top of page
Search

Assessing Reading in Multilingual Students: Whose Norms Are They?

Writer: Andrea Chernin
Andrea Chernin
Oct 19
22 min read

Assessing Reading in Multilingual Students: Whose Norms Are They?


You have a nine-year-old on your caseload who speaks Spanish at home, has been in English-medium classrooms since kindergarten, and reads about two years below where the school thinks she should be. Somebody has handed you a report. It has standard scores in it.


The question you are actually being asked is whether the scores mean what they would mean for the child in the next chair.


That is a question about a normative sample, and assessing reading in multilingual students comes down to reading the sample description rather than the score. The good news is that the publishers write it down.


The uncomfortable news is what it says: every Spanish-language reading and cognitive instrument in common American use is anchored, by one route or another, to an English one. Three publishers say so in print. A fourth says its thresholds came from the English version. The largest reference sample in the whole landscape is a sample in which most children's primary language was English.


None of that makes these instruments unusable. It changes what a score from one of them is a statement about, and that is a thing you can put in a report.


A norm is not a standard. It is a description of what a particular group of people did on a particular set of items at a particular time.

The scale of the thing, in the federal count


The National Center for Education Statistics reports that in fall 2021, 5.3 million public school students — 10.6 percent of all of them — were identified as English learners. Spanish was the most commonly reported home language, at about 4.0 million students, or 76.4 percent of all English learners, which is 8.4 percent of the entire public school population.


So roughly one American public school student in ten is formally identified as an English learner, and three-quarters of those speak Spanish at home. This is not a specialist corner of practice. It is one child in every classroom of ten.


The published assessment record does not look like that number. It looks like a market with one well-documented language in it and a great deal of careful hedging around the edges.


What a norm is a statement about


This is worth being precise about, because the whole post hangs on it.


A norm-referenced score tells you where a person's raw performance falls in a distribution of raw performances collected from a reference group. The score is a position. The position only means something if you know who else is standing in the line.


That is why outdated test norms matter, and why the band around a score matters at least as much as the score, as the standard error of measurement piece set out. With a multilingual student, a third question joins those two: what language, and how much of it, did the reference group have?


Four different answers to that question are currently on the market. All four are defensible. None of them is the answer a report usually implies.


Four ways publishers solve the problem, and all four end in English


One: equate the Spanish items to the English norms


The Batería IV Woodcock-Muñoz is the most widely used Spanish-language psychoeducational battery in the United States, and its publisher describes the architecture with unusual clarity.


From Riverside Insights' own Assessment Service Bulletin Number 1 for the Batería IV:


"Spanish-language calibration data, based on a sample of 601 native Spanish speakers, are equated to the large, nationally representative WJ IV norming sample of 7,416 individuals ranging from 2 to 90+ years of age."

And, later in the same document, flatly:


"The Batería IV tests are parallel Spanish forms of the English tests from the WJ IV; both tests rely on the same set of norms to produce scores such as standard scores and percentile ranks."

Read that twice. There is no separate Spanish normative distribution. Spanish items were calibrated onto the same Rasch scale as the English items, and a student's Spanish performance is then placed in the distribution built from 7,416 people who took the test in English.


The bulletin also contains a sentence that sits awkwardly beside the one above: "These examinees' normative scores will reflect their relative standings in a distribution of their same-language peers of the same age or grade." Both sentences are in the same section, two paragraphs apart. The second is hard to reconcile with the first unless "same-language peers" is being used to mean something like equated-ability peers.


Two: collect a Spanish-speaking sample, then equate it anyway


Pearson took a different route with the WISC-V Spanish, and published the specification on the product page:


"Updated normative sample standardized on 2,200 children aged 6:0–16:11 whose primary language is Spanish and who have attended schools in the U.S. for less than 5 consecutive years. Equated to the English normative sample using IRT approach."

That is a real Spanish-speaking normative sample, and 2,200 is a serious number. Note the two qualifiers, because they do a great deal of work, and I will come back to the second one.


