What is the CELF Preschool-3? A guide to the preschool language test

The CELF Preschool-3 (Clinical Evaluation of Language Fundamentals Preschool, Third Edition) is an individually administered test of spoken language for children ages 3 years 0 months through 6 years 11 months. It measures language skills, such as understanding sentences and using words correctly, rather than intelligence. Published by Pearson in 2020 and authored by Elisabeth Wiig, Wayne Secord, and Eleanor Semel, it is one of the most widely used tools speech-language pathologists rely on when a young child is referred for a possible language delay. This article covers the test's contents, its scoring, how it differs from the school-age CELF-5, and what parents should keep in mind.

According to Pearson, the CELF Preschool-3 "assesses aspects of language necessary for preschool children to meet the language demands of the classroom." In practice, that means it samples two broad channels of communication: "receptive language," a child's ability to understand what is said, and "expressive language," a child's ability to put thoughts into words.

The test is built from a set of short subtests, each targeting a specific skill. The main scaled-score subtests are:

• Sentence Comprehension: the child hears a sentence and points to the picture that matches it, showing how well they interpret sentences of increasing length and complexity.

• Word Structure: the child completes a spoken sentence, which reveals whether they can apply grammar rules such as plurals, verb tenses, and pronouns.

• Expressive Vocabulary: the child names people, objects, and actions shown in pictures.

• Following Directions: the child points to pictures in response to spoken directions that grow longer and more complex.

• Recalling Sentences: the child repeats sentences word for word, a task that draws on both grammar knowledge and verbal memory.

• Basic Concepts, Word Classes, and Phonological Awareness: these subtests cover concept words (such as position, size, and quantity), the ability to see how words relate to each other, and awareness of the sound structure of language through rhyming and syllable tasks.

Beyond these, the examiner and a familiar adult complete rating scales covering the child's social use of language and emerging reading and writing skills, and two supplemental criterion-referenced measures (a Connected Speech Sample and a Pragmatic Activities Checklist) let the clinician observe storytelling and real conversation. Pearson notes that this edition added two new composite scores, an Academic Language Readiness index and an Emerging Literacy index, along with "growth scale values" that track progress over time.

The CELF Preschool-3 is given one-on-one by a trained examiner, most often a speech-language pathologist. Pearson describes the response format as "verbal response to stimulus picture," and the materials are designed around what a three- to six-year-old can actually do: the child looks at colorful stimulus pictures, points, names, and repeats. The print kit even includes a storybook, titled No Juice, used for the narrative task. Nothing requires reading or writing, which matters because most examinees cannot yet read.

The publisher lists a completion time of 15 to 20 minutes for the first level of testing, with total time described as variable because examiners choose which additional subtests a particular child needs. The test can be given with paper materials or digitally through Pearson's Q-interactive tablet system, and it carries a qualification level of B, meaning purchasers need relevant graduate training or certification in assessment.

Like most modern standardized tests, the CELF Preschool-3 reports "standard scores" set to a mean of 100 with a "standard deviation" (a measure of how spread out scores are) of 15. Pearson's score reports state this explicitly for the Core Language Score and each index: a score of 100 "represents the performance of the typical child of a given age." Roughly two thirds of children score between 85 and 115. Individual subtests use a smaller scale with a mean of 10 and a standard deviation of 3.

The headline number is the Core Language Score, which Pearson describes as "a measure of general language ability" derived from three subtests. In the publisher's sample report for a five-year-old, those are Sentence Comprehension, Word Structure, and Expressive Vocabulary. Around it sit several index scores: the Receptive Language Index (built from listening tasks such as Following Directions), the Expressive Language Index (built from speaking tasks such as Recalling Sentences), plus indices for language content, language structure, academic language readiness, and emerging literacy. Reports also include percentile ranks, age equivalents, and confidence intervals, which show the range where a child's true score likely falls.

A preschooler's attention, mood, and comfort with a stranger can all nudge a single day's score, so careful clinicians interpret the confidence interval and the pattern across subtests rather than treating one number as the final word.

The CELF Preschool-3 and the CELF-5 come from the same author team and share the same basic architecture, but they serve different age groups. The preschool edition covers ages 3:0 to 6:11. The school-age edition, described in our companion article on what the CELF-5 is, covers ages 5:0 through 21:11 and was published in 2013.

For children ages 5 and 6, where the two tests overlap, clinicians can choose either. Pearson's technical summary reports that composite scores on the two tests correlate .65 to .76 in that overlap range, so they measure similar ground while remaining distinct instruments. The preschool version uses shorter tasks, more picture support, and content pitched to children who have not started formal schooling, while the CELF-5 adds tasks tied to classroom language, reading, and writing. For children younger than 3, examiners turn to infant and toddler measures instead, such as the parent-report instrument covered in our article on the REEL-4.

