Testing

State testing programs have a long history. New York State administered its first Regents' Examinations as early as 1865. Several other state programs had their beginnings in the 1920s, when new forms of achievement examinations–objective tests–were developed for and introduced in the schools. In 1937 representatives from fifteen state programs and nonprofit testing agencies met, under the leadership of the American Council on Education's Committee on Measurement and Guidance, to discuss common problems. The group continued to meet annually, except during World War II, for decades.

In 1957 President Dwight D. Eisenhower called attention to state testing programs when he indicated the need for nationwide testing of high school students and a system of incentives for qualified students to pursue scientific or professional careers. The subsequent passage of the National Defense Education Act of 1958 (NDEA) not only encouraged but gave financial support to testing, guidance, and scholarship programs.

The growth in number and importance of state testing programs accelerated rapidly in the 1970s and has continued to grow ever since. The growth in the 1970s reflected, at least in part, the enhanced role of states in education policy. In the academic year 1969–1970, the majority–53 percent–of the funding for schools came from local agencies; states contributed 39 percent and the federal government provided 8 percent. A decade later, however, the state share of education funding had increased to 47 percent, and state governments became the dominant source of funding for schools.

With that increased responsibility came demands for some form of accountability for results, and that meant state tests to determine if students were learning. At the same time, there was growing concern among public officials and the public about the quality of schools, fueled in part by the revelation that average scores on the SAT declined between 1963 and 1977. In response to these concerns, and the interest in accountability, a majority of states in the 1970s implemented some form of minimum competency test, which students were required to pass in order to graduate from high school. The number of states conducting such tests rose from a handful in 1975 to thirty-three in 1985.

The wave of state education reforms enacted following the publication in 1983 of A Nation at Risk, the report by the National Commission on Excellence in Education, further accelerated the growth in state testing. That report, which warned of a "rising tide of mediocrity" in America's schools, recommended that states adopt achievement tests to measure student performance, and many states responded to the call. By the end of the 1980s, forty-seven states were operating at least one testing program, up from thirty-nine in 1984.

This growth continued throughout the 1990s as well. The dominant role of states in education policy was symbolized near the beginning of that decade, when President George H. W. Bush called the nation's governors to an extraordinary "education summit" in Charlottesville, Virginia. In the wake of that meeting, the President and the governors agreed to a set of national education goals, which included the pledge that all students would be "competent in challenging subject matter" by the year 2000. The goals were enshrined into federal law in 1994; that same year, President Bill Clinton signed the Improving America's Schools Act, which required states to set challenging standards for student performance and implement tests that measure student performance against the standards. In response to the law, nearly all states revamped their existing tests or developed new tests, and as of 2001, all states except Iowa (where local school districts administer tests) had a statewide testing program; by one estimate, the amount spent by states on testing doubled, to $410 million, between 1996 and 2001.

The No Child Left Behind Act, which President George W. Bush signed into law in 2002, requires a significant increase in state testing. Under the law, states must administer annual reading and mathematics tests in grades three through eight, tests in science in at least three grade levels, and tests at the high school level. At the time of enactment, only nine states met the law's requirements for annual tests in reading and mathematics aligned to state standards.

Types of Tests

Although most state tests consist primarily of multiple-choice questions, there is considerable variation among the states. Thirty-four states include some short-answer questions in at least some of their tests, requiring students to write answers rather than select from among answers already provided, and eighteen states include questions requiring extended responses in subjects other than English language arts. (Nearly all states administer writing tests that ask for extended responses.) Two states, Kentucky and Vermont, assess student performance in writing through the use of portfolios, which collect students' classroom work during the course of a school year. The portfolios are scored by teachers, using common criteria.

The Maryland state test, the Maryland School Performance Assessment Program (MSPAP), is unusual in that it consists exclusively of open-ended questions. Students work in groups for part of the assessment, and many of the questions are interdisciplinary, requiring students to apply knowledge from English language arts, mathematics, science, and social studies. In addition, the test is designed so that individual students take only a third of the complete assessment; as a result, scores are reported for schools, school districts, and the state, but not for individual students.

The MSPAP, like many state tests, was custommade to match the state's standards. In Maryland's case, the test was developed by state teachers; other states contract with commercial publishers to develop tests to match their standards. Such tests indicate the level of performance students attained, but do not permit comparisons with student performance from other states. The types of reports vary widely. Maryland, for example, specifies three levels of achievement: excellent, indicating outstanding accomplishment; satisfactory, indicating proficiency; and not met, indicating more work is required to attain proficiency. The state's goal is for 70 percent of students to reach the satisfactory level and 25 percent to reach the excellent level.

