Listening Comprehension Assessment: Six Tests, Six Tasks
Updated: Nov 1
A referral comes in for a student who decodes adequately and still cannot answer questions about what they just read. Somebody has already suggested a listening comprehension assessment, because the logic is clean: if comprehension fails when the print is taken away, the problem is not the print.
The logic is clean. The measurement is not.
Six instruments in current publication carry a score called Listening Comprehension or a domain called Listening. They are not six versions of one task. They are six different tasks, and a student can land in the average range on one and well below it on another without anything about the student changing.
This post compares them on four factual axes only: what the publisher says the score is built from, the age band the publisher states, the qualification level the seller requires, and what the publisher says about who may administer the instrument. That is reporting. It is not clinical judgement, and there is a flat answer at the bottom to the question of whether any of this lets you identify a disorder.
What you are buying when a score is called "listening comprehension"
Start with the thing nobody says out loud. "Listening comprehension" is not a scored task. It is a label a publisher attaches to a scored task, and the publishers attach it to very different things.
Tiffany Hogan, Suzanne Adlof and Crystle Alonzo made this point in print more than a decade ago. Their 2014 review in the International Journal of Speech-Language Pathology includes a table of standardized tests of listening comprehension for children, listing each instrument against the subtests or domains that supply the score.
"A scan of this table suggests that listening comprehension is a complex construct dependent on numerous cognitive and linguistic processes."
Read down the relevant-subtest column of that table and the point lands harder than the sentence does. One instrument supplies the score from Recalling Sentences in Context. Another from Understanding Spoken Paragraphs plus Sentence Comprehension. Another from Making Inferences.
Another assembles it from Paragraph Comprehension, Sentence Comprehension, Non-literal Language, Meaning from Context and Inference. Another from WH-Question Comprehension and Passives Comprehension. Another from Content Memory and Passage Comprehension. Another from Narrative Comprehension and Oral Narration.
Those are not variations in difficulty. They are different constructs wearing one name.
Six instruments, six tasks
Here is the current state of the six a practitioner is most likely to meet, each taken from the publisher's own product documentation rather than from a secondary summary.
WIAT-4: receptive vocabulary plus oral discourse
Pearson's Wechsler Individual Achievement Test, Fourth Edition carries a publication date of 2020 and an age range of 4:0 to 50:11. Its Listening Comprehension score is a composite of two subtests, and Pearson names them in its own flyer: Receptive Vocabulary and Oral Discourse Comprehension.
That pairing matters. Half of the WIAT-4 listening score is a vocabulary task. If a student's word knowledge is thin, the listening score moves, and nothing in the score name tells you that happened.
Listening Comprehension and Oral Expression together form the WIAT-4's supplemental Oral Language composite, available from PreK through grade 12 and beyond.
KTEA-3: a subtest inside an achievement battery
Pearson's Kaufman Test of Educational Achievement, Third Edition is dated 2014 and spans ages 4:0 to 25:11. Listening Comprehension appears on its stimulus book contents list alongside Reading Comprehension, Reading Vocabulary, Oral Expression and Associational Fluency.
So two Pearson achievement batteries both offer a Listening Comprehension score. They are different instruments with different norm samples six years apart, and a practitioner who owns one and not the other has no way to convert between them.
CELF-5: there is no subtest called Listening Comprehension
The Clinical Evaluation of Language Fundamentals, Fifth Edition was published in Fall 2013 for ages 5:0 to 21:11. Its index scores are Core Language, Content, Structure, Language Memory, Receptive Language and Expressive Language.
Note what is not on that list. The CELF-5 does not produce a score named Listening Comprehension at all. What it has is a Listening domain, and Pearson's own crosswalk populates that domain with four subtests: Word Classes, Following Directions, Semantic Relationships and Understanding Spoken Paragraphs, plus an Observational Rating Scale.
The CELF-5's Reading Comprehension and Writing subtests are restricted to ages 9:0 to 21:11, so the instrument's own age band is not uniform across its parts.
OWLS-II: one scale, four linguistic structures
The Oral and Written Language Scales, Second Edition has a Listening Comprehension scale by that exact name. PRO-ED states the age band as 3-0 through 21-11 and a testing time of 20 to 30 minutes.
The score is not built from a passage task. Each OWLS-II scale assesses four linguistic structures: lexical and semantic, syntactic, pragmatic, and supralinguistic. The Listening Comprehension scale and the Oral Expression scale are the receptive and expressive halves of that design.