The same page describes the instrument as being for children "who speak primarily Spanish, many of whom are in the process of acculturating to the U.S.", and lists among its features "optional language-environment verbal score adjustments." The test is administered in Spanish; examiner instructions are in English.


Three: build criterion thresholds from the English version


Acadience Learning publishes Acadience Reading Español, a Spanish early-literacy measure structured like its English sibling. Its benchmarks and composite score handout, dated December 2024, says this:


"The preliminary benchmarks and cut points for risk were developed based on the corresponding English versions of these measures. The Acadience Reading Español subtests were developed using a similar process and similar design specifications to the corresponding literacy skills in English."

Every threshold in the document is labelled preliminary, and the publisher says continuing research will refine them. The word "norms" does not appear in the handout at all. It is criterion-referenced throughout, which is the honest framing for what it is.


This matters for the same reason it mattered in progress monitoring tools compared: a benchmark is a prediction about a later outcome, not a percentile.


Four: report against same-language peers


The one instrument in this group that claims separate same-language comparison groups is the Woodcock-Muñoz Language Survey III, also from Riverside. Its brochure includes a Spanish-language sample parent report which states that the student's performance on the English version of the tests is analysed in comparison with English-speaking peers of the same grade, and performance on the Spanish version in comparison with Spanish-speaking peers of the same grade.


That is a different and stronger claim than the Batería IV's. It is also the claim I could not verify, and I will say so plainly below rather than leave it sitting here looking confirmed.


The Batería IV's own sample table is the most useful document in this landscape


Here is the part nobody writes for practitioners, and it is sitting in a free PDF.


The bulletin reports the calibration study in full. It ran from December 2017 to April 2018. Six tests were administered to 601 native Spanish speakers aged 2 to 81, about a hundred in each of six age bands. And then it publishes two tables that are, in their quiet way, the most informative pages in the Spanish assessment literature.


Ten states, and 62 per cent of the sample in one of them


The geographic distribution of those 601 examinees:


  • Texas — 375 examinees, 62.4 percent

  • Illinois — 99, 16.5 percent

  • Tennessee — 67, 11.2 percent

  • California — 22, 3.7 percent

  • Virginia — 17, 2.8 percent

  • New York — 7, 1.2 percent

  • Florida and New Jersey — 5 each, 0.8 percent

  • Arizona — 3, 0.5 percent

  • Connecticut — 1, 0.2 percent


Ten states. Three of them hold 90.1 percent of the sample. The English norming study it is equated to drew on 7,416 individuals from 46 states and the District of Columbia.


And one country of origin holding seven examinees in ten


The bulletin also reports country of Hispanic origin, which almost no test document does:


  • Mexican — 421, 70.1 percent

  • Other or mixed — 104, 17.3 percent

  • Puerto Rican — 33, 5.5 percent

  • Salvadoran — 17, 2.8 percent

  • Guatemalan — 12, 2.0 percent

  • Cuban — 9, 1.5 percent

  • Dominican — 5, 0.8 percent


I want to be careful here. This is a calibration sample, not a normative sample. Its job was to establish item difficulties so the Spanish items could be placed on the same scale as the English ones, and 601 people is a reasonable size for that job.


The publisher is not claiming it is nationally representative of Spanish speakers, and it describes the sample as chosen "to ensure a broad representation of sex, parent or examinee education level, and country of Hispanic origin/nativity."


The point is not that Riverside did something wrong. The point is that Riverside published the table, and the table tells you something no score report will: the Spanish item difficulties underpinning this battery were established largely on Mexican-origin Spanish speakers in Texas and Illinois. Hold that beside the testing standards in a moment.


What the equating actually buys you


Something genuinely useful, which is worth saying because the above reads as criticism and is not.


Because English and Spanish items sit on one scale, the battery supports a Comparative Language Index: a direct comparison of a student's language proficiency in each language, from the three parallel English and Spanish test pairs in the Woodcock-Johnson IV Tests of Oral Language.


The bulletin specifies which pairs, and notes that an examiner wanting the index in the report must administer the English tests and commit the results to the scoring program within 30 days of the Spanish administration.


