Download the white paper (PDF)
Summary
When a preschooler can't complete a standardized test as the manual requires, the standard score may not be valid. This guide shows how to write a validity statement, log the supports you tried, word observations respectfully and build the evaluation from parent report, observation, a play based language sample and dynamic assessment.
Read the companion post on invalid standardized test results.
Key points
- IDEA requires that assessments be valid and reliable for their purpose and administered according to the producer's instructions. When that doesn't happen, the standard score may not mean what it seems to mean.
- ASHA's school documentation format includes a validity statement that explains whether the administration and interpretation were valid and describes any procedural modifications.
- Document what you attempted, what supports you tried and what you observed, in neutral language that respects the child.
- Build the evaluation from multiple sources: parent report, observation, a play based language sample and, where it fits, dynamic assessment.
- A language sample is language evidence. Speech sound skills need their own measures, chosen by the SLP.
- The group of qualified professionals and the parent decides eligibility. EASI never decides, diagnoses or qualifies a child.
The question behind this guide
In early November 2026, a school SLP asked colleagues how to document a standardized test that was listed on the consent form but could not be completed with a child who was almost 4 and could not attend to the task. The replies were practical and kind: list what you tried to help the child succeed, avoid wording that sounds like you are blaming the child, and make sure the report accounts for every test named on the consent.
Every preschool clinician has a version of this story. This guide gathers the regulations, professional guidance and research that support good practice when a test breaks down, and offers templates you can adapt.
Part 1: Why tests break down with young children
Standardized tests are built on the assumption that each child takes the test under the same conditions, answering the same items in the same way, so that scores can be compared with a normative group (ASHA, Assessment Tools, Techniques, and Data Sources). That comparison holds only when the child's testing experience resembles what the normative sample experienced.
Young children often cannot meet the demands of a structured test on a given day. They may not yet sit for a series of picture plates, respond to an unfamiliar adult on request, or tolerate a new room. None of that tells you what the child can do with language. It tells you the test conditions did not fit the child that day.
ASHA's guidance on late language emergence is direct: SLPs must ensure that standardized measures show robust psychometric properties, and when a child is not represented in a test's normative group, the test cannot be used to determine the presence or absence of a communication disorder, even when it is administered as instructed. ASHA also describes non normed measures, observation, language sampling, ethnographic interviewing and dynamic assessment as fundamental to telling a language difference from a disorder, and says multiple sources of information are necessary (ASHA, Late Language Emergence).
Part 2: What the regulations require
Under 34 CFR 300.304, the public agency must:
- use a variety of assessment tools and strategies to gather relevant functional, developmental and academic information, including information provided by the parent (300.304(b)(1))
- not use any single measure or assessment as the sole criterion for an eligibility decision (300.304(b)(2))
- provide and administer assessments in the form most likely to yield accurate information on what the child knows and can do academically, developmentally and functionally (300.304(c)(1)(ii))
- use assessments for purposes for which they are valid and reliable (300.304(c)(1)(iii)), administered by trained and knowledgeable personnel (300.304(c)(1)(iv)) and in accordance with any instructions provided by the producer (300.304(c)(1)(v))
- assess the child in all areas related to the suspected disability, including communicative status (300.304(c)(4))
Read together, these provisions point one way. If a test could not be given according to its instructions, the result should not be treated as a valid standard score, and the evaluation must still be sufficiently comprehensive through other tools and strategies (300.304(c)(6)).
The eligibility decision is made by a group of qualified professionals and the parent, who must draw on information from a variety of sources and document and carefully consider it (34 CFR 300.306(a)(1), (c)(1)).
Part 3: The validity statement
What ASHA describes
ASHA's Practice Portal page on documentation in schools outlines a standard assessment report format, drawing on Moore (2010a, 2010b, as cited by ASHA), that includes a validity statement: a statement regarding whether the assessment administration and interpretation procedures were valid for the intended purposes and accurately reflect the student's current functioning and performance, including procedural modifications. The same page lists an explanation of any nonstandard test administration among the elements to document. ASHA's example reads: "Assessment was attempted; however, based on the student's cognitive functioning and/or behaviors, the assessment tool was not appropriate or did not accurately measure the student's performance" (ASHA, Documentation in Schools).
A decision path
- Before testing: Read the administration chapter. Note what the manual allows: demonstration and practice items, repetition, breaks, testing across sessions, basal and ceiling rules.
- During testing: Try the supports the manual allows first. Keep a short log of what you tried and when.
- If the child cannot continue: Stop. Decide whether another session, within the manual's rules, is reasonable.
- After testing: Decide whether the administration stayed within standardized procedures. If it did, report scores with notes. If it did not, report no standard scores and describe performance qualitatively.
- In the report: Write the validity statement, then present the other evidence you gathered.
