Alternative assessment for transitional readers

Bilingual Research Journal, Summer 2002 by Beaumont, Carol, de Valenzuela, Julia Scherba, Trumbull, Elise

Abstract

This paper presents an analysis of one district's performance based assessment of reading comprehension specifically designed for 5th grade transitional bilingual students in their first year of English instruction. In contrast to the common practice of using one assessment measure for all students and modifying assessment conditions for students from diverse backgrounds, they created a new instrument which was administered instead of the district-wide performance based assessment. Findings from the study indicate that: (a) characteristics of second language learning in the areas of phonology, syntax, and semantics were present to varying degrees in 100% of the students' responses; (b) specific knowledge ofthe linguistic characteristics of transitional students was necessary for effective interpretation and accurate scoring of student responses; (c) the transitional students demonstrated the ability to comprehend and interpret English texts at inferential levels when assessed on the transitional assessment measure; and (d) the transitional performance-based assessment results suggested to teachers many possible changes they could make in their instruction to meet the literacy needs of transitional students.

Equitable assessment for English language learners remains one of the greatest challenges of educational reform. The complex interaction of language, culture, educational background, and opportunities to learn make equitable, large-scale assessments particularly difficult to achieve (Figueroa, 1990; Garcia & Pearson, 1994; Solano-Flores & Nelson-Barber, 2001; Ulibarri, 1985). Often, English language learners (ELLs) are assessed through the same mechanisms as their English-only counterparts. This practice compromises the accuracy of results and may lead to inappropriate interpretations about English language learners' proficiency in reading and academic content areas (Garcia, 1991; Miramontes, Nadeau & Commins 1997; Valdis & Figueroa, 1994).

The purpose of this paper is to describe student performance on a specially designed assessment of reading comprehension created for native Spanish-speaking students in their first year of English instruction. This study was guided by the following questions: (a) How do scoring and student evaluation practices in this assessment account for linguistic characteristics of first-year English language learners? (b) What evidence exists that reflects first-year English language learners' ability to process English texts at an inferential level? (c) How can this transition assessment inform curriculum and instruction for these students?

Assessment Reform and English Language Learners

In the decade-long debate on school reform, some educators favor new methods of assessment to remedy persistent problems in the American schooling. They have suggested that complex, cognitively demanding assessments can drive improvements in curriculum and instruction (e.g., Darling-Hammond, 1994; Mitchell, 1992; O'Day & Smith, 1993). They have also pointed out the overrepresentation of ELLs and other students from non-dominant cultures among poor performers on standardized tests (Figueroa, 1990; Valdes & Figueroa, 1994; Williams, 1996) and the consequent reinforcement and extension of social and educational inequalities (DarlingHammond, 1994, Duran, 1989; Oakes, 1985).

ELLs' lack of full English proficiency is an obvious reason for their poor performance on standardized tests (LaCelle-Peterson & Rivera, 1994; Cummins, 1989; Genesee & Hamayan, 1994), but it reflects a limited understanding of the factors contributing to a student's performance on an assessment or standardized test. To begin to grasp the relationship between English proficiency and test performance requires a deep understanding of the developmental nature of second language acquisition, the difference between surface fluency and the ability to participate in content area instruction in English, the time required to achieve academic fluency, the kinds of language conventions used on typical assessments and tests, and the language skills necessary for success on standardized tests. Because the length of time required for full English proficiency is often underestimated (Collier; 1988; Cummins, 1989; Valdes, 1998), and students may appear to be more proficient with English than they are, they may be expected to take tests in English long before they are fully proficient with the kind of academic language needed to perform well. All of these factors should be considered in designing appropriate assessment and instruction.

Standardization itself rules out any contextualization of assessment, meaning that linguistic differences among students cannot be accounted for adequately. The practice of assessing English language learners with the same mechanisms as their English-only counterparts may seriously compromise the validity of results and lead to misleading interpretations and unfair decisions affecting their futures (August & Hakuta, 1997; Garcia, 1991; LaCelle-Peterson & Rivera, 1994; Miramontes, Nadeau & Commins, 1997; Vald6s & Figueroa, 1994).