The battery also carries a structured Language Background Information section. It asks the examiner to classify the examinee as a native English speaker, a second-language learner of English, a native English speaker learning a foreign or heritage language, or a simultaneous bilingual individual.


It also records the languages spoken at home, with peers and in the classroom, and the amount of time spent in prior language programmes including schooling outside the United States. It can report Cognitive Academic Language Proficiency levels, on Cummins' 1984 distinction between everyday and academic language proficiency.


That is more structured attention to language history than any English achievement battery asks for. It is a good design. It just does not change what the norms are.


The WISC-V Spanish excludes the student you are most likely to see


Back to that second qualifier, because it is the most practically consequential sentence in this whole post.


The WISC-V Spanish normative sample consists of children "whose primary language is Spanish and who have attended schools in the U.S. for less than 5 consecutive years."


Now picture your caseload. The student who arrives at an educational therapist's door with a reading problem and a Spanish-speaking home is very often a student who has been in American schools since kindergarten. Six years. Eight years. Her whole school life.


The child who has been in US schools for eight years is, by the publisher's own specification, not in the group her score is being compared against.

That is not a defect in the instrument. The sample was defined that way deliberately, and for a good reason: a child recently arrived from a Spanish-medium school system is exactly the child for whom an English-language cognitive battery is least appropriate, and that child is who this test was built for. The exclusion is the design.


It does mean that for the long-resident bilingual student — the modal educational therapy client — neither the English instrument nor the Spanish instrument has a reference group that looks like her. Which is the actual finding, and it is a finding about samples, not about children.


The largest Spanish reading sample is mostly English-primary students


Amplify publishes mCLASS Lectura, a Spanish early-reading measure, and its technical manual is the most detailed sample documentation anyone in this area has released. It is also the most surprising.


The manual describes a 2020–21 calibration study and a 2021–22 field study yielding three further samples. The section introducing the norming sample reports 32,933 students, and describes them:


"In addition, 30.0% of the students in the sample were considered English Learners (ELs), 52.6% of the students identified English as their primary language, and approximately 17.4% of the sample lacked language data."

A majority of the children in the reference sample for a Spanish reading measure identified English as their primary language.


The sample used to establish the cut scores is smaller and more striking still. Of the 1,864 students in that sample, 25.4 percent were considered English learners and 71.9 percent identified English as their primary language; 46.3 percent reported Spanish as the primary language at home against 42.4 percent English.


The manual explains why, and the explanation is entirely reasonable: these were dual-language programmes. Students in kindergarten through grade 2 were receiving 80 percent or more of their literacy instruction in Spanish, and students in grades 3 to 5 about half of it.


So the cut scores on a Spanish reading measure were set on a group of children who were mostly English-primary at home and were being taught to read in Spanish at school. If your student is Spanish-primary at home and has been taught to read only in English, she is close to the mirror image of that group.


One documentation note, offered as an observation rather than a criticism: the manual's regional-coverage sentences carry the previous sample's label — the paragraph describing the norming sample opens by naming Sample A, and the paragraph reporting Sample C's 3,995 students opens by naming Sample B. The figures are tied unambiguously to the tables they introduce, but anyone quoting a sample size from this document should quote it against its table.


The same manual, incidentally, documents that its Spanish word bank drew on "language assessments designed for use with Spanish-English bilinguals in the U.S.", naming the bilingual version of the Expressive One Word Picture Vocabulary Test among its sources. The word "norms" appears nowhere in the manual; "normative" appears once.


What the testing standards ask for, and why no sample can supply it


The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014) is the governing document, and it is readable free as an open-access PDF. The AERA catalogue page is the canonical reference for the edition.


Counting words in it, which is the fastest way to find out what a document actually prioritises: "linguistic" appears 51 times, forms of "translate" 32 times, "interpreter" 20, "language proficiency" 18, "English language learner" 10, "bilingual" 8, "Spanish" 5, "multilingual" 4. "Dual language" appears zero times.


