Workplace Listening Assessment for Employees, Managers and Support TeamsWhat you took away, against what was actually said. Two figures, never added.
Forty short exercises about listening at work. Eighteen are spoken excerpts, read aloud by your own device, after which you sort a handful of statements three ways: the speaker said it, only implied it, or neither. Eight are spoken messages with written answers that differ only in which detail reached you; the rest are written exercises on what a message said, asked, or left open. The report is a warrant table: one row per part, with what it carried, its reliability, its result or the printed refusal of one, and a plain-words column saying what that row lets you claim and what it does not. At the top sit two figures that are never averaged: what you caught, and what you kept. Free to sit; the report is the product.
The first listening assessment that scores the two mistakes apart
The Workplace Listening Assessment is a thirty-minute performance-based assessment for anybody whose work runs on conversations, calls and messages. It measures the gap between what a speaker said and what the listener took from it, scored as two directions that are never added, and prints what each row of its report lets you claim.
Free listening quizzes ask what you know about paraphrasing and barriers, with no audio, and return one score on remembered advice. Self-report listening scales ask how well you think you listen, and the published evidence is that self-reports of listening converge only weakly with how the speaker or an observer saw the same conversation. This assessment plays you speech and scores what you took from it. Nothing on it is self-report, and nothing on it is keyed on knowing a word.
In a conversation, two different things can go wrong. You can take on board something that was never said: an implication heard as a statement, a plan built on a promise nobody made. Or you can lose something that was said: a commitment demoted to a maybe, a condition dropped. A single listening score averages the two, which makes a listener who takes everything on board and a listener who doubts everything look the same. This report prints caught and kept as two figures, each with its 68 and 95 per cent bands, and names the shape of the pair in a phrase you would say out loud.
The eighteen spoken excerpts are short and in plain words: a team lead moving a call, a colleague covering a shift, finance explaining a payment condition, a manager who liked a proposal but did not approve it. After each, you sort four or five statements as said, implied, or not said, with three labelled buttons and no dragging. Eighty-one statements in all, forty-one of them genuinely said, so that doubting everything cannot win. Every statement carries a declared share of ordinary listeners who sort it correctly, and a subtle one is worth more.
The sort is ordinal and the scoring uses that: calling an implied statement said is one step, calling a never-said statement said is two, and the second is priced at three times the first, with the ratio printed above the figures so you can disagree with it. Eight spoken messages then test which detail survived the gap between hearing and acting: a condition, an ask that was never put as a request, a date nobody mentioned. Six select-every-line exercises, four matching exercises and four ordering exercises cover the same four parts in writing: said against inferred, the commitment and its condition, the ask underneath the words, and what was not said.
Every spoken excerpt has a written equivalent. It appears on its own when your device cannot read aloud, and on request on any device. An exercise that reached you in writing is reported in its own row, apart from the listening figures, and never counts against you: which exercises were spoken to you is a fact about the device, not about your listening.
Below the two figures: the warrant table itself, with every refusal printed where a figure would have gone; the step-against-leap count; the four parts with every exercise listed; what zero means against the declared priors; the stronger side paired with what it costs; one boxed if-then change; and the movement a re-sitting in four weeks would have to clear to count as real. At Rs 249 the sitting is free and the report is the product.
What you walk away with
What you caught: the things that were only implied or never said, not taken on board. What you kept: the things that were said in words, not lost. Each is chance-corrected so that zero is the ordinary listener, each carries its 68 and 95 per cent bands drawn and written, and the two are never added, because a listener who takes everything on board and a listener who doubts everything would average to the same number.
The habit is read off the shape rather than the level: I add things that were not said; I drop things that were said; I keep what was said and add nothing; I both add and drop. A shape is named only when the bands are clear of the ordinary listener, and refused otherwise.
One row per part: the exercises it carried, its share of the headline and why, its assumed reliability, its result or the printed refusal of one, and what that lets you say, with the half a certificate would leave you to assume printed beside it. There is no total down any column and no percentile anywhere.
The sort is ordinal, so the report counts one-step slips (said against implied) apart from two-step slips (said against not said), prices the leap at three steps, and prints the ratio above the figures where it can be disagreed with.
A spoken exercise that reached you in writing, because the device could not speak or you asked for the words, leaves the listening figures entirely and is reported in its own row. Nothing about taking that route costs you a point.
