Lynch School Researchers See Uneven Impact from Tests

Lynch School Researchers See Uneven Impact from Tests

Studies gauge effect of MCAS, other programs on students, teachers

MCAS and other state testing programs have an uneven impact on students and are unlikely to foster positive overall changes in learning and motivation, according to one of two recent testing-related studies by the National Board on Educational Testing and Public Policy at Boston College.

The other newly released NBETPP study indicates that state testing programs lead teachers to change what they teach and how they teach - sometimes in ways that go against their professional judgment.

The two reports were issued earlier this month by the board. Based in the Lynch School of Education's Center for the Study of Testing, Evaluation and Educational and Policy, NBETPP provides information on the uses and outcomes of educational testing for decision making purposes, paying special attention to groups identified as historically underserved by the educational system.

In a comparative study, the first of its kind, 360 teachers and administrators from school districts in Kansas, Michigan and Massachusetts were interviewed. The goal of the study was to identify the effects of state-level standards-based reform on teaching and learning, paying particular attention to the state test and associated stakes. In a state considered to have low stakes, such as Kansas, there may be no apparent rewards or sanctions related to test scores, whereas in high-stakes states like Massachusetts, for example, student graduation and school accreditation may be tied to the results.

The NBETPP findings show that the impact of high-stakes testing on students is uneven, with high-achieving and suburban students most likely to be motivated and low-achieving and at-risk students most likely to be demoralized. The report also reveals a consistently greater impact on students and educators at the elementary level, regardless of the stakes attached to the test results. Some of these effects were positive, say the researchers, but others produced a classroom environment that was test-driven and unresponsive to students' needs.

Marguerite Clarke, a senior member of the Boston College research team, said the results are "of particular concern in the current policy climate since the accountability requirements of the federal 'No Child Left Behind' Act are placing an even greater testing burden on the early and middle grades. In order to increase the chances of deep, rather than superficial changes in student knowledge, testing should be in the service of learning, not in control of it.

"The fact that some of the biggest differences in impact are not between states, but within states highlights the complexities involved in implementing a one-size-fits-all reform in different contexts and with different populations. The findings also illustrate that increasing the stakes attached to the test results does not necessarily bring about improvement in teaching and learning, but can adversely affect the quality of classroom practice and have a negative impact on at-risk student populations."

Based on the results of the study, NBETPP recommended that test results should not be used to compare teachers and schools unless student demographics and school resources are equated, and the latter are adequate to produce high student performance. The authors also recommended states be flexible in the options available to students for demonstrating achievement so that all have a chance to be successful.

The other study, the broadest ever to be conducted on the topic, was based on a nationally representative sample of teachers. The survey sample was designed to reflect the views of teachers in states in which low, moderate or high stakes are attached to test results.

A majority of the teachers were positive in their opinions of their state's curricular standards, and the vast majority indicated that their district's curriculum was aligned with the state test. However, a substantial majority of teachers at each grade level, but particularly elementary teachers, indicated that state testing programs have led them to teach in ways that contradict their ideas of sound instructional practices.

Across all stakes and grade levels, about four in 10 respondents indicated that teachers in their school could raise test scores without really improving learning. Roughly three-quarters of all teachers, regardless of stakes or grade level, found that the benefits of the testing program were not worth the time and money involved.

"These findings call into question what state test results are really telling us," said Lisa Abrams, a member of the research team. "It is undeniable that these tests are having a profound impact on what happens in the classroom, and in many cases are affecting instruction in ways that contradict the intent of state education reform policies."


Return to April 24 menu

Return to Chronicle home page