Standard 3.12 asks for subgroup evidence the samples cannot carry


Standard 3.12 requires test developers and users to describe the methods establishing the adequacy of an adaptation and to document evidence for the validity of score interpretations. Its comment then gets specific, and this is the sentence to hold beside the Batería IV's origin table:


"For example, if a test is translated and adapted into Spanish for use with Central American, Cuban, Mexican, Puerto Rican, South American, and Spanish populations, the validity of test score interpretations for specific uses should be evaluated with members of each of these groups separately, where feasible."

The comment continues: "Where sample sizes permit, evidence of score accuracy and precision should be provided for each group, and test properties for each subgroup should be included in test manuals."


Six populations, each to be evaluated separately where feasible. In the Batería IV calibration sample, the entire non-Mexican, non-other/mixed portion is 76 people spread across four named origins — nine Cubans, five Dominicans, twelve Guatemalans, seventeen Salvadorans, thirty-three Puerto Ricans.


The standard's escape hatches are doing the work: "where feasible", and "where sample sizes permit". They do not permit. That is not a publisher failing to meet a standard; it is a standard describing a study nobody has funded.


Standard 3.13 is the sentence that should be on every report


Standard 3.13 itself is one line: "A test should be administered in the language that is most relevant and appropriate to the test purpose."


Its comment is where the real content is, and it contains three claims practitioners should be able to quote.


First, that the most proficient language may not be the language of instruction: "In some cases, test takers' most proficient language in general may not be the language in which they were instructed or trained in relation to tested constructs."


Second, and this is the sentence I would print and pin up:


"Determination of a test taker's most proficient language for test administration does not automatically guarantee validity of score inferences for the intended use."

Third, the clause that names the thing the whole market talks around: "individuals may be more proficient in one language than another, but not necessarily developmentally proficient in either", and "disconnects between the language of construct acquisition and that of assessment also can compromise appropriate interpretation of the test taker's scores."


Choosing the language of administration is a decision about which of two imperfect comparisons you want. The standards say that out loud. The score reports do not.


Standard 3.14 is about interpreters, and it sets a high bar


Standard 3.14 requires that where testing needs an interpreter, the interpreter follow standardised procedures and be "sufficiently fluent in the language and content of the test and the examinee's native language and culture."


The comment states that individuals with limited proficiency in the language of the test "should ideally be tested by professionally trained bilingual/bicultural examiners", with an interpreter as the fallback, and makes the test user responsible for selecting a qualified one and for documenting the procedure.


The relevant point for an educational therapist is the one about documentation, which connects straight to what the codes do and do not require of session notes and progress reports. If an interpreter was involved anywhere in the assessment you are reading, that fact belongs in the record and very often is not in it.


The federal record declined this question on purpose


There is one federal practice guide written about teaching English learners in the grades educational therapists mostly work in: Teaching Academic Content and Literacy to English Learners in Elementary and Middle School, NCEE 2014-4012, from the Institute of Education Sciences and the What Works Clearinghouse.


Its four recommendations and their evidence ratings:


  • Teach a set of academic vocabulary words intensively across several days — Strong

  • Integrate oral and written English language instruction into content-area teaching — Strong

  • Provide regular structured opportunities to develop written language skills — Minimal

  • Provide small-group instructional intervention to students struggling in literacy and English — Moderate


The screening recommendation was dropped deliberately


The 2014 guide updated a 2007 one, and the update removed something. The guide says so itself:


"The guide no longer includes a separate recommendation on universal screening. The panel chose not to pursue the rapidly changing issue of universal screening and formative assessment in this practice guide."

It refers readers back to Recommendation 1 of the 2007 guide for screening measures. That document is now nineteen years old, and it is the live federal reference for how to screen an English learner's reading.


"Such systems are not available"


Recommendation 4 is where the guide does address identification, and it is unusually blunt about the state of the field:


"While it would be ideal to assess English learners along with native English speakers using universal screening systems with some evidence of reliability and predictive validity in literacy, the panel notes that such systems are not available in the areas of oral language and reading and listening comprehension for students in grades K–8."

Validated measures do exist in foundational skills, the panel says — the decoding and phonemic territory covered in phonological awareness assessment and in dyslexia screening tools. For oral language and comprehension, they are not there. The panel adds that the screening process "should not end with a brief screening measure that focuses on decoding and phonemic skills", because a flag on foundational skills does not mean foundational skills are the only need.


