Weak Tests Drive Math, Science Curricula

K-12 Testing

A report from the National Institute for Science Education (NISE) and another on the Third International Math and Science Study (TIMSS) reinforce the conclusion that commonly used tests undermine, rather than support, high quality student learning in math and science.


Science and math teaching in the U.S. continues to focus too heavily on memorizing facts and practicing computation, according to a June 1997 NISE report. These practices run counter to the content standards developed by the math and science teaching profession.


Preparation for multiple-choice standardized tests appears to be a major cause of the problem. In the January 1998 NISE Brief, James Ridgeway, Director of the Mathematics Assessment Resources Service at Michigan State, points out that "approximately 40 percent of mathematics classes and 20 percent of science classes gave heavy emphasis to preparing students for standardized tests, which have been shown to focus on lower level knowledge and skills." Additionally, "high school teachers in classes with higher proportions of minority students were more likely than others to emphasize preparing students for standardized tests."


The report explains that "since almost all 'high-stakes' assessments. . . assess mathematical technique on a narrow range of tasks, a serious conflict exists between new educational goals and current assessment practices." Those standardized tests pose "a substantial threat to educational reform."


As an alternative, he describes "balanced assessments" that include not only "technical exercises," which can be assessed with multiple-choice items, but far greater use of assessment tasks requiring students to conduct investigations, solve nonroutine problems, plan, and review -- which cannot be assessed with the sorts of tests now used by most states. These recommendations are similar to those found in the Principles and Indicators for Student Assessment Systems of the National Forum on Assessment (order form, p. 15).


TIMSS Results

Ridgeway's recommendation appears to be supported by analyses of U.S. students' performance on the Third International Mathematics and Science Study (TIMSS). While U.S. grade 4 students performed well and grade 8 students were in the middle, students in their final year of high school scored sig-nificantly lower than students in most European nations in both math and science. (Most Asian nations, which scored high at grade 8, did not participate in the high school study.)


While previous international studies were questioned for having a sampling bias that excluded many teenage students in other nations while including more U.S. students, TIMSS may have largely overcome such bias. However, many nations still did not test fully adequate samples of their populations.


Also, in many European nations the final year of high school for college-bound students is equivalent to the first or even second year of college in the U.S. Nations with high scores tended to have older students in the final year of high school, and those students were more likely to study math at the end of high school than were comparable U.S. students. Additionally, they were more likely to have studied calculus, which was included on the exams. Yet these factors only explain part of the relatively low U.S. performance.


Weak Curriculum

One identified cause of steadily declining U.S. perform-ance in math from grade 4 to grade 12 is a curriculum which emphasizes breadth over depth and does not emphasize conceptual understanding or application of knowledge. Tests in the U.S. reinforce and at times even control this comparatively low-level curriculum.


Improving Mathematics in Middle School: Lessons from TIMSS and Related Research, written by U. of Pittsburgh cognitive psychologist Edward Silver, criticizes typical U.S. math curricula and instruction as being "unfocused," "excessively repetitive," and lacking intellectual challenge. Excessive fragmentation of knowledge is a major problem. (A presumption of standardized multiple-choice tests is that knowledge can be "decomposed" into small, measurable fragments.) The report also criticized tracking, low expectations, especially for low-income and minority-group students, and the lack of professional development opportunities for teachers.


A study of classroom practices, by J. Stigler and J. Hiebert in Phi Delta Kappan, analyzed videotapes of schools in Germany, Japan and the U.S. It showed U.S. teachers were less likely to encourage students to actively engage in developing concepts and spent far and away the most classroom time on practicing routine problems.


Similar studies have not yet been released for science -- but they are likely to suggest similar problems. If anything, most standardized science tests are even worse (see article, p. 8). Most commercial and state tests neither encourage nor measure student ability to think and use knowledge in math or science.


Adding reinforcement to this picture, Education Week (February 18, 1998) reported on another study, Facing the Consequences, which analyzed grade 4 and 8 TIMSS results. That report concluded that low U.S. scores on TIMSS stem not from a failure to teach "basics" but rather "a focus on traditional 'basics'." Traditional basics were defined as focusing on computation, while "basics" include conceptual understanding and application.


With traditional tests increasingly controlling the curriculum, high quality student learning is not likely to increase. Unfortunately, as has often been the case, low scores are likely to lead to an increased emphasis on the very kinds of testing that are part of the problem.


 NISE Brief, U. Wisconsin, 1025 W. Johnson St., Madison, WI 53706; (608) 263-9250.

 Improving Mathematics is available on the Web at www.ed.gov/inits.html

 Stigler & Hiebert, "Understanding and Improving Classroom Mathematics Instruction: An Overview of the TIMSS Video Study," Phi Delta Kappan, September, 1997.

W. Schmidt, et al., Facing the Consequences, Kluwer Academic, 1998.