Support log
Use this log during the session so the report can describe exactly what happened.
| Support attempted | Allowed by the manual? | When | Child's response |
|---|---|---|---|
| Rapport building play before testing | |||
| Familiar adult in the room | |||
| Quieter room or reduced visual distractions | |||
| Short movement breaks | |||
| Fidget or seating change | |||
| Extra demonstration or practice items | |||
| Reinforcement between items | |||
| Second session at a different time of day |
Respectful wording
Describe behavior, not character. Compare:
- Less helpful: "The child was noncompliant and refused to test."
- More helpful: "[Child] looked at 3 picture plates, then moved to the toy shelf. After a movement break and a return to the table with a parent nearby, [child] looked at 1 more plate before leaving the table again."
The second version gives the team useful information about attention and engagement without blaming the child.
Templates you can adapt
Discontinued test: "The [test] was attempted on [date] and [date]. Supports included [list]. [Child] engaged with [number] items before leaving the task. Because the test could not be administered according to the standardized procedures, results are not considered a valid estimate of [child]'s skills, and standard scores are not reported. [Child]'s communication was assessed through the measures described below."
Modified administration: "The [test] was administered with the following modifications, which are outside the standardized procedures: [list]. Results are described qualitatively and are not reported as standard scores."
Test listed on consent but not attempted: "The [test] was listed on the consent for evaluation. It was not administered because [reason]. Information in this area was gathered through [other measures]."
Check your district's procedures for tests listed on consent, prior written notice and reevaluation timelines. These vary.
Part 4: Building the evidence base
Parent report
IDEA names information provided by the parent as part of the evaluation (34 CFR 300.304(b)(1)), and existing data reviews include information from parents (34 CFR 300.305(a)(1)(i)). Ask what the child says at home, how the child asks for things, whether familiar and unfamiliar listeners understand the child, and what situations are hard.
Observation
Observe in a natural setting when possible, such as a preschool classroom, a play group or a play routine at home or in the clinic. Note how the child starts interactions, responds to peers and adults, and uses words, gestures and sounds.
A play based language sample
ASHA describes language sampling as eliciting spontaneous language in contexts such as free play, conversation, narration and expository speech, then deriving measures to complement standardized assessment (ASHA, Assessment Tools, Techniques, and Data Sources).
Choosing the context. Southwood and Russell (2004) compared conversation, free play and story generation in a small sample of 5 year olds and found that free play produced more utterances than story generation, while story generation produced longer utterances and conversation and story generation produced a greater share of complex structures. For a preschooler who has just walked away from a test table, free play with a familiar partner is often the most natural start.
Choosing the partner. Guo and Eisenberg (2015) used parent elicited conversational samples during play with 3 year olds. A parent can be an excellent partner. Coach the parent briefly: follow the child's lead, comment more than you question, and wait.
Choosing the length. Heilmann, Nockerts and Miller (2010) found that measures of productivity, lexical diversity and utterance length were quite reliable even in short samples of conversation and narrative. Guo and Eisenberg (2015) concluded that 7 to 10 minutes of parent elicited talk is desirable for 3 year olds, and cautioned that very short samples may lead to misdiagnosis. For high stakes decisions, plan for more talk. See How Long Should a Language Sample Be?. For MLU counting rules, see how to calculate MLU.
Choosing the measures. Common measures include mean length of utterance, number of different words and total number of words. Eisenberg and Guo (2013) found that the percentage of grammatical utterances differentiated 3 year olds with and without language impairment in their study of 34 children. Rice and colleagues (2010) published MLU reference values in six month intervals for children 3 to 9 years with and without language impairments. If you compare to reference values, match the procedures used to collect them, and report the context, length and number of utterances analyzed.
Recording. Record the sample so you can review the transcript later and so your measures rest on what the child actually said.
Planning sheet for the play sample
| Decision | Your plan |
|---|---|
| Context (free play, book sharing, routine) | |
| Partner (parent, clinician, both) | |
| Materials (child's favorites, a few novel toys) | |
| Target length (minutes and utterances) | |
| Measures to calculate | |
| Reference values to compare, and their procedures | |
| Who reviews the transcript |
Dynamic assessment
Dynamic assessment uses a test, teach, retest approach to see how the child responds to teaching. Peña, Iglesias and Lidz (2001) studied culturally and linguistically diverse preschool children and found that measures from a mediated learning approach to word learning distinguished children with typical and low language ability better than static pretest measures. ASHA notes that dynamic assessment can be used in conjunction with standardized assessments and language sampling (ASHA, Assessment Tools, Techniques, and Data Sources).
Part 5: Speech sounds are a separate question
The test in the original question assessed speech sounds. A language sample answers questions about language: vocabulary, sentence length and grammar. It does not replace a speech sound assessment.
