Editorial - March 2007 - Volume 107 (3)

The Great American Education Pastime: High Stakes Testing

Gerald Kulm, Texas A&M University

This time of year heralds the beginning of high stakes testing in most American schools. Although math has had the center stage, science is beginning to be tested in some states as early as 5th grade. In many schools across the country, teachers have now turned their primary attention and class time to “benchmarking,” practice tests, and test-taking strategies. The teaching and learning of new content has essentially come to an end for the remainder of the school year.

It is hard to predict the extent to which testing is likely to continue its influence in controlling the curriculum and content. It is already very apparent that testing has narrowed the mathematics curriculum, taken time away from instruction, and threatened innovation by creative teachers. A few years ago, the Third Mathematics and Science Study (TIMSS) revealed that, compared with the top-ranked countries, the typical U. S. 8th grade mathematics textbook attempted to “cover” nearly twice as many topics. This attempt by publishers to satisfy the diverse desires of educators and policymakers in 50 states and school boards in about 15,000 school districts, results in a curriculum that some have characterized as “a mile wide and an inch deep.” In the last few years, the pressures to test this curriculum have resulted in a curriculum that is far narrower than a mile, but still an inch deep. Many of the objectives that appear in state standards are deemed as untestable. This really means they are not possible to assess with paper-and-pencil, easily scored tests. In today's schools, untestable means untaught. The tested objectives are primarily low level ideas and skills that can be assessed through quick response paper and pencil items.

There is little question that increased testing takes time away from instruction. The actual days spent on review, practice test, test-taking strategies, and the test itself are only part of the problem. In most cases, the test is given a month or more before the end of school, effectively shortening the number of days in the school year. Students know that content presented during this time will not be tested, so teachers often fill the time with less demanding or supplementary material. Teaching to the test is a by-product of the emphasis on test results. It is a rare and confident teacher who is willing and able to address important mathematics or science topics, basing instructional decisions on how students learn best, and taking the time to develop individual students' understanding and proficiency. Instead, many teachers are forced either by edict from administrators, or fear of poor student performance, to use direct teacher-centered and controlled instruction in order to move through the required objectives. Strategies such as problem-based instruction, inquiry approaches, or in-depth applications of content are out of the question, or severely limited.

Is there any way to slow the ever-increasing demand for accountability and high stakes testing? American education seems to follow pendulum swings of extremes when reforms are introduced. It is difficult to predict precisely, but testing seems to be nearing the extreme, providing hope that within a few years, policymakers and legislators will be open to alternative ideas when it becomes obvious that the emphasis on testing is not producing the desired results, especially for students who are most at risk. When that time comes, what should the mathematics and science education community offer as a better approach? Historically, when these shifts have taken place, the positive aspects are discarded along with the negative. Surely, we understand that accountability is an important component of the system. The idea of alignment of curriculum, instruction, and assessment is a valuable design strategy. The assessment approaches, such as rubrics and diagnostic benchmarking, that have been developed during the past decade are useful tools. But accountability can be accomplished without serious damage to instruction. Appropriate sampling and clear reporting of specific performances can provide sufficient information about the effectiveness of curriculum, teaching, and systemic characteristics. It is unnecessary to test every student every year to determine whether individual schools or teachers are performing well. Alignment and benchmarking are tools that belong in a professional development setting. The considerable resources, both time and money, that are now devoted to testing would be better directed at improving teacher knowledge about mathematics and science content as well as knowledge about how and to what extent their students learn specific concepts and skills.

Research and development efforts in assessment, instead of further refining strategies to do efficient large-scale testing, could be devoted to exploring and developing effective ways for classroom teachers to diagnose and assess, real-time, key student misconceptions and difficulties as the lesson begins, as well as determining to what extent student understand content as they complete an activity, and what areas require further discussion or re-teaching. Assessment items that are linked to key concepts, available at the teacher's fingertips through technology-supported student feedback systems, would enhance instructional effectiveness. In short, the majority of assessment time and effort must be re-focused at the classroom level. Teacher development would be aimed at applying sophisticated and user-friendly tools that provide the data teachers need for instructional decisions. This type of assessment is in contrast to the current approach of using testing mainly to provide data for those outside the classroom for policy decisions. The sooner this shift is made, the more likely that instruction can begin to reflect the research-based knowledge we have gained about teaching and learning mathematics and science.