Other states, meanwhile, use commercially available tests that provide comparative information. The most commonly used tests are the Stanford Achievement Test, 9th Edition (or SAT-9), published by Harcourt Brace Educational Measurement, and the Terra Nova, published by CTB-McGraw-Hill. These tests are administered to representative samples of students, known as a norm group, and provide information on student performance compared with the norm group. For example, results might indicate that a student performed in the sixty-fifth percentile, meaning that the student performed better than sixty-five percent of the norm group. To provide both information on performance against standards and comparative information, some states employ hybrid systems. Maryland, for example, administers a norm-referenced test in grades in which the MSPAP is not used. Delaware, meanwhile, has embedded an abbreviated version of the SAT-9 within its state test.

Purposes

State tests are used for a variety of purposes. The most common is to provide information to parents and the public about student, school, and school system performance. The expectation is that this information can improve instruction and learning by pointing out areas of weakness that need additional attention.

In addition to providing reports to parents about their children's performance, forty-three states issue "report cards" on schools that indicate school performance; twenty states require that these report cards be sent to parents. The No Child Left Behind Act requires all states to produce school report cards and to disseminate them to parents.

States also place consequences for students on the results of tests. As of 2001 four states (Delaware, Louisiana, New Mexico, and North Carolina) make promotion for at least certain grades contingent on passing state tests, and another four states are planning to do so by 2004. In other states, school districts set policies for grade-to-grade promotion, and many districts use state tests as criteria for determining promotion from grade to grade. A 1997 survey of large school districts, conducted by the American Federation of Teachers, found that nearly 40 percent of the districts surveyed used standardized tests in making promotion decisions at the elementary school level, and 35 percent used tests in making such decisions at the middle school level. Although the survey did not indicate which tests the districts used, the report noted that statewide tests were among them.

More commonly states use tests as criteria for high school graduation. Seventeen states, as of 2001, make graduation from high school contingent on passing state tests, and another seven are expected to do so by 2008. These numbers are similar to those recorded in the early 1980s at the height of the minimum-competency era. Yet the graduation requirements first implemented in the late 1990s are different than those of the earlier period because the tests are different. Unlike the previous generation of tests, which measured basic reading and mathematical competencies, many of the newer tests tend to measure more complex skills along with, in many cases, knowledge and skills in science, social studies, and other subjects.

In most states with graduation test requirements, the tests are administered in the tenth or eleventh grade, and students typically have multiple opportunities to take the tests before graduation. In some states, such as New York, Tennessee, and Virginia, the graduation tests are end-of-course tests, meaning they measure a particular course content (such as algebra or biology) and are administered at the completion of the course.

States also use tests to reward high-performing students. For most of its existence, the New York State Regents' Examination was an optional test; students who took the test and passed earned a special diploma, called a Regents' Diploma. Beginning in 2000, however, the state required all students to take the examinations. Other states, such as Connecticut, continue to use tests to award special diplomas to students who pass them.

Some states, such as Michigan, provide scholarships for students who perform well on state tests. There, the state awards $2,500 scholarships to students attending Michigan colleges and universities who score in the top level of all four high school tests–mathematics, reading, science, and writing (the scholarships are worth $1,000 for students attending out-of-state institutions).

In addition to the consequences for students, states also use statewide tests to determine consequences for schools. As of 2001 thirty states rate schools based on performance, and half of those states use test scores as the sole measure of performance (the others use indicators such as graduation and attendance rates, in addition to test scores). In Texas, for example, the state rates each school and school district in one of four categories, based on test performance: exemplary, recognized, acceptable/academically acceptable, and low-performing/academically unacceptable. To earn a rating of exemplary, at least 90 percent of students–and 90 percent of each group of students (white, African American, Hispanic, and economically disadvantaged)–must pass the state tests in each subject area. Those with at least an 80 percent pass rate, overall and for each group, are rated recognized; those with a 55 percent pass rate are rated acceptable.

In eighteen states high-performing schools can earn rewards. In some cases, such schools earn recognition from the state. In North Carolina, for example, the twenty-five elementary and middle schools and ten high schools that register the highest level of growth in performance on state tests, along with those in which 90 percent of students perform at or above grade level, receive a banner to hang in the school and an invitation to a state banquet in their honor. Schools and teachers in high-performing or rapidly improving schools also receive cash awards in some states. In California, for example, 1,000 teachers in the schools with the largest gains on state tests receive a cash bonus of $25,000 each. Another 3,750 teachers receive $10,000 each, and another 7,500 receive $5,000 each. Schools that demonstrate large gains receive awards of $150 per pupil.