Federal special education law lists "speech or language impairment" as a disability category, defined in the regulations as a communication disorder "that adversely affects a child's educational performance." Preschoolers can qualify for services under this category, and a norm-referenced language test given by a qualified examiner is usually part of the evidence. The CELF Preschool-3 is well suited to that role because it was normed on 700 children ages 3:0 to 6:11 drawn from 41 states, and because Pearson publishes classification accuracy data: at a cutoff of 1.5 standard deviations below the mean, the test showed "sensitivity" (the ability to catch true cases) of .89 and "specificity" (the ability to clear typical children) of .87, which the publisher calls the optimal balance.

Two cautions belong alongside those numbers. First, the American Speech-Language-Hearing Association is clear that a standardized test is one component of a comprehensive evaluation that should also include a case history, a hearing screening, and informal observation. ASHA also advises against "cognitive referencing," the old practice of requiring a gap between IQ and language scores before granting services. Second, no test is perfect: even at the optimal cutoff, roughly one in ten children with genuine language disorder scores above the line, and a similar share of typically developing children scores below it. Eligibility decisions should never rest on a single score.

The CELF Preschool-3 has real strengths. Its index scores show internal consistency reliability of .90 to .93 in the normative sample, test-retest stability of .80 to .93, and documented agreement with related measures such as the PLS-5 and the CELF-5. Its picture-based tasks fit young attention spans, and the growth scale values give clinicians a defensible way to measure progress during therapy.

Its limitations are equally worth naming. Individual subtest reliabilities run lower (.73 to .92), so single-subtest interpretations are shakier than composite ones. Like any norm-referenced language test, it samples behavior in a structured setting on one day, and it reflects a child's exposure to the language and dialect of the test. For children learning more than one language or speaking a nonmainstream dialect, scores must be interpreted with particular care.

Above all, the CELF Preschool-3 is a language test, and only a language test. A low Core Language Score means the child struggled with language tasks relative to age peers; it says nothing direct about reasoning, memory for visual patterns, or overall intellectual ability. Related instruments make the same point from other angles: oral language batteries such as the OWLS-II also measure language proficiency rather than intelligence. Measuring intelligence requires a test built and normed for that purpose. For adults who want that kind of measure, the Reasoning and Intelligence Online Test, developed by RIOT IQ with psychometrician Dr. Russell T. Warne for ages 18 and up, uses 15 subtests across six cognitive indices (verbal reasoning, fluid reasoning, spatial ability, working memory, processing speed, and reaction time), takes about 52 minutes, and reports scores on the same mean-100, SD-15 scale. It does not replace an individually administered diagnostic evaluation, and you can take the RIOT IQ test online.

No. It measures spoken language skills (understanding and producing words, sentences, and directions) in children ages 3:0 to 6:11. It is not designed to measure intelligence, and a language score should never be read as an IQ.

Ages 3 years 0 months through 6 years 11 months, according to Pearson. Children ages 5 and 6 can be tested with either the preschool edition or the school-age CELF-5.

Index scores have a mean of 100 and a standard deviation of 15. Pearson's research found that a cutoff of 1.5 standard deviations below the mean (a score of about 77 or 78) gave the best balance of sensitivity and specificity for identifying language disorder, though clinicians weigh scores alongside other evidence.

It is individually administered by professionals with graduate training in assessment, most commonly speech-language pathologists. Pearson assigns it qualification level B.

Pearson lists 15 to 20 minutes for the first level of testing; the full battery is variable because examiners select additional subtests based on the referral question.

No. Under federal special education rules, eligibility requires a comprehensive evaluation, and ASHA recommends combining standardized scores with a case history, hearing screening, observation, and informal measures.

1. Pearson. (2020). Clinical Evaluation of Language Fundamentals Preschool-3 (CELF Preschool-3) [Product page]. pearsonassessments.com

2. Pearson. (2020). CELF Preschool-3 technical summary. pearsonassessments.com

3. Pearson. (2020). CELF Preschool-3 Q-interactive sample score report. pearsonclinical.ca

4. Pearson. (2013). Clinical Evaluation of Language Fundamentals, Fifth Edition (CELF-5) [Product page]. pearsonassessments.com

5. American Speech-Language-Hearing Association. (n.d.). Spoken language disorders [Practice portal]. asha.org

6. U.S. Department of Education. (2017). 34 CFR 300.8(c)(11): Speech or language impairment. sites.ed.gov

New McGuffey First Reader (1901), via University of Washington Libraries / Wikimedia Commons (public domain)