OPUS: passages, and the publisher says memory
The Oral Passage Understanding Scale is the only one of the six built specifically and solely for this construct. PAR describes it as a measure of auditory comprehension that evaluates a person's ability to listen to passages read aloud and recall information about them, across 17 passages divided into six item sets. PAR states an age range of 5 to 21 years and an administration time of 10 to 20 minutes.
And then PAR says the part most score reports leave out.
"It also measures memory skills, which are integral to listening comprehension."
That is the publisher of a listening comprehension test telling you in its product copy that the score is partly a memory score. It is conormed with the CASL-2 and shares the same underlying Integrative Language Theory, which is why the two can be interpreted together and why neither is interchangeable with a passage task from a different family.
TILLS: yes, no, or maybe
Brookes Publishing's Test of Integrated Language and Literacy Skills covers ages 6 to 18. Its Listening Comprehension subtest is Subtest 6, and its stated purpose is narrow and specific.
"To assess a student's ability to comprehend the complex syntax of academic language and to draw inferences allowed by the text."
The task is a set of very short stories, and the student answers yes, no, or maybe. The examiner's directions instruct the student to say maybe when the story does not clearly supply the answer, and repetition of the stimulus is not permitted except on the practice item.
A three-option inferential judgement on deliberately complex syntax is a different measurement from recalling a passage, which is a different measurement from defining words. All three are sold as listening comprehension.
Pearson publishes a crosswalk between its own two instruments
This is the part worth sitting with, because it is not an inference drawn by a critic. It is Pearson's own marketing document.
A Pearson flyer titled for speech-language pathologists sets the WIAT-4 side by side with a CELF-5 assessment process, organised by Language Domain. In the row labelled Listening, the CELF-5 column lists Word Classes, Following Directions, Semantic Relationships, Understanding Spoken Paragraphs and an Observational Rating Scale. The WIAT-4 column lists Listening Comprehension, broken out as Receptive Vocabulary and Oral Discourse Comprehension.
The flyer says the WIAT-4 can help you support and augment comprehensive oral language skills assessment tasks, and in several rows the CELF-5 column simply reads N/A.
Two observations follow, and they pull in opposite directions.
The first is that the publisher is explicitly positioning an achievement battery as coverage for a language domain. The second is that the two columns do not contain the same tasks, which the table itself makes plain. A crosswalk that lines up two columns of different content is a claim about coverage, not a claim about equivalence, and it is easy to read it as the second.
One more thing about that flyer is worth counting. Across its roughly 600 words, the word "speech" appears zero times, despite the document being addressed to speech-language pathologists in its own headline.
Every one of them is Level B, and the definition of Level B is circular
Here is where the purchasing gate stops helping you.
Every instrument above sits at qualification level B at the seller that publishes a level for it. The WIAT-4 is Level B at Pearson. The KTEA-3 is Level B at Pearson. The CELF-5 is Level B at Pearson. The CASL-2 is Test Level B at PRO-ED. The OWLS-II is Test Level B at PRO-ED, and PRO-ED's own description line states Qualification Level: B. The OPUS is Qualification Level B at PAR.
An achievement battery a learning specialist may already own and a clinical language battery administered by speech-language pathologists carry the same letter. The letter does not mark the boundary. It was never designed to.
And on the Pearson product page where that letter appears, the definition offered for it is this.
"Level B: This approval level enables you to buy our assessments requiring A or B qualification levels."
Level B is defined as the level that lets you buy Level B things. Only Level A carries a substantive definition on that page, as the level for assessments that require "no professional degree, accreditation, organization membership, or license/certificate," and Level C is defined as the level that lets you buy everything.
No profession is named. No training is specified. Nothing on that page distinguishes a vocabulary subtest from a clinical language battery.
This is the same finding this site reported about achievement test qualification levels, arriving from the opposite direction. There the problem was that one instrument carries different letters at different sellers. Here the problem is that six different kinds of instrument carry the same letter, so the letter cannot tell you whose instrument it is.
The age bands do not line up either, and one of them disagrees with itself
Two sellers of the OWLS-II describe its age coverage differently.
PRO-ED's product description gives a single band: ages 3-0 through 21-11. PAR's page splits it by scale, stating that the Listening Comprehension and Oral Expression scales assess individuals ages 3-21 years while Reading Comprehension and Written Expression are designed for individuals ages 5 to 21 years.
Both statements are about the same instrument. Only one of them tells you that the listening scale and the reading scale do not cover the same ages, which is exactly the fact you need if you are comparing a listening score against a reading score for a five- or six-year-old.
