Tests and Teaching Quality: Interim Report (2000)

Improving the quality of teaching in elementary and secondary schools is now high on the nation's educational policy agenda. Policy makers at the state and federal levels have focused on initiatives designed to improve the abilities of teachers already in schools and increase the numbers of well-qualified teachers available to fill current and future vacancies.

As part of their efforts, many policy makers have prescribed tests as a measure of the quality of teachers and teaching, and there is strong interest in requiring teachers to pass a test to earn a license. Moreover, Congress and the President have raised the profile of licensure tests by requiring states and institutions of higher education to report passing rates on such tests.

In response to the increased interest in tests for teachers, the U.S. Department of Education asked the National Research Council (NRC) to study the issue. The NRC convened the Committee on Assessment and Teacher Quality, which has begun a 20-month investigation of the technical, educational, and legal issues surrounding the use of tests for licensing teachers. This interim report covers the first nine months of the committee's study and focuses on existing tests and their use.

The purpose of licensure is to protect the public from harm by setting minimal qualifications for beginning practitioners. To license teachers, states seek a variety of evidence that candidates possess such qualifications, including coursework in state-approved teacher education programs at the undergraduate or graduate level, a major or minor in the intended teaching field, and student-teaching experience. Forty-one states require prospective teachers to pass one or more tests.

States vary considerably in what they test, how they test it, and the level of performance they require for passing. Some states use examinations that assess basic skills, while others assess subject-area knowledge, pedagogical knowledge and skills, or a combination of these types of measures. Test format also varies. Some exams rely on multiple-choice items; others use open-ended questions or a combination of these or other formats (such as portfolios or more performance-based measures). Even when states use the same test, they set different scores for passing.

As a result of these differences in testing practices and standards, it is not feasible to make meaningful comparisons in passing rates among states. At the institutional level, there are variations in policies regarding who is admitted to and graduated from teacher education programs and when these programs require or allow candidates to take the tests. Thus comparisons of passing rates among institutions are difficult to interpret as well.

While licensure policies and testing requirements are intended to set minimal qualifications for public school teachers, two-thirds of the states allow waivers from state licensure requirements for one or more tests. Waivers allow districts to fill vacancies when not enough licensed teachers are available. As a result, classrooms may have teachers who have not satisfied all of their state's testing requirements. In some states, the numbers of teachers with such waivers is substantial.

To construct a licensure test, test developers often begin by collecting data, which ranges from the solicitation of informed judgment to conducting formal surveys, and conducting an analysis to determine the knowledge and skills that a minimally qualified beginning teacher would need. These determinations vary from state to state. Once the test is developed, states set passing scores. Typically, the passing score is based on recommendations from panels of educators who are asked to estimate the level of performance on the test a minimally qualified candidate would be expected to achieve.

Most of the validity evidence currently available for teacher licensure tests is based on judgments about whether the test is likely to assess the knowledge and skills it was intended to measure and whether such knowledge and skills are necessary for beginning teachers to possess. This evidence helps indicate test quality and helps assure policy makers and the public that the test results indicate that teachers are likely to possess the knowledge and skills judged necessary for teaching. However, some tests have been criticized for failing to adhere to professional guidelines for development and validation.

The information licensure tests provide may be deemed necessary, but it is not sufficient to determine whether teachers will be effective in the classroom. Currently, there is little research on the relationship between teachers' test scores and their teaching performance. Such research is difficult to conduct, but it is important. Such research would provide a better understanding of what teacher licensure tests measure.

Even under the best of circumstances, teacher licensure tests, like tests for other professions, cannot be expected to measure everything that is important for practice. Good teachers can explain ideas in ways that different students understand; they are compassionate, resourceful, committed, honest, and persistent in their efforts to help children learn. All of these things are important to teaching, but difficult to measure. A single test or set of tests can only measure some of the characteristics associated with competent teaching.

Blacks and Hispanics generally earn lower scores than whites on licensure tests for teachers. Consequently, blacks and Hispanics tend to have lower passing rates. The same is true for licensing tests in other professions. The disparities in passing scores have contributed to problems faced by schools that want to hire a diverse teaching force and have led to legal challenges. Evidence is needed to determine whether the disparities in average scores and passing rates among groups on these tests are due to actual differences in mastery of the knowledge and skills the tests were designed to assess rather than something else.

As a result of its preliminary explorations and discussions, the committee has reached five conclusions:

• Licensure tests are designed to provide useful information about the extent to which prospective teachers possess the literacy and mathematics skills and/or the subject-matter and pedagogical knowledge that states consider necessary for beginning teaching.

• Teacher licensure tests assess only some of the characteristics that are deemed to be important for effective practice. They are not designed to predict who will become effective teachers.

• There is currently little evidence available about the extent to which widely used teacher licensure tests distinguish between candidates who are minimally competent to teach and those who are not.

• Comparisons of passing rates among states are not useful for policy purposes because of the diversity of testing and licensure practices.

• Test instruments, pass/fail rules, and other licensing requirements and policies that result in large differences in eventual passing rates among racial/ ethnic groups pose problems for schools that seek to have a diverse teaching force.