A boxed if-then sentence drawn from the lower of your two figures, naming a specific situation and a specific replacement behaviour. Beside it, a re-sitting date four weeks on and the movement each figure would have to clear to count as real change rather than measurement error.
Inside your report
Illustrative sample — your report is generated from your own responses.
Distant slip to adjacent slip, 3 to 1. An implied statement called said costs a third; a never-said statement called said costs the whole.
A value judgement, printed where it can be disagreed with. Nothing below averages the two figures into one.
I keep what was said and add nothing
Both 68 per cent bands sit clear of the ordinary listener. A shape is named only when they do; otherwise the page says the side cannot be told from ordinary.
About two chances in three that a re-sitting would land between 30 and 57; nineteen in twenty between 17 and 70.
Contributors: 29 lure statements as keyed, 11 slipped, weighted by how few ordinary listeners get each right. The furthest slip: “The room has been booked for the whole afternoon”, keyed not said, pressed said: two steps, weight 0.66.
About two chances in three that a re-sitting would land between 11 and 48; nineteen in twenty between -7 and 67.
Contributors: 33 genuine statements as keyed, 8 slipped, weighted by how few ordinary listeners get each right. The furthest slip: “The desk plan can be found on the shared drive”, keyed said, pressed not said: two steps, weight 0.43.
These two figures are never added, averaged or netted. A listener who takes everything on board and a listener who doubts everything would average to the same number, and that number would be false for both.
| Part | Exercises it carried | Share, and why | Omega, SEM | Result, or the refusal of one | What this lets you say — and what it does not |
|---|---|---|---|---|---|
| Caught | 40 heard statements keyed implied or not said, in 18 excerpts | One of the two headline figures, whole; never added to kept. | 0.94, SEM 13 | +44; 68% 30 to 57; 95% 17 to 70 | You can say you did not take on board things that were never said, in eighteen short excerpts heard at most twice. You have not shown you can hold that line in a live conversation with a stake in it. |
| Kept | 41 heard statements keyed said, in the same excerpts | The other headline figure, whole; never added to caught. | 0.95, SEM 19 | +30; 68% 11 to 48; 95% -7 to 67 | You can say that what was committed to in words reached you and stayed. You have not shown you would act on it a week later. |
| Said against inferred | 10 of 10: 6 heard, 4 read by design, 0 read by device | None of the headline; a part on its own. | 0.81, SEM 16 | +45; 68% 30 to 61 | You can say you can tell a statement from your own inference when the two are put side by side. You have not shown you notice an inference while a conversation is still moving. |
| The commitment and its condition | 10 of 10: 6 heard, 4 read by design, 0 read by device | None of the headline; a part on its own. | 0.81, SEM 18 | +7; 68% -11 to 25 | You can say you heard what was committed to and the condition attached. You have not shown you can hold a conversation under pressure, or that you would chase the condition before the date. |
| The ask underneath the words | 7 of 10 answered: 5 heard, 2 read by design | None of the headline; a part on its own. | 0.75, SEM 24 | Refused: near the ordinary listener, untested. Only 7 exercises answered, under the 8 a figure needs. | Nothing on this row: the figure is refused, and the placement word is printed where it would have gone. You have not shown you would answer the ask, or read a hint from a speaker this instrument did not sample. |
| What was not said | 10 of 10: 7 heard, 3 read by design | None of the headline; a part on its own. | 0.81, SEM 21 | +38; 68% 18 to 59 | You can say you notice the date, the owner or the decision a short message left out. You have not shown you would ask for it. |
| Delivered in writing | 1 excerpt, 4 statements, reached you as text | None. A device that cannot speak costs you nothing. | none | Reading pair too thin for a figure; placement words printed | You can say what you took from the words on the page; nothing here is a listening claim. A reading figure says nothing about listening, and the page never converts one into the other. |
Omega is McDonald’s omega in Spearman-Brown form from an assumed inter-item correlation of .30, stated as an assumption until live data exists. A row under 8 answered exercises or omega under .70 carries a placement word and no figure. No percentile appears anywhere, because there is no norm group.