Across the six, the stated bands are 4:0 to 50:11 for the WIAT-4, 4:0 to 25:11 for the KTEA-3, 5:0 to 21:11 for the CELF-5 with its reading and writing parts starting at 9:0, 3-0 through 21-11 for the OWLS-II, 5 to 21 for the OPUS, and 6 to 18 for the TILLS.
An eighteen-year-old is in range on five of the six. A four-year-old is in range on three. A twenty-four-year-old is in range on one.
The practical consequence is that your instrument choice for an older adolescent or a young adult is made for you by the age bands long before any question of construct arises. The same reasoning applies to the norm recency problem these publication dates imply, and the CELF-5 is now more than a decade old.
Only one of the six says who may administer it
Here is the single most useful sentence in this whole area, and it is in the TILLS sampler rather than in any ethics code.
"TILLS can be administered by any professional who has received training on working with children and adolescents with disabilities and administering and scoring individualized standardized assessments. This includes speech-language pathologists, special educators, reading specialists, learning disability specialists, neuropsychologists, and educational psychologists."
That is a publisher answering the question directly instead of delegating it to a letter grade. It names a training standard, then names six professions.
Educational therapist is not one of them. Across roughly 15,700 words of the TILLS sampler, the phrase "educational therapist" appears zero times.
Read that carefully, because it is easy to over-read. The governing clause in the sentence is the training standard, not the list, and the list is introduced with "this includes" rather than a restriction. A publisher's illustrative list of professions is not a prohibition, and nothing on that page says an educational therapist may not administer the instrument.
But it is also not nothing. The one instrument in this group whose publisher troubled to say who it is for named six professions and did not name ours. If you are deciding whether to add a language instrument to your kit, that absence is worth more than the letter B.
What AET's own documents say, and do not say
If the publishers will not draw the line, the next place to look is the body that governs the practice. The answer there is thinner than practitioners expect, and it is thin in an informative way.
The AET Code of Ethics runs to roughly 3,000 words. Counted as whole words, "listening" appears zero times. "Oral language" appears zero times. "Comprehension" appears zero times. "Pathologist" appears zero times.
"Language" appears three times in total. "Speech" appears twice, and one of those two is a bibliographic reference to ASHA in the acknowledgements for the e-practice sections.
Which leaves exactly one sentence in the Code that bears on this post, and it is the parenthetical in Section II.I.D.
"seek interpretation of assessment data from professionals in related fields (e.g. medical, psychological, speech/language, neuropsychological)."
The clause immediately before it, II.I.C, is the one that governs administration: educational therapists "use only those assessment instruments for which they have been adequately trained." II.I.B adds that they select and use appropriate instruments "recognizing their limitations with respect to reliability, validity, and bias."
Put the three together and AET's position is consistent, if compressed. Administration is governed by your training. Interpretation of language data is to be sought from the speech and language field. The Code does not list which instruments fall on which side, and it does not use the word "listening" once.
The AET Fact Sheet is thinner still on this subject. In roughly 760 words it contains the word "language" zero times. Its much-quoted four-item DO NOT list is: diagnose; administer cognitive, intelligence, or psychological tests unless otherwise qualified to do so; practice psychotherapy; prescribe medication.
A language test is not on that list, and neither is a listening comprehension subtest of an achievement battery. The absence is not permission. It is silence, and this site has now found the same silence on session notes, on vocabulary, on scope of practice and on informed consent.
What the research says about the score you actually get
Two free full-text papers are worth reading before you put a listening comprehension number in a report, and both complicate the number in ways the score report does not.
The passage contributes more variance than the child
Young-Suk Grace Kim and Yaacov Petscher published an explanatory item response analysis of listening comprehension in Annals of Dyslexia, volume 71, issue 2, using data from 529 second graders.
They partitioned the variance in item responses five ways. The result, from their unconditional model, is the single most sobering number in this area.
Between-item differences accounted for 33 percent of the variance. Between-passage differences accounted for 18 percent. Between-person differences accounted for 8 percent. Between-classroom differences accounted for 2 percent, and between-school differences for 0 percent.
The child accounted for 8 percent. The passage accounted for more than twice that, and the particular item more than four times it.
They went further on the passage effect and found that narrative versus expository genre explained almost all of the between-passage variance, with a pseudo R-squared of 0.98 for passage variance in the conditional model.
State the boundary in the same breath, because the finding is easy to over-read. This is one study, of one set of passages, with second graders, modelling item-level accuracy rather than composite standard scores.
