. Cheap and on the wrong path. Still searching. A single score would hide all four. The efficiency figure compares what your route cost with the cheapest fully correct route and with what the typical respondent's route would cost, so it is not a reward for doing less: the full-credit move is the cheapest option at only five of the sixteen rungs, and at six of them the cheapest move is worth nothing.
The ladder does not branch, and the report says so on the page rather than leaving you to assume. Every respondent meets the same authored position at every rung, which is exactly what makes it safe: a fixed position can be checked to contain a genuinely good move, and all sixteen do. Nobody can be handed a place with nothing sensible left to do and then told it was their own doing. Eight further situations and four selection exercises cover the fix that worked without anyone knowing why, the colleague who has already decided what it is, and the check that would have come out the same either way.
What you walk away with
Two chance-corrected figures with their bands, never averaged, because a cheap wrong route and an expensive right one are different people and a single number hides both.
One row per capability, two dots joined by a line, sorted so the largest disagreement between the two readings sits at the top. No radar; the shapes differ so it survives greyscale.
What you did at each of the sixteen rungs, what it was worth, what it cost, and what the cheapest fully correct move would have cost.
How often you chose a check that could have come out the other way, against the ones that would have confirmed what you already knew.
Lands it cheaply, gets there the long way, moves cheaply to the wrong place, or still searching — with what to do about the one you got.
Inside your report
Illustrative sample - your report is generated from your own responses.
Sorted by the size of the gap. The distance between the two dots is the finding, and no radar can show it.
Rung two went wrong and rung three was full marks from where that left you. That is recovery credit, drawn.
Built for
- Employers hiring support, maintenance, operations, quality and analyst roles where the job is finding out what is wrong
- Technicians, analysts and support staff who want a portable measure of how they diagnose rather than what they know
- Training providers and apprenticeship programmes teaching troubleshooting, where nothing currently measures the reasoning
Find out how you narrow down a cause you have never seen before
32 scored exercises - about 40 minutes - a full bespoke report with correctness and efficiency reported apart, a gap dumbbell, and every rung with what it cost.
₹749 (incl. GST) · assessment and full report, nothing further to pay
Frequently asked questions
How a person narrows down an unknown cause: whether they form a hypothesis before testing, buy the information that actually splits the possibilities, know when they have enough to act, and hand the problem on with what they eliminated. It does not measure technical knowledge of any system, any product or tool, how fast someone works, or general intelligence.
None. All four faults are fictional and ordinary - a shared printer, a tracking screen, hot water on a top floor, a file that rejects rows - and every option is about diagnostic strategy rather than about a technical fact. That is what makes it usable across industries and what stops it becoming a test of who has met this equipment before.
Because they dissociate, and the dissociation is the insight. Somebody who reaches the right cause by an expensive route and somebody who moves cheaply down the wrong path would score identically on a combined figure, and they need opposite advice. The report names the four patterns, prints both numbers with their bands, and sorts the capability rows by how far the two readings disagree.
No, and the report says so in plain words rather than letting you assume otherwise. Every respondent meets the same authored position at every rung. That is the safe design on a fixed form: an authored position can be checked to contain a genuinely good move, and all sixteen do, so nobody is handed a state with nothing sensible left to do and then told it was a consequence of their own earlier choice.
Rs 749 in India including GST, or US$7.99 elsewhere, one time, for one full sitting and report. Pre-employment suites carrying anything comparable are subscription-only and start in the hundreds of dollars a month with no single-test purchase; the free troubleshooting quizzes online test product knowledge. Organisations can use AssessAll credits at 25 credits per person.
Each one takes a single capability, puts you inside the situations where it is actually tested, and scores your choices against published evidence — with a report designed for that capability alone, not a template. They span hiring, compliance, education, operations and personal skill.
Browse the catalogue →Methodology: Measures applied diagnostic judgment through original ladder, situational and selection items keyed to published construct areas. Construct statement: it measures how a person narrows down an unknown cause, whether they form a hypothesis before testing, buy the information that actually splits the possibilities, know when they have enough to act, and hand the problem on properly when it is not theirs, and it does not measure technical knowledge of any system, any product, any tool, speed, or general intelligence. Declared response instruction: behavioural tendency throughout, every stem asks what the respondent is most likely to do next. Item format: four self-contained fictional faults, each worked through four fixed-state rungs. The state at each rung is authored and identical for every respondent, so the instrument does not branch and no rung is read as a consequence of an earlier answer; every rung is authored to carry at least one move that earns full credit, which makes it impossible to reach a position with no good move left. Every move carries an authored cost in time, disruption and other people's attention, and the full-credit move is deliberately not the cheapest one on most rungs, so that always choosing the cheapest move cannot produce a good result. Scoring design: two independent readings. The correctness figure comes from the graded value of each move, chance-corrected against the authored per-option choice rates, so that a move most people pick counts for less than one they do not. The efficiency figure normalises the cost the respondent actually incurred between the cost of the cheapest full-credit path and the cost expected from a respondent choosing at the authored rates, so zero means no better than the way people typically choose and one means the cheapest correct path. The two are reported side by side and never averaged. Construct areas drawn on: hypothesis-driven versus exhaustive search in troubleshooting; expected information gain and the value of a test that splits the possibility space; confirmation bias and positive-test strategy; the availability habit of replacing whatever failed last time; premature closure and anchoring in diagnostic reasoning; the distinction between correlation in a record and cause; work on the cost of not yet knowing set against the cost of testing; satisficing and stopping rules under time pressure; research on escalation and handover quality showing that the handover is where a diagnosis is most often lost; the sunk-cost effect in persisting with a hypothesis that has been discarded; and calibration research on how confident diagnosticians are against how often they are right. All items are original works, all four faults are fictional, no product, tool, vendor or trademarked instrument is named or reproduced, and no affiliation with any source is claimed. AssessAll original design.