ROANOKE TIMES
Copyright (c) 1997, Roanoke Times
DATE: Wednesday, April 2, 1997 TAG: 9704020027
SECTION: VIRGINIA PAGE: C-1 EDITION: METRO
SOURCE: MATTHEW BOWERS LANDMARK NEWS SERVICE
The stakes are high, so the tests must be fact-based and objective, some educators say. Others counter that multiple-choice tests don't really show what students know.
New tests to be given this month in classrooms across Virginia have triggered an old debate over the best way to gauge how well public schools teach and their students learn.
Students in grades three, five, eight and 11 will take the tests as part of a new initiative to measure achievement and learning.
One test will compare Virginia students with peers across the country; the other will determine whether students are learning what's supposed to be taught in Virginia's beefed-up curriculum.
The state is proposing that beginning with this fall's high school freshmen, the class of 2001, students will have to pass the 11th-grade state-standards test to graduate. Most questions will be multiple choice.
This prompts some educators to renew the long-held complaint that these tests don't really show what students know or how well they think, analyze or solve problems. Instead, some testing experts say, multiple-choice exams measure mostly lower-level intellectual skills, such as memorization.
"For politicos and policy wonks, this is right on the money,'' said Edward Kifer, a University of Kentucky professor who specializes in testing. "I don't think you could confuse what you get from schooling with what you find out on a test.''
Kifer supports alternatives to multiple-choice tests, such as open-ended questions that enable students to show their reasoning, and group activities from which students can draw inferences.
Others - including the state Board of Education, which picked Virginia's tests - counter that the fact-based, multiple-choice test is more objective, less expensive and easier to administer. The results are more defensible than those of open-ended tests, board members say. The tests fit well with Virginia's new tougher, fact-based curriculum standards that have second-graders learning about ancient Egypt and China, among other things.
Test results will be reported starting in spring 1998, and they will be a factor in accrediting public schools.
Proposed standards for accreditation would make the 11th-grade test a hurdle for high-school graduation beginning with this fall's freshman class.
The board in an April 1996 resolution said standardized and machine-scoreable assessments are "a proven means of measuring student achievement in the acquisition of knowledge and skills.'' It also said that nontraditional assessments ``-such as observations, projects, portfolios and demonstrations - carry too many problems: cost, time and unproven methods for providing results that are reliable, valid, generalizable, and equitable.''
``It's a controversial area,'' concluded Tonya Moon, a University of Virginia assistant professor who worked on the committee that sought bids for the state testing contract.
Actually, the new Virginia Assessment System contains two parts.
One test is a standard, nationally used one that will compare Virginia students in grades three, five, eight and 11 with peers across the country by weighing results against a national norm. Students in grades three, five, eight and 11 will take this test, the Stanford Achievement Test Series, Ninth Edition - commonly called the "Stanford 9'' - sometime in April.
Scores will be hard to compare with past years - it's a different test, plus it's based on a new standard.
The second part of the state tests will assess progress by students and schools in incorporating the new Standards of Learning, the statewide curriculum standards passed in 1995. Test writers from Harcourt Brace Educational Measurement of San Antonio are using the state standards to write tests for grades three, five, eight and 11 in the core areas of English, history, math and science. Computer and technology knowledge also will be tested in grades five and eight, and writing short essays will be required of students in grades five, eight and 11.
The first test will be a field test only - a test of the test questions - that won't count for or against students. It will be given during the week of April 28-May 2. Most students in each age group will be tested in only one subject area. No results will be reported, but they'll be used to refine the test and set standards for passing it.
Test results will be reported starting in spring 1998.
Since the results will be a factor in school accreditation and since passing the 11th-grade test will be a high school graduation requirement, some argue the necessity for objective, multiple-choice tests with easy-to-defend results, the University of Virginia's Moon said.
How good a test is depends on what it's intended to do, said Jack Campbell, a professor at Mount St.Mary's College in Emmitsburg, Md. Multiple-choice tests are good for some things, he said.
``Basically, if it comes down to how much content do they know, what basic knowledge do they know, these do it very well,'' Campbell said.
The public schools in his state, however, went in another direction in 1993. They use what's called "authentic assessment,'' or more open-ended testing using writing samples, conducting scientific experiments in groups and answering questions that have more than one right answer, such as using math to design a garden.
The idea is that such skills - cooperation, problem-solving and so on - are what employers say are needed, said Ronald Peiffer, assistant state superintendent for school and community outreach.
``It's not just that they know how to do math computations, but it's important to know they can use that in real-life situations,'' Peiffer said.
Teachers were specially trained to score the complicated tests. Parents complained about low scores, since standards were set high to provide a target for improvement. And the tests are pricey - $24 per child as opposed to $5 for an "off-the-rack'' national-norm-type test. But Peiffer compared the cost to the $6,000 a year it costs to educate each child, and to the helpful information it provides schools on how the students are doing.
"That's not a very large investment to get some quality control there,'' Peiffer said. "No other standardized test we've had before has driven instruction, but this has.''
That's just one failure of basic multiple-choice tests, said the University of Kentucky's Kifer.
"One of the things you can't do is get the kids to give you reasons'' for their answers, Kifer said.
"Because finally it's not what you can regurgitate, but how you think about things that serves you well. I would rank multiple-choice tests last. And I would put writing tests, portfolios and performance tests way ahead of them.'' the kids.''
Echoing the view that multiple-choice tests don't accurately reflect what children know or can do is FairTest, a testing industry watchdog group in Cambridge, Mass. Worried about scores, teachers can feel pressure to favor teaching simply facts over problem-solving and analysis, said Monty Neill, an associate director.
"If your standards only call for memorization and recall, fine,'' Neill said. "If your standards call for you to analyze information, synthesize, evaluate and apply information, these won't do that.'
"I would assume, in the end, that we'd want students to use knowledge in the real world.''
LENGTH: Long : 128 linesby CNB