Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 5
Testing English-Language Learners in U.S. Schools: Report and Workshop Summary 1 Committee Findings One participant in the workshop summed up the most sobering, and paradoxical, set of messages that could be taken from the presentations and discussion to illustrate the difficulties facing policy makers. It is crucial, he noted, to include all students in testing designed to hold teachers and administrators accountable for the education they are providing these students.3 However, testing students whose language skills are likely to significantly affect their test performance will yield inaccurate results, unless it is English-language skills that are being tested. Accommodations can help students demonstrate what they know, but they can easily be misapplied; inappropriately applied accommodations–that is, ones that are not well designed to suit the content or skills being measured – can distort test results. It is not fair or sensible to test students for high-stakes purposes unless they have truly been provided with the classroom supports they need and real opportunities to learn the material in question. On the other hand, in the absence of reasonable substitutes for such tests, excluding English-language learners from them can have negative effects on students' school careers as well because educators may not recognize and attend to the particular needs of students who are not included. 3 The questions raised by excluding students from testing programs are not unique to English-language learners; they also pertain to students with disabilities and others.
OCR for page 6
Testing English-Language Learners in U.S. Schools: Report and Workshop Summary While most participants clearly agreed that the goals can seem to conflict, and that there is no one obvious solution, others pointed out that “the perfect must not be the enemy of the good.” As a number of workshop discussions demonstrated, a substantial amount of sound research and practice have already yielded important insights, not only about pitfalls but about the details of how students learn a second language, and about the factors that can make schooling and testing these youngsters more or less successful. Moreover, as one participant noted, Massachusetts, the first state to mandate that every bilingual student's progress in learning English be assessed every year, did so in 1971. The goal for this mandate was to ensure that the students were being adequately taught, but the state has not found an appropriate testing instrument for this purpose and has delegated this responsibility to the districts. The state's long wait for a perfect instrument, he argued, may not have been necessary since a variety of tools are available to help administrators track students' performance. RESULTS FROM THE WORKSHOP The workshop provided committee members with a clear picture of English-language learners in U.S. schools and some of the factors that affect both their educational needs and decisions about testing them. These are presented in detail in Part Two of this report and summarized here. U.S. schools face a significant challenge in educating a growing population of immigrants and others who are not proficient in English. Spanish-background elementary students concentrated in high-poverty schools comprise the majority of these students. A decades-long set of legal precedents, beginning with the Civil Rights Act of 1964, has led to a clear expectation that English-language learners be provided with the same educational opportunities as other students, and that they be held to the same academic standards as other students. English-language learners' academic needs are complex and variable. They need to develop not only mastery of conversational English, but also mastery of the academic spoken and written English necessary to do the academic work for which they are ready. Accomplishing the latter takes four to seven years, on average. Moreover, while their English skills are developing they also need to continue to make progress in other subjects and to receive appropriate and challenging instruction that prepares them
OCR for page 7
Testing English-Language Learners in U.S. Schools: Report and Workshop Summary to meet required standards. Disentangling students' progress in English from their academic performance in other areas is difficult because oral and written English are tools used in all kinds of academic work, but educators need effective ways of monitoring all aspects of these students' academic progress. Current state and district policies regarding the inclusion of English-language learners in testing, the use of accommodations for these students, and other issues vary significantly across the country. Accommodations can be used to compensate for limitations in a student's English proficiency and thereby make it possible to collect valid information about that student's knowledge or skills. On the other hand, if misused, accommodations can render testing results inaccurate. The appropriateness of accommodations depends both on the particular language status of the student and on the purposes for which he or she is being tested. Two very different kinds of information about English-language learners ' linguistic progress are needed. Educators need to know which students lack proficiency in English and may need academic supports or need to be accommodated in or excluded from testing. They also need ways of monitoring these students' linguistic progress after they are identified as English-language learners. The workshop discussions also yielded a clear sense that more research is needed in a number of areas related to the testing of English-language learners. The most pressing need identified was for more long-term studies that track the progress of students beyond their reclassification as proficient in English and removal from bilingual or English as a Second Language (ESL) programs. Several participants noted that statistical tracking of language-minority students is difficult and has been spotty. Those who drop out of school or move, for example, can disappear from record-keeping systems, and the progress of students who are never enrolled in such programs is similarly unrecorded. The absence of these students can make the statistical picture of the support system inaccurate. Moreover, since students are frequently moved out of language support programs before they are completely proficient in English, particularly academic English, studies of their academic performance, career trajectories, and continuing language development after leaving the program would be very useful. Greater understanding of the effects of different accommodations was also cited as an important need. Committee members and workshop participants clearly articulated the
OCR for page 8
Testing English-Language Learners in U.S. Schools: Report and Workshop Summary sometimes frustrating tensions between the goals of including students in testing for the sake of system accountability and adhering to testing standards to ensure the accuracy of testing results. The seeming paradoxes, however, may be most acute when a test is used for more than one purpose. Including English-language learners in testing is often the only way to capture information about these students' progress that will enable administrators and policy makers to make informed decisions about their education. Because of past policies that excluded many of these students from testing, there is an unacceptable dearth of information about them. On the other hand, inappropriate testing only yields misleading information that can lead to incorrect decisions about individual students and mistaken assumptions about groups of students. A point that was the theme of a 1999 NRC report offers an important corrective to the apparent dilemma posed by the testing of English-language learners. As High Stakes: Testing for Tracking, Promotion, and Graduation (National Research Council, 1999a) made clear, states and school districts can include English-language learners in large-scale assessments and use their scores for system accountability even in cases where it would be inappropriate to use such scores to make promotion or graduation decisions about individual students, as, for example, when students have not had an opportunity to acquire the knowledge and skills the tests measure. The particular concerns about fairness that are paramount in the individual testing context are not necessarily a hindrance to collection of the aggregated data that is needed for system accountability. (Even in the collection of such aggregated data, the inclusion of students with limited English proficiency may yield invalid results, depending on the circumstances. This point is discussed later in the report.) Many of the workshop discussions seemed to lead toward a shared sense that the real goal in testing English-language learners is to strike a workable, real-world balance tailored to particular circumstances—to find ways to have true accountability for the successful education of these students without letting assessments turn into punishments for students or teachers. POINTS TO CONSIDER Perhaps because the variations in student populations and local circumstances were raised so frequently during the workshop, attention centered on questions that could profitably be asked about specific testing programs to help educators achieve the hoped-for balance.