Built for
- Employees whose day runs on calls, stand-ups and corridor conversations, who want to know whether what they carry away is what was said
- Managers and team leads who hear commitments, conditions and hints all day and want the two kinds of listening mistake measured apart
- Support, service and back-office teams who act on spoken instructions from colleagues and need the condition to survive the gap between hearing and doing
- Anybody who has sat a free listening quiz, scored well, and still been told they do not listen
Find out what you take from a conversation, in both directions
40 exercises across five formats · about 30 minutes · the sitting is free and the report is the product: two figures never added, a warrant table with every refusal printed, the step-against-leap count, one boxed change and a re-sitting threshold. Report Rs 249 in India inclusive of GST, or US$2.99 elsewhere. Sit it again in four weeks.
Take free · full report ₹249 (incl. GST)
Frequently asked questions
No. Every excerpt is short, slow and in plain words, and no statement is keyed on knowing a word; a fluent speaker can score low and a less fluent one high. It measures one thing: the gap between what a speaker said and what you took from it, in four parts. It does not measure hearing, accent, attention as a trait, empathy, or whether you are a good colleague, and the report's refusal block says so before any figure is printed.
The free Active Listening quiz tests what you know about paraphrasing, question types and the named barriers, with no audio and one score. This is its performance-based upgrade: you hear speech and are scored on what you took from it, in two directions that are never averaged. The sales listening assessment measures listening inside a sales conversation, with buying signals as the construct; this one is set among colleagues, managers, suppliers and support teams. The workplace English assessment measures proficiency; this one never keys a statement on vocabulary. The report names all three on the page.
Because they are different habits with different fixes, and they would cancel. Taking on board something that was only implied or never said costs a plan built on a promise nobody made. Losing something that was said costs a commitment or a condition. A listener who takes everything on board and a listener who doubts everything would average to the same number, and that number would be false for both of them. The report prints caught and kept apart, names the shape of the pair, and prices a two-step slip at three times a one-step slip with the ratio printed above the figures.
The written equivalent appears automatically, and you can ask for it on any device at any time. An exercise that reached you in writing is recorded as read rather than heard, leaves the two listening figures entirely, and is reported in its own row of the warrant table as a reading figure. Nothing about taking that route counts against you: which exercises were spoken to you is a fact about the device, not about your listening.
The sitting is free and the report is the product: Rs 249 in India inclusive of GST, or US$2.99 elsewhere, one time. Forty exercises across five formats take about thirty minutes, and each spoken excerpt plays at most twice. The report names a re-sitting after four weeks of the boxed change and prints the movement each figure would have to clear to count as real change rather than measurement error. No total is printed across the four parts, no percentile appears anywhere, and you are ranked against nobody.
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: Forty original items across five formats: eighteen spoken excerpts each followed by four or five statements sorted SAID, IMPLIED or NOT SAID (eighty-one statements: forty-one said, twenty-one implied, nineteen not said, so just over half are genuine); eight spoken messages with four written answers that differ only in which detail survived; six select-every-line items on a written message; four match-the-following items pairing a spoken line with the most a listener can honestly report from it; and four ordering items on how much four phrasings commit the speaker to. DECLARED RESPONSE INSTRUCTION: one instruction for the whole instrument, keyed accuracy against what was actually said: sort, pick, tick, match or order according to what the speaker put into words, implied, or left out. No item asks what you would do, and no item is self-report. CONSTRUCT STATEMENT: the gap between what was said and what the listener took from it, in four parts: said against inferred, the commitment and its condition, the ask underneath the words, and what was not said. It is NOT a measure of hearing, English proficiency, accent, attention as a trait, empathy, whether you are a good colleague, or how you would listen in a conversation this instrument did not describe; it is not a sales-listening product and not a language test, and no statement is keyed on vocabulary. FRAMEWORK ANCHOR: the Common European Framework of Reference for Languages (Council of Europe, 2001; Companion Volume 2020), listening-comprehension descriptor scales, used for the structure of what a listener does with speech (overall oral comprehension, understanding conversation between other people, understanding announcements and instructions, identifying cues and inferring). The CEFR is a published framework free to cite; no endorsement by the Council of Europe is implied, and this is a workplace listening instrument, not a language-level test. SCORING DESIGN, C22 asymmetric-lure detection with a declared take-up prior, reported as D18 the Warrant Table: