Hayward, Stewart, Phillips, Norris, & Lovell 3

At-a-Glance Test Review: Clinical Evaluation of Language Fundamentals-4 (CELF-4)

Name of Test: Clinical Evaluation of Language Fundamentals-4 (CELF-4)
Author(s): Semel, E., Wiig, E., and Secord, W.
Publisher/Year (Please provide original copyright as well as dates of revisions): PsychCorp 1980, 1987, 1995, 2003
Forms: Form 1 ages 5-8 years and Form 2 ages 9-21 years
Age Range: 5 years, 0 months, to 21 years, 11 months
Norming Sample
Total Number: 2 650
Number and Age: 200 students at each age level from 5 years to 17 years, 50 students in each age years from 17 to 21
Location: 47 states in four regions
Demographics: age, gender, race/ethnicity, geographic region.
Rural/Urban: not specified
SES: stratified by four parent education levels: < Grade 11, Grade 12 completion, college, and university
Other (Please Specify): “children receiving special services” constituted 9% of sample with 7% diagnosed of sample diagnosed with language disorders; numbers consistent with data from National Dissemination Center for Children with Disabilities (2003) and US Office of Education Program. The authors note that previously CELF-3 did not include these children in the standardization sample.
Summary Prepared By (Name and Date): Eleanor Stewart 12 Jul 07, additions 16 Jul
Purpose of Test: to identify, diagnose and monitor language and communication disorders in students aged 5 to 21. (p.1)
Theoretical Model: not provided
Theory: The developmental stages model uses milestones identified from extensive research.
1.  concepts and following directions
2.  word structure
3.  recalling sentences
4.  formulated sentences
5.  word classes (1 and 2)
6.  sentence structure
7.  expressive vocabulary
8.  word definitions
9.  understanding spoken paragraphs
10.  semantic relationships
11.  sentence assembly
12.  phonological awareness: 17 tasks criterion referenced (meets age or not)
13.  rapid automatic naming (RAN): criterion referenced (normal, slower, non-normal for speed and normal, more than normal, and non-normal for number of errors)
14.  word associations: criterion referenced
15.  number repetition (1 and 2)
16.  familiar sequences (1 and 2)
17.  pragmatics profile
18.  observational rating scale: in accordance with US education legislation.
Phonological awareness includes: detection, identification, blending, segmenting across word, syllables and phonemes.
Who Can Administer: “Diagnosticians who have been trained and are experienced in administration and interpretation of individually administered, standardized tests” (Semel, Wiig, & Secord, 2003, p. 1).
Administration Time: Time varies with number of subtests administered as well as age of student and other student characteristics.
Test Administration (General and Subtests): There are a total of 18 subtests but not all subtests are administered. This is where the assessment model is useful for selecting from among the subtests.
The record form includes the specific instructions for each subtest.
Test Interpretation: Chapter 3 provides a detailed interpretation of results for assessment levels 1 and 2. Chapter 4 continues with a description of levels 3 and 4 which includes “when to administer subtests to evaluate related clinical behaviors”, “criterion-referenced subtest scores”(phonological awareness, word associations, etc.), authentic and descriptive assessment measures (pragmatic profile and ORS).
Standardization: Age equivalent scores (by subtest scores) Grade equivalent scores Percentiles Standard scores (and subtest scaled scores) Stanines Normal Curve Equivalents
Other (Please Specify) Core Language and Index Scores which are standard scores in Appendix D.
Criterion-referenced subtest scores (Appendix G) association (age 5-21 years), phonological awareness (age 5-12 years, 11 months), rapid automatic naming in time per second and error (5-21 years,11 months), and pragmatic profile (5-21 years,11 months)
Reliability:
Internal consistency of items: Composite scores yielded high alpha coefficients in the range from .89 to .95. Subtest coefficients were lower in the range from .70 to .91.
Test-retest: 320 students were retested with high correlations on composite scores (demonstrated with .90+ reported for all age groups). Subtest scores were less robust with ranges from .60 to .90. The average interval in days was 16.
Inter-rater: Two raters were chosen randomly from a pool of 30 raters who had trained under the supervision of the test developers. Agreement for subtests that required scoring judgments were reported. High agreement was evidenced (.90 to .98).
Other: Reliability for clinical groups was reported (405 students in four groups: language disorder, mental retardation, autism, and hearing impairment). Authors include inclusion criteria and demographic information.
SEMs and confidence intervals were provided for 68%, 90%, and 95% levels.
Validity:
Content: Based on literature review, expert panel review of content, clinician feedback.
Criterion Prediction Validity: studies, high correlations with CELF-3.
Construct Identification Validity: Factor Analysis research. The Buros reviewer states, “ comprehensive intercorrelational and factor analytic analyses confirmed the basic construct validity of the instrument” (Langlois & Samar, 2005, p. 222).
Differential Item Functioning:
Other: In clinical validity studies, CELF-4 differentiates clinical populations of children with language disorders, autism, mental retardation, and hearing impairment.
Summary/Conclusions/Observations: *The authors were attentive to the U.S. legislative changes that required testers to address “curricular goals and academic benchmarks” (p. 198) thus specifically creating the rationale for the Observational Rating Scale (ORS).
*There is less of an appeal for Canadian clinicians who must deal with provincial variations. I wonder how much variation there is given the differences in resources and, perhaps, in working definitions of clinical categories? Is there such information?
Clinical/Diagnostic Usefulness: The CELF-4 is intended as a diagnostic tool and it is used by clinicians for exactly that purpose. That the sensitivity has been proven is one of the strengths of this test as clinicians are most concerned about accurately identifying children with language problems particularly when funds are limited as they most often are. The test authors and developers have provided us with a strong tool that can be linked to classroom language performance.

References

Langlois, A., & Samar, V. J. (2005). Review of the Clinical Evaluation of Language Fundamentals-4. In R.A. Spies, and B.S. Plake, (Eds.), The sixteenth mental measurements yearbook (pp. 217-223). Lincoln, NE: Buros Institute of Mental Measurements.

Semel, E., Wiig, E. & Secord, W. (2003). Clinical evaluation of language fundamentals -4 (CELF-4). San Antonio, TX: PyschCorp.

To cite this document:

Hayward, D. V., Stewart, G. E., Phillips, L. M., Norris, S. P., & Lovell, M. A. (2008). At-a-glance test review: Clinical evaluation of language fundamentals-4 (CELF-4). Language, Phonological Awareness, and Reading Test Directory (pp. 1-3). Edmonton, AB: Canadian Centre for Research on Literacy. Retrieved [insert date] from http://www.uofaweb.ualberta.ca/elementaryed/ccrl.cfm.