That second half is the practice implication, and it lands in exactly the gap reading comprehension assessment tools describes for monolingual students, only wider.


One more word count, because it is startling for a document of 45,000 words about teaching English learners: the phrases "native language", "home language" and "first language" appear zero times each. "Primary language" appears eleven times, "bilingual" twenty-three, "Spanish" twenty-six. It is a guide about instruction in English, and it keeps that scope strictly.


The clearinghouse that rates tools has no language domain


The National Center on Intensive Intervention maintains the three tools charts that practitioners use as a shortcut to technical adequacy: Academic Screening, Academic Progress Monitoring, and Instructional Intervention Tools.


Searching all three charts as published: "Spanish" appears zero times, "bilingual" zero times, "multilingual" zero times. There is no language-of-administration filter, no Spanish-language tool listed, and no rating criterion that touches the language of the normative sample.


The only occurrences of "English learner" anywhere on the three chart pages are navigation links — to NCII's separate resource pages, including "Implementing DBI for English Learners" and "Taxonomy of Intervention Intensity: Considerations for English Learners". Those resources are real and worth reading. They are guidance documents, and they sit outside the rating rubrics entirely.


This is the same shape as the finding in assistive technology for dyslexia: NCII addresses the topic in its guidance and not in its ratings, so a tool's chart entry tells you nothing about it. If you have been using the charts as a filter, as reading program evidence ratings suggested you should, this is one dimension the filter does not see.


One more thing worth noticing about the market


Riverside markets the Woodcock-Johnson V as "the only co-normed platform measuring cognitive ability, academic achievement, and oral language", and its brochure states the battery "is normed from 2022–2024 and has a sample size of almost 6000."


The word "Spanish" appears in that brochure zero times. The word "language" appears nineteen.


The brochure announces a WJ V Nonverbal, seven nonverbal tests from the cognitive battery, "Coming November 1st, 2026", described as "Designed for English Learners, students with language or hearing impairments, and individuals for whom language may mask ability" and claimed as "Normed on representative samples, including EL and language-impaired groups." The brochure publishes no numbers for those subgroup samples.


And it gives this instruction: use it standalone, or "pair with WJ V COG or Batería IV for bilingual or complex evaluations."


Pair a 2026 nonverbal battery, normed 2022 to 2024, with a 2019 Spanish battery whose items are equated to a 2014 English norming study. That is three generations of norms in one evaluation, recommended by the publisher of all three, and the practitioner is the one who has to write the interpretation. Readers of outdated test norms will recognise the shape.


Two smaller observations from the same catalogue. Riverside's Arabiya is advertised as "The First Comprehensive Psychoeducational Battery in Arabic", built with "cultural and linguistic precision" — and its page publishes no normative sample information of any kind. Logramos, the Spanish achievement battery, offers "comparisons of performance to local and national groups" without naming the group anywhere on the page.


Two languages have comprehensive batteries in this catalogue. The published-norms detail available on the page runs from a free 20-page bulletin with full sample tables at one end to a marketing paragraph at the other.


What AET's Code of Ethics says about any of this


Almost nothing, and the almost is the interesting part.


The AET Code of Ethics, Revision #9, 2025, runs to roughly 2,978 words. Counted through:


  • "bilingual" — 0 occurrences

  • "multilingual" — 0

  • "Spanish" — 0

  • "English learner" — 0

  • forms of "translate" — 0

  • "interpreter" — 0

  • "norm" — 0

  • "language" — 3

  • "cultural" — 2, "culture" — 1


Of the three uses of "language", one sits inside "speech/language" in the list of related professions at II.I.D, one is the name of the American Speech-Language-Hearing Association, and one is a real clause: II.II.E, which asks educational therapists to develop effective communication with parents and guardians, "avoiding or interpreting technical terminology, using the primary language of the home and other modes of communication when appropriate."


So the profession's only language clause is about explaining results to a family, not about producing them.