every sorted statement is a LURE (implied or not said) or GENUINE (said) and carries prior_i, the declared share of ordinary listeners calling it correctly; w_i = 1 - prior_i; caught = sum(w_i x credit_i) / sum(w_i) over lure statements, kept = the same over genuine statements; E_caught and E_kept are the same sums with the expected credit under the declared prior row in place of credit_i; each is reported as 100 x (figure - E) / (1 - E), so zero is the ordinary listener. Credit is ordinal: a line sorted as keyed earns 1, an adjacent slip (said against implied) earns two thirds, a distant slip (said against not said) earns nothing, a declared cost ratio of three to one printed on the page. The two figures are NEVER netted; the headline is the pair and the shape names the habit. Only excerpts delivered as speech enter the two figures; an excerpt delivered in writing, because the device could not speak or the reader asked for the words, is reported apart as a reading figure and never counted against the reader. Each of the four parts carries a chance-corrected accuracy over all its answered exercises against the declared marginals, with McDonald's omega in Spearman-Brown form from an assumed inter-item correlation of .30, every figure with its SEM and 68 and 95 per cent bands, and a part under eight answered exercises or omega under .70 refused a figure and given a three-way placement word where the figure would have gone. An unanswered exercise leaves the numerator, the denominator and the chance term; an empty sitting scores exactly zero and is not reportable. SOURCES: Grice, Logic and conversation (in Cole and Morgan, Syntax and Semantics 3, 1975), for implicature, the distinction between what is said and what is implied that is the spine of the sorting task; Bransford and Johnson, Contextual prerequisites for understanding (Journal of Verbal Learning and Verbal Behavior, 1972), for comprehension depending on context supplied by the listener; Bartlett, Remembering (Cambridge University Press, 1932), and Brewer and Treyens, Role of schemata in memory for places (Cognitive Psychology, 1981), for schema-driven recall of things that were never present; Loftus and Palmer, Reconstruction of automobile destruction (Journal of Verbal Learning and Verbal Behavior, 1974), and Loftus, Planting misinformation in the human mind (Learning and Memory, 2005), for suggested detail being remembered as witnessed; Deese, On the prediction of occurrence of particular verbal intrusions in immediate recall (Journal of Experimental Psychology, 1959), and Roediger and McDermott, Creating false memories (Journal of Experimental Psychology: Learning, Memory, and Cognition, 1995), for a strongly implied word being confidently remembered as having been presented, which is why the implied statements are the instrument's central lures; Gernsbacher, Language Comprehension as Structure Building (Erlbaum, 1990), for comprehension as building a structure and the cost of mapping new information onto it; Clark and Brennan, Grounding in communication (in Resnick, Levine and Teasley, Perspectives on Socially Shared Cognition, 1991), for what two people must establish before either can take a statement as shared; Weger, Castle Bell, Minei and Robinson, The relative effectiveness of active listening in initial interactions (International Journal of Listening, 2014), for the finding that active-listening responses made speakers feel more understood than advice or simple acknowledgement, while conversational satisfaction and social attraction were no higher than after advice, so that a listening manner and listening accuracy are distinct things and this instrument measures only the second; Bodie, Jones, Vickery, Hatcher and Cannava, Examining the construct validity of enacted support (Communication Monographs, 2014), and Bodie, Issues in the measurement of listening (Communication Research Reports, 2013), for self-reports of listening behaviour converging only weakly with the speaker's and an observer's view of the same conversation, the limit that is why this instrument is performance-based rather than self-report; Janusik, Conversational Listening Span: a proposed measure of conversational listening (International Journal of Listening, 2005), and Building listening theory: the validation of the Conversational Listening Span (Communication Studies, 2007), for listening as a span measured by what is retained from live speech; Council of Europe, Common European Framework of Reference for Languages: Learning, teaching, assessment (2001) and its Companion Volume (2020), for the listening-comprehension descriptor scales named above; Haladyna, Downing and Rodriguez, A review of multiple-choice item-writing guidelines (Applied Measurement in Education, 2002), for line-length and cue control; Jacobson and Truax, Clinical significance (Journal of Consulting and Clinical Psychology, 1991), for the re-measurement threshold; Gollwitzer and Sheeran, Implementation intentions and goal achievement: a meta-analysis (Advances in Experimental Social Psychology, 2006), for the if-then form of the one change. All forty items are original works written for this instrument; no item from any published listening test, language test, communication inventory or commercial product is reproduced or adapted; no listening instrument, communication-style profile or branded method is named or used; no competitor is named; and the instrument is not affiliated with or endorsed by the Council of Europe or by any author or body named above. The sitting is free at Rs 249 under the platform's pay-first floor and the report is the product.