澳门跑狗论坛

States that are designing science tests to satisfy the No Child Left Behind Act should set up broad 鈥渟ystems鈥� of assessment, in which questions are aligned with standards and curriculum and new teachers are thoroughly trained in using the exams to improve student learning.

Those are some of the central recommendations of a report by the National Research Council that offers guidance to state officials on how to craft the science tests that the sweeping federal law will require in the 2007-08 school year.

Those tests are likely to have what the report calls a 鈥減owerful influence鈥� over how the subject is taught in the coming years. Establishing entire systems of assessment鈥攁s opposed simply to drawing up tests to comply with the law鈥攊s crucial for states, in order to promote a broad, conceptual understanding of science and its processes, according to the study, released last month.

Further information on the report, is available from .

A committee of testing experts and university researchers convened by the research council produced the report, 鈥淪ystems for State Science Assessment.鈥� The NRC, located in Washington, is a division of the congressional chartered National Academies, which provides research for the government, scientists, engineers, and the public.

The authors say the tests should be built around 鈥渙rganizing principles鈥� or 鈥渂ig ideas鈥� of science, such as evolution and molecular theory, to give students a stronger sense of how different aspects of the discipline connect.

Too often, science tests offer facts in isolation, focusing 鈥渙n the item,鈥� said Mark R. Wilson, one of the report鈥檚 editors.

鈥淵ou really have to step back and focus on the information,鈥� he said. 鈥淲hat purpose are you trying to serve by presenting the item on the test?鈥�

Multiple Assessments

The report notes that states have several options for meeting the No Child Left Behind Act鈥檚 requirement for science assessments at least once in each grade span of 3-5, 6-9, 10-12, and measuring the progress of individual subgroups of students. States should consider using more than one test, with pieces of assessments written at the state and district levels, the report says.

Science tests should include both multiple-choice questions and those asking for written responses, which the authors say are better able to gauge understanding of overall scientific concepts and processes.

鈥淎ny one test, task, or assessment situation is an imperfect measure of what students understand and can do,鈥� the report says. 鈥淚ncluding different types of measures in the system can 鈥� paint a richer picture of student achievement and can tap into the complexity of each science standard more fully.鈥�

The difficulty of designing tests around those district and state partnerships makes it unlikely that states would try it, said Greg Hall, a committee member and Washington state鈥檚 assistant superintendent for assessment and research. Still, he said, many state officials may try to be creative in generating their assessments, in part because under the NCLB law, states do not have to use science-test results as a factor in determining whether schools face such penalties as giving students the option to transfer to other schools, as they do in reading and mathematics.

鈥淵ou might be able to take some more risks in what you do,鈥� Mr. Hall said. 鈥淵ou may see more innovative approaches.鈥�

As of the 2004-05 academic year, 22 states had science tests in elementary, middle, and high school that were aligned to their state standards, according to an analysis by 澳门跑狗论坛鈥檚 most recent Quality Counts report. (States are required to have content standards for science in place by the coming school year)

Better-Trained Teachers

Yet if state officials hope to use science tests to foster achievement in the subject, the research council鈥檚 report says, they will have to improve their performance in a critical area: teacher training. Not only do teacher colleges need to produce instructors with solid grounding in the subject matter鈥攁 long-standing concern of state and school officials鈥攂ut they also need to provide teachers with more-specific training in how to use tests to improve their instruction, according to the NRC.

To that end, schools of education should include more courses on testing, and science testing in particular, in their curricula, the report urges. And states, which are responsible for setting licensing requirements for teachers, should make competence in student assessment mandatory, it says.

The goal is to have teachers who can use science tests to gauge 鈥渉ow well students are learning,鈥� Mr. Wilson said, as well as to grade students鈥� work.

Arthur E. Wise, the president of the National Council for Accreditation of Teacher Education, commended the research council for pressing for improving the training of teachers in using science tests. But he noted that his Washington-based organization, which currently accredits more than 600 teacher colleges nationwide, has implemented new, stronger requirements for institutions to provide aspiring teachers with more expertise on how to administer and evaluate tests.

鈥淲e鈥檙e already there,鈥� Mr. Wise said. He acknowledged, however, that NCATE鈥檚 effort is 鈥渁 major new approach,鈥� and that teacher colleges 鈥渁re still figuring out new and more effective ways鈥� to train teachers in testing.

The report also recommends that the federal government take a more active role in helping states devise strong standards and tests. Federal agencies can award grants for research aimed at producing models for states to follow, the authors say. They also suggest that the U.S. Department of Education require every state to have an independent body to evaluate the content of its science standards. Such evaluations should be made public, they add, and be included in the department鈥檚 review of NCLB compliance.

Federal Checkup

Lawrence S. Lerner, a retired professor at California State University-Long Beach, who has studied state science standards for years, found the idea of a stronger federal role in reviewing them appealing.

Since the passage of the 3陆-year-old No Child Left Behind law, states have been drafting new science standards 鈥渧ery vigorously,鈥� and they are longer and more detailed, he said. Yet the quality of the documents that Mr. Lerner said he had seen varied from 鈥渕odel standards, to really awful, to everything in between.鈥� While he had not reviewed states鈥� science tests in as much depth, Mr. Lerner said he suspected he would find similar disparities in those exams.

Generally, states鈥� standards in science were strongest at the elementary level and gradually became weaker at the middle and high school levels, a weakness he attributed partly to lack of subject-matter expertise among some of the standards-setters. Demanding a more thorough review of standards and tests would benefit students, said Mr. Lerner, who did not serve on the NRC committee.

鈥淭he question is, to what extent would the states stand for that much federal involvement?鈥� he said.