It does not license the claim that a published instrument's composite is 8 percent about the child. What it does establish is that genre and item selection are not neutral, and that two instruments sampling different text types are not measuring the same thing even when both are well built.
The construct's weight changes as the student gets older
Hogan, Adlof and Alonzo's 2014 review assembles longitudinal evidence on how much of reading comprehension each half of the simple view explains, and the shift across grades is large.
The unique variance explained by word recognition fell from 27 percent in second grade to 13 percent in fourth grade to 2 percent in eighth grade. Over the same span the unique variance explained by listening comprehension rose from 9 percent to 21 percent to 36 percent.
By eighth grade, they report, all of the reliable variance in reading comprehension could be explained by the listening comprehension factor, citing the finding that the two had become a unitary construct by that point.
That has a direct practice consequence and it is not a clinical one. If you work with older students, a listening comprehension measure is carrying far more of the explanatory load than the same measure would for a second grader. The instrument has not changed. Its relevance has.
It also means the familiar move of comparing listening comprehension against reading comprehension to isolate a decoding problem gets weaker with age, because by eighth grade the two measures are converging on the same factor. That is a different caution from the ones in the manuals, and none of the six manuals makes it.
Can a listening comprehension score identify a language disorder?
No.
Not from any of the six instruments above, administered by an educational therapist, learning specialist or executive function coach.
A below-average Listening Comprehension score tells you the student performed below the normative sample on a particular task with a particular set of passages or items. It does not name a cause, it does not establish a disorder, and the construct is dependent enough on memory, vocabulary, genre and item set that attributing it to a single underlying ability is not supported by the measurement.
Identification of a language disorder is the business of a qualified professional in that field. AET's Fact Sheet puts diagnosis first on its DO NOT list, and the Code of Ethics directs you to seek interpretation of assessment data from the speech and language field by name. The honest version of the finding is a referral and a description, not a label.
Nothing here is legal advice. Scope questions that turn on your own state's licensing arrangements belong with that state's agency and, for independent practice, with an attorney. This post reports what publishers and professional bodies publish, nothing more.
A sequence, not a decision
What follows is the order of questions, not a recommendation of an instrument. The decision about which instrument fits a particular student is yours, and it depends on facts this post does not have.
Ask what question the score is for. "Is comprehension failing when decoding is removed" is a different question from "is receptive vocabulary thin" and from "can this student hold a passage in mind." Three of the six instruments answer one of those well and the others poorly.
Read the subtest composition before the score name. A composite that includes a vocabulary task will move with vocabulary. A composite built from passage recall will move with memory, which PAR says in its own product copy.
Check the age band against your actual student, scale by scale. The OWLS-II example shows that one instrument's scales can cover different ages, and that two sellers can describe the same instrument's coverage differently.
Check the publication date against the decision you are making. The newest of the six was normed in 2020 and the oldest in 2013.
Decide whether the question is yours to answer at all. If the answer you need is whether a language disorder is present, the instrument choice is the wrong problem and a referral is the right one.
Say in the report which instrument produced the number and what task it was. A reader who knows your listening score came from Receptive Vocabulary plus Oral Discourse Comprehension can use it. A reader who sees only "Listening Comprehension: 84" cannot.
Where this sits in the rest of the record
This is the ninth instrument comparison on this site, and the pattern it belongs to is now large enough to be worth laying out, because each prior post supplies a fact this one leans on.
Reading comprehension instruments disagree with each other for the same structural reason, which is why two scores on one student diverge. Vocabulary instruments overlap this construct directly, and the WIAT-4's listening composite is half a vocabulary task.
Achievement test qualification levels established that the letter is a storefront rule and not an industry scale. Written expression instruments sit on the same discipline fault line, split between occupational therapy and the language professions.
Outdated test norms is the post to read on what the 2013 and 2014 publication dates above actually cost you. Standard error of measurement is the one on why a single listening score is a band rather than a point, and grade equivalent scores is the one on the metric you should not report it in.
Phonological awareness assessment and dyslexia screening tools cover the decoding side of the comparison this post is about. Oral reading fluency norms and reading comprehension for older students carry the age-band argument further.
Executive function scope of practice is the closest precedent for the discipline-boundary question here, and what educational therapy is is the scope map behind it. The psychoeducational report post covers the case where somebody else produced the score and you are reading it.
Assessing multilingual learners matters more here than on most instrument posts, because a listening measure is a language measure before it is an academic one. Remote administration matters because a passage read aloud over a connection is not the standardised condition the norms were collected under.