ASHA describes speech sound assessment as using standardized instruments, as appropriate, along with other sampling procedures to evaluate production in single words and connected speech, and describes stimulability testing as a way to see how well a child imitates sounds the child does not yet produce correctly (ASHA, Speech Sound Disorders: Articulation and Phonology). Choose those measures for this child, and apply the same validity thinking: if a speech sound test could not be given as standardized, document that, and describe what you observed.
This guide and the EASI workflow below focus on language.
Part 6: Writing it up
A clear report for this child might follow this order:
- Reason for referral and background, including parent concerns.
- Assessments attempted, including the validity statement for any test not given as standardized.
- Parent report and observation.
- Language sample: context, partner, length, utterances analyzed and measures, each from a reviewed transcript.
- Dynamic assessment, if used.
- Speech sound findings from the measures you chose.
- Summary of strengths and needs, and how communication affects participation in the child's preschool or daily routines.
ASHA's documentation guidance notes that eligibility is defined by state and district and decided by the IEP team, and that recommendations for eligibility are a team decision typically not included in the evaluation report (ASHA, Documentation in Schools).
Part 7: Where EASI fits
EASI is language sample analysis software made by iTherapy, LLC. You can record a play sample live using the interactive backgrounds and high interest pictures built in, upload a recording, or bring in an existing transcript. EASI analyzes English language samples. The workflow is Draft, Review, Attest, Score. EASI drafts the transcript. You play each line, confirm who spoke, fix any words, and attest each line. Only then, when the sample supports it, does EASI calculate language measures such as mean length of utterance, number of different words, vocabulary diversity and the Index of Productive Syntax for you to review. A short sample is reported as a short sample. You decide whether it is adequate, which norms apply and what the results mean. Learn more about EASI language sample analysis software.
MySLP, the secure AI assistant built into EASI, answers SLPs' clinical, scientific and eligibility questions. You can ask it how to word a validity statement or plan a play sample, then review and edit what it suggests. Report drafting (in beta) can bring your reviewed findings together into a draft that you edit and sign.
EASI is HIPAA and FERPA compliant, with a BAA and insurance. EASI never decides eligibility, diagnoses or qualifies a child. You and your team stay in charge.
Get EASI, with MySLP included, at NorthernSpeech.com/EASI. Northern Speech Services is the exclusive distributor.
References
- American Speech Language Hearing Association. (n.d.). Assessment tools, techniques, and data sources [Practice Portal]. https://www.asha.org/practice-portal/resources/assessment-tools-techniques-and-data-sources/
- American Speech Language Hearing Association. (n.d.). Documentation in schools [Practice Portal]. https://www.asha.org/practice-portal/professional-issues/documentation-in-schools/
- American Speech Language Hearing Association. (n.d.). Late language emergence [Practice Portal]. https://www.asha.org/practice-portal/clinical-topics/late-language-emergence/
- American Speech Language Hearing Association. (n.d.). Speech sound disorders: Articulation and phonology [Practice Portal]. https://www.asha.org/practice-portal/clinical-topics/articulation-and-phonology/
- Assistance to States for the Education of Children With Disabilities, 34 CFR 300.304, 300.305, 300.306. https://www.law.cornell.edu/cfr/text/34/part-300
- Eisenberg, S. L., & Guo, L. Y. (2013). Differentiating children with and without language impairment based on grammaticality. Language, Speech, and Hearing Services in Schools, 44(1), 20 to 31. https://doi.org/10.1044/0161-1461(2012/11-0089)
- Guo, L. Y., & Eisenberg, S. (2015). Sample length affects the reliability of language sample measures in 3 year olds: Evidence from parent elicited conversational samples. Language, Speech, and Hearing Services in Schools, 46(2), 141 to 153. https://doi.org/10.1044/2015_LSHSS-14-0052
- Heilmann, J., Nockerts, A., & Miller, J. F. (2010). Language sampling: Does the length of the transcript matter? Language, Speech, and Hearing Services in Schools, 41(4), 393 to 404. https://doi.org/10.1044/0161-1461(2009/09-0023)
- Peña, E., Iglesias, A., & Lidz, C. S. (2001). Reducing test bias through dynamic assessment of children's word learning ability. American Journal of Speech Language Pathology, 10(2), 138 to 154. https://doi.org/10.1044/1058-0360(2001/014)
- Rice, M. L., Smolik, F., Perpich, D., Thompson, T., Rytting, N., & Blossom, M. (2010). Mean length of utterance levels in 6 month intervals for children 3 to 9 years with and without language impairments. Journal of Speech, Language, and Hearing Research, 53(2), 333 to 349. https://doi.org/10.1044/1092-4388(2009/08-0183)
- Southwood, F., & Russell, A. F. (2004). Comparison of conversation, freeplay, and story generation as methods of language sample elicitation. Journal of Speech, Language, and Hearing Research, 47(2), 366 to 376. https://doi.org/10.1044/1092-4388(2004/030)