What the Code does give you is II.I.B, which is short and does more work than its length suggests. Educational therapists are to "select and use appropriate assessment instruments, recognizing their limitations with respect to reliability, validity, and bias."


Bias is named. The Code does not define it, does not mention language once in connection with it, and does not say what recognising a limitation requires you to do. But the obligation to recognise it is there in the one clause that governs instrument selection, and everything in this post is a limitation of exactly that kind.


Alongside it: II.II.B (develop our own knowledge and cultural competency), II.II.C (recognise and address biases that may impact our work with clients and their families), II.II.G (recognise and respect cultural diversities in the implementation of professional practices), and II.II.I (facilitate understanding of the realistic limitations of each role). The preamble's third principle commits educational therapists to practice that is inclusive and non-discriminatory, naming national origin explicitly.


That is a profession that has written down the value and left the method entirely to the practitioner — the same pattern psychoeducational evaluation for educational therapists found on interpretation and educational therapy informed consent found on consent.


Can these scores tell you whether it is a language difference or a learning disability? No.


Flatly no, and it is worth saying so in one place.


Nothing in this post gets you to that answer, and none of the instruments above claims to. A score from a Spanish battery equated to English norms, a score from an English battery, and both together produce a profile, not a determination.


The distinction between a difference arising from language exposure and a disability is an identification question, it requires expertise and a team, and in schools it is governed by statute. It is not something an educational therapist settles from a score, and it is not something this article can help with.


What the published record does support is a sequence — the same move that achievement test qualification levels and math assessment tools for dyscalculia used, which is to give the decision order rather than the decision.


Five questions to ask of any manual before you report its score


  • Was there a separate normative sample in this language, or were the items equated to an English one? The Batería IV answers equated; the WISC-V Spanish answers both, a Spanish sample and then equating; Acadience Reading Español answers neither, because it is criterion-referenced.

  • How was the reference group defined by language? Years in US schools, primary language at home, language of instruction, English learner status — these are different variables and publishers use different ones.

  • How large is the subgroup my student would belong to? Not the headline sample size. The cell.

  • What language was my student taught to read in, and for how long? Then compare that to the reference group's. The mCLASS Lectura cut-score sample was being taught in Spanish; a student taught only in English is not in that population, whatever language she speaks at home.

  • Is the comparison I am about to make inside this instrument, or across two instruments? Across two, you are stacking two reference groups and two error bands, and Standard 2.4 of the testing standards asks for a separate error term for differences, as the standard error of measurement piece set out.


You cannot make the norms fit the child. You can describe, in a sentence, whose performance the number is being measured against — and that sentence is more useful to the next reader than the number.

What is out of scope here


This post is about what test manuals and publishers say about their own normative samples. Several adjacent questions are governed by law, regulation or contract rather than by publisher documentation, and they are not covered above:


  • Identification of a student as an English learner, and the assessments used for it

  • Eligibility for special education, and whether a language difference may be the determinant factor

  • Required English language proficiency assessment, and the consortia and instruments used for it

  • A school's obligations to a multilingual student or family, including interpretation and translation of documents

  • Which language a school is required to assess in


Those are statutory and regulatory questions. They belong with an attorney, with your state education agency, and in a school context with the district's own counsel. Nothing here is legal advice.


If what you actually need is the authority to make an identification decision, that is a question about your own credential, and the scope boundaries in educational therapist vs. learning specialist vs. executive function coach and in what is educational therapy are the place to start.


Where the published record runs out


Several things in this area could not be verified from primary sources, and it is more useful to list them than to paper over them.


The WMLS III's Spanish normative sample could not be retrieved. Riverside's product page and brochure describe the instrument and include a sample report stating that the Spanish version is compared against Spanish-speaking peers, but no technical manual or technical summary with a Spanish sample size is publicly readable, and a publisher webinar handout on the instrument's norms gives none either.


Two independent attempts failed. So the same-language comparison claim above is reported as what the sample report says, and the sample behind it is not asserted here. If that claim is load-bearing for a report you are writing, ask Riverside for the technical manual's norming chapter and keep the reply.