OCR for page 9
Testing English-Language Learners in U.S. Schools: Report and Workshop Summary Is the assessment valid for the purpose for which it is being used? Well designed tests have clearly articulated purposes, and using them in other ways—either selecting students for testing according to incorrect criteria or using the scores to make inferences for which they were not validated—is likely to yield inaccurate results. This point is of particular importance in the context of decisions about testing English-language learners. Validity can be affected in subtle ways by language deficiencies, and identifying the level of proficiency that makes including a student in a particular test fair is not straightforward. What is known about the effects of particular accommodations being considered in the setting in question? The useful question is not “are accommodations fair and effective?” but rather, “if we want to assess the mathematics skills of these students, with these levels of English proficiency, what accommodations might help us do so accurately?” Both research findings and the experiences of others who have faced similar situations can be of use in answering these kinds of questions. What are the possible long-term implications of a particular policy? While it is clear that a variety of real-world factors influence decisions about inclusion and accommodations, thoughtful consideration of possible long-term effects (such as effects on graduation rates, or the long-term impact of spending priorities, for example) can be a useful corrective to short-term political considerations. Imagining, for example, how a policy that is being contemplated might work in a district whose population of English-language learners doubles in the next 10 or 20 years could be a useful way of focusing on its possible unintended consequences. The committee took note of Ursula Casanova's reminder to the group that while issues of inclusion and accommodation are sometimes similar for students with disabilities and English-language learners, it is important to recognize the differences as well. Being literate in two languages, she noted, is an academic advantage, perhaps one less valued in the United States than in other countries, but not a status to be penalized. Unfortunately, however, for many English-language learners in U.S. schools literacy in two languages is a goal, not a reality. Because the population of English-language learners is growing at a time when reliance on testing for many kinds of educational purposes is also growing, the importance of thoughtful planning for testing this population is greater than ever. Many new tests have been developed and used in the past decade or two, but not all of them are accomplishing their stated goals. The goals for testing have grown
OCR for page 10
Testing English-Language Learners in U.S. Schools: Report and Workshop Summary out of ambitious goals for teaching all students to high standards; the risk to be averted is that tests can have the effect of working against those ambitious goals if their misuse leads to poor educational decisions for some students. RECOMMENDATIONS The committee, having deliberated on what it heard at the workshop and on supplementary information, has identified three recommendations for ways of improving on current uses of tests and data related to English-language learners. With regard to assessment, it is clear that educators and policy makers need information about language-minority students' developing skills in English and about their developing academic skills and knowledge in other subject areas. These two needs should be understood as distinct and addressed separately. Research suggests that doing so is a particular challenge because language skills are easily confounded with other constructs in testing (Abedi et al., 1995). Nevertheless, these two needs can be addressed. Recommendation 1: Researchers should continue to target the separate needs for improved ways of assessing the developing language skills of English-language learners and improved ways of assessing their academic progress, regardless of their level of proficiency in English. They should focus on improving or expanding existing assessment tools or creating new tools for both purposes.4 Recommendation 2: Test developers, educators, and policy makers should make better use of existing knowledge of how these areas can be assessed by, for example, taking pains to use assessments that avoid unnecessary linguistic complexity; adapting existing tests of English proficiency to assess incremental progress; using a variety of appropriate means to assess students' academic and linguistic progress; and avoiding inappropriate practices such as arbitrary cut-offs in the length of time that English-language learners can receive testing accommodations or other supports. In part because of past policies of excluding English-language learners from large-scale data collection efforts, information about their educational status has been spotty and less reliable than it should be. The committee 4 The committee notes that some work of this kind has already been undertaken.
OCR for page 11
Testing English-Language Learners in U.S. Schools: Report and Workshop Summary recommends that the collection of data on these students be improved in several ways. Recommendation 3: Policy makers and funding agencies should consider the need for consistent longitudinal data about the progress of English-language learners through school and beyond at both the national and state levels. Researchers, policy makers, and the public need valid information that yields insights into how these students fare whether or not they are enrolled in instructional programs that target their linguistic and cognitive needs as second-language learners. Information is also needed about English-language learners after they leave school and move into the workforce, and after they stop receiving any educational supports. More data are also needed to improve our understanding of variations in the kinds of bilingual services and testing accommodations children receive and their effects; how long they are eligible to receive them; and how their progress is affected by factors such as the age at which they entered the system. It is the committee's hope that policies regarding the testing of English-language learners will grow more consistent and accurate as the ramifications of different approaches come to be better understood. Moreover, if data collection is improved and more is learned about the progress of these students through school, strategies for attending to their developing language skills and to their academic achievement should become more focused and successful than they are now.
Representative terms from entire chapter: