Source URL: View this document on the ASHA website
This study illustrates how important it is for an evaluator to be familiar with both the typical language practices of the communities they work in as well as the bias inherent in many of the standardized tests that are used to determine disability.
Background: Two studies compared the performance of Puerto Rican and African American Head Start children on presumably familiar (description) and unfamiliar (one-word labeling) test tasks.
Hypotheses: Peña and Quinn wanted to see if familiarity of the task presented influenced the performance of the children being tested.
Methodology and Participants: Fifty (41 typically developing and 9 of “low language ability”) Puerto Rican and African American children enrolled in Head Start programs from the same community were each given two language assessments. The first test was the Expressive One Word Picture Vocabulary Test- Revised (EOWPVT-R) and the second was the Comprehension Subtest of the Stanford Binet Intelligence Scale (SBIS). The EOWPVT-R is a traditional vocabulary test, which asks children to label pictures in a flipbook while the Comprehension Subtest of the SBIS examines the function of objects. It had been noted in the literature that the home language socialization practices of Standard American English (SAE) speaking, middle class (socioeconomic status, SES) individuals often focused on labeling tasks while parent-child interactions in culturally and linguistically diverse groups tended more toward describing the function of objects.
Conclusion: Results indicated that the EOWPVT-R did not distinguish typically developing (TD) children from children with language impairment. In fact, 90% of the children scored below the EOWPVT norms and would have been classified as language impaired. The Comprehension Subtest of the SBIS however did distinguish between the two groups of children, lending support to the authors’ hypothesis that task familiarity impacts performance on language tests.
Relevance to the Field: This study illustrates how important it is for an evaluator to be familiar with both the typical language practices of the communities they work in as well as the bias inherent in many of the standardized tests that are used to determine disability. The results of this research proved that labeling and vocabulary tests do not distinguish between typically developing and language impaired children from culturally and linguistically diverse backgrounds. It also demonstrated that asking children from these backgrounds to describe the function of objects, rather than labeling objects is less biased against them and may provide more useful information about their language development. Diagnosing language impairment after only administering standardized vocabulary tests, such as the EOWPVT-R, is likely over-referring, misdiagnosing, and subjecting these children to inappropriate and inadequate placement in special education classrooms.
Peña, E. D., & Quinn, R. (1997). Task familiarity: Effects on the test performance of Puerto Rican and African American children. Language, Speech, and Hearing Services in Schools, 28, 323-332.