States also use test results to intervene in low-performing schools. In 2001 twenty-eight states provide assistance to low-performing schools. In most cases, such assistance includes technical assistance in developing improvement plans and financial assistance or priority for state aid. In some cases, such as in North Carolina and Kentucky, the state sends teams of expert educators to work intensively with schools that perform poorly on state tests. These experts help secure additional support and can recommend changes, such as replacing faculty members.

Twenty states also have the authority to levy sanctions on persistently low-performing schools. Such sanctions include withholding funds, allowing students to transfer to other public schools, "reconstitution," which is the replacing of the faculty and administration, and closure. However, despite this authority, few states have actually imposed sanctions. One that did so was Maryland, where in 2000 the state turned over the management of three elementary schools in Baltimore to a private firm.

The No Child Left Behind Act of 2001 contains a number of provisions to strengthen the role of state tests in placing consequences on schools. Under the law, states are required to set a target for proficiency that all students are expected to reach within twelve years, and to set milestones along the way for schools to reach each year. Schools that fail to make adequate progress would be subject to sanctions, such as allowing students to transfer to other public schools, allowing parents to use funds for supplemental tutoring services, or reconstitution.

Effects of Statewide Testing

The rapid growth in the amount and importance of statewide testing since the 1970s has sparked intense scrutiny about the effects of the tests on students and on classroom practices.

Much of the scrutiny has focused on the effects of tests on students from minority groups who tend to do less well on tests than white students. In several cases, advocates for minority students have challenged tests in court, charging that the testing programs were discriminatory. In one well-known case, African-American high school students in Florida in the 1970s challenged that state's high school graduation test–on which the failure rate for African Americans was ten times the rate for white students–on the grounds that the African Americans had attended segregated schools for years and that denying them diplomas for failing a test preserved the effects of segregation. In its ruling, the U.S. Court of Appeals for the Fifth Circuit upheld the use of the test, but said the state could not withhold diplomas from African-American students until students had no longer attended segregated schools (Debra P. v. Turlington).

The court also held that students have a property right to a diploma, but that Florida could use a test to award to diplomas provided that students have adequate notice of the graduation requirement (four years, in that case) and that the test represents a fair measure of what is taught. Although the decision applied only to the states in the Fifth Circuit, the standards the court applied have been cited by other courts and other states since then.

In addition to the legal challenges, state tests have also come under scrutiny for their effects on student behavior–specifically, on the likelihood that students at risk of failing will drop out of school. There is some evidence that dropout rates are higher in states with graduation-test requirements. But it is unclear whether the tests caused the students to decide to drop out of school.

Many testing professionals have also expressed concern that the use of tests to make decisions like promotion or graduation may be inappropriate. Because test scores are not precise measures of a student's knowledge and skills, test professionals warn that any important decision about a student should not rest on a single test score; other relevant information about a student's abilities should be taken into account.

There has been a great deal of research on the effects of state tests on instruction. Since a primary purpose of the state tests is to provide information to improve instruction and learning, this research has been closely watched. The studies have generally found that tests exert a strong influence on classroom practice, but that this influence was not always salutary. On the positive side, the studies found that tests, particularly those with consequences attached to the results, focused the attention of students and educators on academic performance and created incentives for students and teachers to raise test scores. In addition, tests also encouraged teachers to focus on aspects of the curriculum that may have been underrepresented. For example, in states like California that introduced writing tests that assessed students' written prose (as opposed to multiple-choice tests that measured writing abilities indirectly), teachers tended to spend more time asking students to write in class and exposing them to a broader range of writing genres.

On the negative side, studies also found that state tests often encouraged teachers to focus on the material on the test at the expense of other content that may be worthwhile. In states that used exclusively multiple-choice basic skills tests, researchers found that many teachers–particularly those who taught disadvantaged students–spent a great deal of class time on drill and practice of low-level skills, as opposed to instruction on more complex abilities that the tests did not assess. At the same time, researchers found that in some cases teachers devoted a greater proportion of time to tested subjects and less to subjects not tested, like history and the arts, and that teachers spent class time on test-preparation strategies rather than instruction in academic content. The heavy influence of tests on instruction has led some commentators to question whether gains in test scores represent genuine improvements in learning or simply "teaching to the test."