No English achievement battery's language-inclusion rule could be verified. Pearson's and other publishers' technical manuals are not public documents, so this post does not state whether bilingual students were included in or excluded from the normative samples of the WIAT-4, the KTEA-3 or the Woodcock-Johnson achievement batteries, or on what criterion.


That is the single most useful missing fact in this whole area, and it is in the manuals on your shelf. If you own one, the norming chapter's participant exclusion criteria will answer it in a paragraph.


The 2007 federal English learner guide was not re-read for this post. The 2014 guide refers readers to its screening recommendation, and that reference is reported here as the 2014 guide states it, not from the 2007 document itself.


Riverside's Arabiya and Logramos publish no sample information on their pages, so nothing is claimed about either instrument's norms beyond the absence itself.


There is no published comparison of these instruments' reference groups anywhere. The tables in this post were assembled from four publishers' own documents because nobody has put them side by side. If a publisher or researcher has done this work in a place I could not reach, the right move is to read theirs rather than mine.


Outbound sourcing for this post covers 13 distinct domains — four publishers, the testing standards, two federal agencies, the national clearinghouse and AET.


That is a small number by this blog's usual standard, and it is a property of the subject: the complete universe of primary sources here is the people who made these tests, the standards that govern them, and the federal record. Padding it would have meant citing other people's summaries of documents I can quote directly.


Common questions


Does administering a test in Spanish make the result valid for a Spanish-speaking student?


No, and the testing standards say so explicitly. Standard 3.13's comment states that determining a test taker's most proficient language "does not automatically guarantee validity of score inferences for the intended use", and adds that a student may be more proficient in one language than another without being developmentally proficient in either. Language of administration is one decision among several, not a fix.


The report says the student was tested on the Batería IV. What are those standard scores compared against?


Per Riverside's own bulletin, the Batería IV's standard scores and percentile ranks come from the same norms as the Woodcock-Johnson IV — a norming sample of 7,416 individuals from 46 states and the District of Columbia. The Spanish items were calibrated on a separate sample of 601 native Spanish speakers and equated onto that scale. There is no separate Spanish normative distribution.


My student has been in US schools since kindergarten. Is the WISC-V Spanish appropriate?


That is a clinical judgement and not mine to make, but the specification is public and relevant: the normative sample consists of children whose primary language is Spanish and who have attended US schools for less than five consecutive years.


A student with eight years in US schools falls outside the group the norms describe. Whoever administers it should be in a position to say what they think that means for the score.


Why do the Spanish early-reading benchmarks keep saying "preliminary"?


Because the publisher is being accurate. Acadience Learning states that Acadience Reading Español's benchmarks and cut points "were developed based on the corresponding English versions of these measures" and that continuing research will refine them. A preliminary criterion threshold is a reasonable thing to publish and a reasonable thing to use, as long as you describe it that way when you report against it.


Do the NCII tools charts tell me whether a tool works for English learners?


No. The words "Spanish", "bilingual" and "multilingual" appear zero times across all three charts, there is no language-of-administration filter, and no rating criterion addresses the language of the normative sample. NCII does publish separate guidance on intensive intervention for English learners, including a companion to its Taxonomy of Intervention Intensity, but that guidance is not part of the ratings.


What does AET require of me here?


The Code of Ethics never uses the words "bilingual", "multilingual" or "norm", and its only mention of language concerns communicating with families in the primary language of the home, at II.II.E. The clause that applies is II.I.B: select and use appropriate assessment instruments, "recognizing their limitations with respect to reliability, validity, and bias." The Code names the obligation and leaves the method to you.


Related reading


This post sits inside an assessment-literacy cluster built the same way — from the publishers' and standards bodies' own documents.



Get the practitioner list


If you want the next one of these when it lands — assessment documents read closely, with the sample tables actually opened — you can join the Illuminate educator list. It is written for educational therapists, learning specialists and executive function coaches, and nothing on it is for parents.

Recent Posts

See All
Standard Error of Measurement: What That Band Means

Every score report prints a confidence interval and almost no practitioner acts on it. What the standard error of measurement is, why the band's width is a software setting rather than a property of t

 
 
 

Comments

bottom of page