States Clear Initial Hurdle on ELL Tests

Includes updates and/or revisions.

All states and the District of Columbia have now ushered in new English-language-proficiency tests to comply with No Child Left Behind Act requirements for those still learning the language. Still, it’s too early to tell if a passing score predicts that a student will do well in a regular classroom or on other mandatory state tests, according to a nationwide examination of such assessments released today.

“We see systematic improvement” in the new generation of English-proficiency tests, Jamal Abedi, a professor of education at the University of California, Davis, and the editor of the report, said in a phone interview. “But we don’t know how they actually translate into performance of English-language learners.”

The report, “English Language Proficiency Assessment in the Nation: Current Status and Future Practice,” released by Mr. Abedi’s university, is a collaboration by 32 experts on testing and ELLs to provide an overview of the new era of testing for English proficiency.

Under the nearly 6-year-old NCLB law, states were, for the first time, required to gauge the progress of second-language learners in grades K-12 every year in learning English. The federal education law requires states to assess those students in reading, writing, speaking, and listening—while the previous generation of tests were mostly designed to assess only speaking and listening.

The tests are better than the ones that were commonly used prior to passage of the NCLB law because they are aligned with state standards for English-proficiency—which many states had to craft to comply with the act—and state content standards, such as for mathematics, Mr. Abedi explained.

In addition, he said, the new generation of tests are designed to assess “academic English”—the version of the language children need in order to learn subjects in school—rather than the social English used on the playground.

Prodded by NCLB

Experts say the federal education law’s requirements to test English progress in a comprehensive way is a positive move, even though much more work needs to be done to ensure that the tests are valid and meaningful.

“These tests wouldn’t be as good as they are without those [NCLB] requirements,” said Stanley N. Rabinowitz, the director of assessment and standards-development services for WestEd, a San Francisco-based research agency. He is conducting a separate examination of a number of the new English-proficiency tests.

Mr. Rabinowitz agreed with Mr. Abedi that the states have done a good job in coming up with tests that evaluate academic English. At the same time, he said, one limitation of the UC-Davis report is that it is written primarily by the people who devised the new tests. “I won’t call it public relations, but it is the best foot forward,” he said. “Not all tests have been received and are working as well as this report indicates.”

In the case of Florida, noted Candace A. Harper, an associate professor of education at the University of Florida, Gainesville, the report’s description of the state’s English-proficiency test is much more “glowing” than what teachers had to say about the test in her state.

For example, she and several graduate students interviewed 12 teachers and 16 students about Florida’s test, the Comprehensive English Language Learning Assessment. They found that the listening section was broadcast over the public address system in one school’s cafeteria, while at another school, an English-as-a-second-language teacher administered that section to students in small groups. Such inconsistencies in implementation raise questions about the validity of the scores, Ms. Harper said.

Other experts also said it’s important to learn more about whether the English-proficiency tests are working as the developers intended and how useful they are to educators.

“What do you do with the scores?” said Diane Staehr Fenner, a research scientist at the Center for Equity and Excellence in Education at George Washington University in the nation’s capital.

“You see all these numbers. Is the teacher trained on how to read these scores? How do you translate the scores into meaningful classroom instruction?” said Ms. Staehr Fenner, who previously was an assessment specialist for ELLs in the Fairfax County, Va., schools.

Federal Efforts

The UC-Davis report is being released as the U.S. Department of Education is trying to respond to state officials’ pleas for help in writing and using English-proficiency tests and for setting targets under the federal education law, called “annual measurable achievement objectives,” for students to progress in English and attain proficiency in the language.

Separately, state officials and federal education officials have, at times, been at odds over how best to include English-learners in the regular mathematics and reading assessments that all students must take. (“Tussle Over English-Language Learners,” Jan. 31, 2007.)

Meanwhile, states—with federal assistance—have quietly been taking steps to put English-proficiency tests in place. To date, the Education Department has provided a total of $10 million to four consortia of states to craft new tests, which about half are using, and has followed up with some additional grants. The rest of the states are using tests designed specifically for their states or by commercial developers.

The Education Department released in October a draft of a framework for creating English-language-proficiency standards and tests. Federal officials soon will invite states to participate in a pilot project to use the framework in evaluating their English proficiency standards and tests, according to Kathryn M. Doherty, a special assistant to the department’s deputy secretary.

In addition, Ms. Doherty said, federal officials are writing official guidance for states on how to set targets for the achievement objectives, which federal officials expect to release in the spring.

At the state level, educators involved in the implementation of English-proficiency tests said the UC-Davis report helps to put into perspective what is happening with such tests in their own states.

“Having a report of this type earlier could have conserved resources so that states would not have had to independently collect current information to make decisions about which English-language-proficiency assessment they would select,” said Teddi Predaris, the director of the office of services for English-language learners in the 164,000-student Fairfax County, Va., school system.

Ms. Predaris served on a committee to select a new English-proficiency test for Virginia to replace the one that most districts in the state are now using, the Stanford English Language Proficiency Test, to comply with the NCLB law. After comparing four tests, the committee recommended switching to a test known as Assessing Comprehension and Communication State to State for English Language Learners, or ACCESS for ELLs, starting next school year.

The Virginia board of education approved that selection in September, making Virginia the 16th state to adopt the test produced by the World-Class Instructional Design and Assessment Consortium. It is the English-proficiency test most widely used by states.

Blog: Learning the Language

Veteran Education Week reporter Mary Ann Zehr tackles difficult policy questions, explores learning innovations, and shares stories about different cultural groups on her beat.

Read the blog.

“The ACCESS for ELLs test is the best thing out there right now because it is carefully aligned with the content-area assessments and is an accurate, valid, and reliable assessment,” Ms. Predaris said.

In Florida, Ms. Harper said she was concerned by the UC-Davis report’s finding that Florida is behind many other states in putting an English-proficiency test in place. While Florida just carried out its English-proficiency test for the first time last school year, California, for instance, administered such a test before passage of the NCLB law.

Ms. Harper, who serves on an advisory committee to the Florida education department for implementation of Florida’s test, believes it is premature to require that the test be used by schools to make decisions such as when to release children from English-acquisition programs.

“How do these scores reflect what students can do in content-area classrooms?” she said, “We don’t know that. We have very shaky data. Even the assessment people [on her committee] were saying, ‘This is like building on mud.’ ”

Ms. Harper’s concerns are warranted, said Mr. Abedi, the report’s editor. A state should have at least two years’ worth of data based on full implementation of its test before using it for any high-stakes decisions, he added. Now, Mr. Abedi said, most of the data available is from field-testing or pilot-testing, not based on full implementation across a state over a significant amount of time.