persiapan tugas lte

Testing strategies involving least biased language assessment of bilingual children.

Increased attention is being given to strategies for the least biased language assessment of culturally and linguistically diverse children for whom English is a second language. One main reason is the concern that the administration of language tests in English using current norms is inappropriate for students whose dominant language is other than English. This article discusses a variety of alternative assessment approaches, including renorming a test for the specific population being tested, using dynamic assessment techniques to assess a child's modifiability, using other nonstandardized measures, and modifying standardized tests. Each of these alternatives has advantages and disadvantages. In addition, results from a preliminary investigation using modifications of four subtests of the Clinical Evaluation of Language Fundamentals-Third Edition (CELF-3) are presented to further the dialogue about language assessment issues. Twenty-eight bilingual Latino children, ages 8 and 9 years, were administered the CELF-3 and the modified version of the test in counterbalanced order. Clinical implications for creating least biased language assessment strategies are included as the authors provide a tutorial on the broad topic of language assessment of culturally and linguistically diverse children for whom English is a second language.
Considerable attention has been given in recent years to the least biased language assessment of culturally and linguistically diverse children for whom English is a second language (Brice, 2002; Cheng, 1991; Goldstein, 2000; Hamayan & Damico, 1991; Kayser, 1995, 1998; Langdon, with Cheng, 1992; Langdon & Saenz, 1996; Lidz & Pena, 1996; Pena, Iglesias, & Lidz, 2001; Pena & Quinn, 1997; Restrepo, 1998; Roseberry-McKibbin, 2002; Tzuriel, 2001; Ukrainetz, Harpell, Walsh, & Coyle, 2000). The consensus is that the administration of language tests in English using current norms is inappropriate for students whose dominant language is not English.
There are a number of reasons for this consensus. Typically developing students who are acquiring English as a second language may score lower than monolingual peers on standardized tests involving academic language skills, even when they are fluent in conversational contexts (Cummins, 1984). Tests that are normed in English may include members of different ethnic and racial groups, but the percentage of culturally and linguistically diverse students included in norming samples typically is no more than that included in the population of the United States as a whole. In such cases, for many English standardized tests, Asian and Latino children's scores are still compared primarily with those of middle class, monolingual Euro-American children.
Another issue with standardized testing is the fact that children from low-income backgrounds frequently score lower on standardized tests than do children from middle or upper class backgrounds (Fazio, Naremore, & Connell, 1996; Fujiura, 2000; Rhyner, Kelly, Brantley, & Krueger, 1999). In 2001, more than 27% of Latino families with children under the age of 18 lived in poverty, compared with 16% of families in the United States as a whole (U.S. Census Bureau, 2002). Culturally and linguistically diverse Latino children are therefore often at risk for scoring low on standardized tests by virtue of income alone. Because of these factors, researchers have advocated a variety of alternative approaches.
The issue of least biased assessment in the public schools is complicated by the fact that state or local administrative systems may specify eligibility criteria for speech-language services that involve minimum standard scores, percentiles, or minimum standard deviations below the mean on standardized testing, although the use of standardized tests is not mandatory (Roseberry-McKibbin, 2002). Given the large numbers of English language learners in the public schools in some states, the variety of languages they speak, and the difficulty at times of obtaining interpreters for testing, many speech-language pathologists continue to use, and rely on, standardized tests and their norms in English as an important aspect of their test battery. Perhaps not coincidentally, areas with large numbers of culturally and linguistically diverse students have a disproportionately high number of children enrolled in special education programs (Articles, Aguirre-Munoz, & Abedi, 1998). In fact, children referred for special education assessments may be assessed using the same battery of tests, regardless of their characteristics (Baker, Plasencia- Peinado, & Lezcano-Lytle, 1998). Given the continuing emphasis upon standardized testing, especially in the public schools, investigations of alternative methods for this population are needed.
ALTERNATIVES TO STANDARDIZED TESTING AND NORMS
A number of alternatives have been advocated for standardized testing in English with culturally and linguistically diverse children. They have included renorming a test for the specific population being tested, using dynamic assessment techniques to assess a child's modifiability (Lidz & Pena, 1996; Pena et al., 2001; Pena & Quinn, 1997), using other nonstandardized measures (Brice, 2002; Cheng, 1991; Goldstein, 2000; Hamayan & Damico, 1991; Kayser, 1998; Langdon, with Cheng, 1992), and modifying standardized tests (Goldstein, 2000; Kayser, 1998; Langdon & Saenz, 1996; Mattes & Omark, 1991). Each of these alternatives has advantages and disadvantages.
Renorming
Renorming a standardized test using the target population at issue has been advocated as an alternative to using norms based upon the population of the United States. The issue of establishing appropriate norms is an important one. Genesee and Upshur (1996) stated that an appropriate norming group should include individuals of the same ethnic and cultural backgrounds, linguistic background, age, gender, and educational level and type. The group on which the test is normed should also be relatively homogeneous. Although renorming has not been used frequently, at least one district, Santa Ana Unified School District in Southern California, has renormed the Expressive One-Word Picture Vocabulary Test-Revised (EOWPVT-R; Gardner, 1985) for its students (D. Blevins, personal communication, November 7, 2002) and has extensively used the renormed test to assess children within the district (see Note).
Renorming a test has several advantages. In a large and relatively homogeneous school district like Santa Ana, large numbers of bilingual pupils share the same cultural and linguistic backgrounds and economic status, making the time and effort expended in renorming potentially worthwhile. When the performance of particular students in the district is compared to that of their peers, speech-language clinicians can be relatively confident that students who score in the low range on the test rank substantially below peers of the same background.
Renorming also possesses a number of disadvantages. The process requires substantial coordination and time to assess enough typically developing children to provide meaningful norms. In a small or relatively unpopulated geographical area or highly diverse school district, there may not be enough speech--language pathologists or students sharing the same cultural/linguistic backgrounds to collect meaningful local norms. In addition, applicability of the norms is limited to the population upon which it has been renormed (D. Blevins, personal communication, November 7, 2002).
Tests that are both translated and renormed also should have their reliability and validity established for the intended uses (Baker et al., 1998). The difficulty in finding exact equivalencies for English words and concepts may change the absolute and relative difficulty of test items, seriously compromising a test's reliability (Kayser, 1998; Roseberry-McKibbin, 2002). Translated items also may not correspond as clearly with pictured items on a vocabulary test as the original English vocabulary words. In addition, grammatical structures in different languages may not be equivalent or may be acquired at different ages. Finally, differences in life experiences may not easily translate from one culture …

Language Tests in Secondary and Further Vocational Education Are Problematic, Dutch Researcher Says
ScienceDaily (Nov. 15, 2010) — Teachers regularly test the wrong things or they test students for the wrong purpose. We can handle this differently and better, states Dutch researcher Uriël Schuurs. Teachers need tests to evaluate what a student has already learned and what he still needs to learn. A student also takes tests to make progress in his own learning process. However, research by the Expertise Centre for the Dutch Language (Expertisecentrum Nederlands) has shown that there is a gap between what teachers want to test and what standard language tests have to offer.
________________________________________
Measuring students' learning performances generally serves two goals: teachers test their students as a form of feedback and as a part of the learning process (assessment for learning) or teachers test students to examine them and to evaluate the progress of the learning process (assessment of learning). Schuurs studied recent academic literature and asked teachers about language tests in secondary and further vocational education to gain insight into the current practice of testing.
Hardly any standardisation
In theory, there is a clear difference between assessment for learning and assessment of learning. Depending on the purpose of the test, different requirements can be defined. In practice it seems, however, that teachers hardly make a distinction between the different test types. Especially in further vocational education, various test purposes are thrown together in a heap and hardly any use is made of standardised tests. The tests that are being used are often developed by teachers themselves and generally do not comply with reasonable reliability requirements. As a result, the value of the test remains unclear, which hinders students, teachers and the vocational sector. In secondary vocational education, little innovative language testing takes place.
Student's learning process needs better support
Schuurs notices that many teachers expect more from a test than the purpose for which it was developed. Teachers want one test to serve multiple purposes. For instance, they want to be able to judge the performance of a student and at the same time they want to identify the student's weaknesses and the pathway the student should follow. But one test cannot bring all these things to light, says Schuurs. This approach to testing means that opportunities are overlooked. A student's learning process can be supported better if teachers start to examine better.
Reintroduce central exams in further vocational education
On the basis of his research, Schuurs concludes that for further vocational education, the best approach would be to (re)introduce central exams based on national standards. In further vocational education, teachers, in particular, have high expectations that cannot be met with the tests. The converse is true for secondary vocational education, where tests could be used more as part of the learning process. This could, for example, be realised by having students create portfolios, regularly giving them feedback and making them evaluate each other's work (peer evaluation).
Transition from secondary to further vocational education needs improvement
Education policy is aimed at facilitating so-called continuous learning pathways, which enable students to effortlessly move from secondary to further vocational education. Schuurs contends that the difference in testing methods between secondary and further vocational education hinders a smooth transition between the two. In his research, Schuurs provides teachers with tips on how to choose the right testing methods.
The study Measuring learning performances in (v)mbo: assessment for learning and assessment of learning was commissioned by the Dutch Programme Council for Educational Research, a research unit of NWO, and was carried out by Dr U. Schuurs and Prof L. Verhoeven of the Expertise Centre for the Dutch Language (Expertisecentrum Nederlands) in Nijmegen.