Three more are adjacent rather than central. Morphology instruction and spelling assessment both deal with constructs sold under one name and measured several ways. The science of reading is where the simple view of reading, which this post's research section depends on, gets its popular framing.
And two on practice rather than measurement. Session notes is where a finding like this one has to land to be useful, and dyslexia defined is the clearest example on this site of one word meaning five things to five bodies.
The boundary this post describes at the level of a single score is compared document by document in Educational Therapy Referral to SLPs and OTs, which covers what each body publishes about referral and written consent, and notes ASHA's own statement that speech-language pathology services are not prescribed or supervised by another professional.
Where the published record runs out
Several things a practitioner would reasonably want are not in the public record, and it is worth naming them rather than filling the gap with a confident sentence.
No published crosswalk exists between these six scores. Pearson's flyer lines up two of its own instruments by domain, not by equivalence, and no publisher offers a conversion between a WIAT-4 listening score and an OWLS-II one.
PAR does not state its qualification level definitions on the product page. The level letter is rendered to signed-in users, so the definition behind PAR's B could not be read for this post. PRO-ED states its level requirement for purchase but its tier definitions sit behind a separate qualifications process.
Five of the six publishers do not say who may administer. Only Brookes answers the question in plain words. The rest delegate it to a letter whose own definition, in Pearson's case, is circular.
AET has published nothing on language assessment. Not a position statement, not a guidance note, not a clause naming the instruments. Its single relevant sentence is a parenthetical in II.I.D.
The Kim and Petscher variance partition has not been replicated on a published instrument. It used study passages with second graders. Whether a commercial composite behaves the same way is untested in public.
ASHA's Scope of Practice could not be read for this post. Its policy page served an automated-client challenge on two attempts, so no claim about ASHA's scope language is made here. If that boundary is load-bearing for a decision you are making, open the document yourself rather than relying on any summary, including this one.
Frequently asked questions
Is listening comprehension the same thing on every test?
No. Six current instruments build the score from different tasks: receptive vocabulary plus oral discourse on the WIAT-4, a four-subtest Listening domain on the CELF-5, four linguistic structures on the OWLS-II, passage recall on the OPUS, and a yes-no-maybe inference task on short stories with complex syntax on the TILLS. A student can score differently across them without changing.
What qualification level do I need to buy a listening comprehension test?
Every instrument in this comparison sits at level B at the seller that publishes a level: the WIAT-4, KTEA-3 and CELF-5 at Pearson, the CASL-2 and OWLS-II at PRO-ED, and the OPUS at PAR. Pearson's own on-page definition of Level B is that it enables you to buy assessments requiring A or B qualification levels, which names no profession and no training.
Can an educational therapist administer a language test?
AET's Code of Ethics II.I.C says educational therapists use only those assessment instruments for which they have been adequately trained, and II.I.D says to seek interpretation of assessment data from professionals in related fields including speech and language. The Code names no instruments. Brookes is the only publisher of the six that states who may administer, and its list of six professions does not include educational therapists.
Does a low listening comprehension score mean the problem is not decoding?
Not on its own, and the inference weakens with age. Hogan, Adlof and Alonzo report that the unique variance in reading comprehension explained by word recognition falls from 27 percent in second grade to 2 percent in eighth, while listening comprehension rises from 9 percent to 36 percent, and that by eighth grade the two measures converge on a single factor. The listening-versus-reading comparison is strongest in the early grades.
Does the passage affect the score?
Yes, and more than most score reports suggest. In Kim and Petscher's analysis of 529 second graders, between-passage differences accounted for 18 percent of the variance in item responses against 8 percent for between-person differences, and narrative versus expository genre explained almost all of the passage-level variance. That is one study with study passages, not a published composite, but it is a reason to compare instruments on text type.
Which listening comprehension test should I use?
This post does not answer that, deliberately. Choosing an instrument for a particular student is a clinical judgement about that student, and the four axes compared here — subtest composition, stated age band, qualification level and who the publisher says may administer — are the factual inputs to that judgement rather than a substitute for it. Read the subtest composition first and the score name last.
If this is the work you do
Illuminate publishes for practitioners: educational therapists, learning specialists and executive function coaches who need the published record rather than a summary of it. If the distinctions in this post are the ones you work with daily, the roles comparison is the place to start, and the practitioner list is where the next one arrives.
Join the educator list for new practitioner-facing comparisons as they publish.




Comments