The report states:
"Evidence continues to mount demonstrating that high-stakes testing undermines, rather than enhances, efforts to improve education for all children.
The picture that emerges from several studies is of a nation severely hurting its educational system while failing to provide help to schools that need it, thereby harming the nation’s children – all in the name of “accountability.” High-stakes testing puts narrow, flawed instruments at the center of education and leads to intensive teaching to the exams, which does not result in real learning gains. At the same time, many children are less motivated, are denied a high-quality educational experience, and become more likely to leave school before graduating. While state-mandated exams have been the major culprit, the federal government’s imposition of high stakes on schools and districts will compound the problem (see stories on NCLB, pp 23- 28).
Amrein and Berliner
Last year, Professors Audrey Amrein and David Berliner of Arizona State University reported that gains states report on their own high-stakes tests do not correlate with results from other exams, such as the National Assessment of Educational Progress or SAT and ACT college entrance tests (see Examiner, Spring 2002). In states with graduation tests, scores on these other exams often declined or grew less quickly relative to the nation as a whole. Their students were apparently less well prepared and less likely to go to college than their peers in non-high-stakes states.
In a second report, Amrein and Berliner examined dropout and graduation rates in the 16 states that used high school exit exams in the 1990s. They found that the graduation rate decreased in 10 states after high school exit exams were implemented and increased in only five states. Similarly, dropout rates increased in 8 states and decreased in 5. They also found that General Equivalency Diploma (GED) enrollments tended to increase and the age of GED examinees decreased, indicating that more students had left school before graduating.
The authors also examined news clips from 26 states with high stakes for students or schools, to consider other consequences. This evidence is suggestive, not systematic. The authors found tendencies toward greater grade retention (a policy that fails to improve student learning while harming children); more students being expelled, in some cases apparently to drive out low scorers; and increased exemptions of students with disabilities or limited English proficiency.
Amrein and Berliner also reviewed other published research that found high stakes cause increased teaching to the tests and narrowing curriculum to fit them, with particular harm to low-income students.
Boston College Study
The National Board on Testing and Public Policy at Boston College reported that three-quarters of surveyed teachers said state testing programs were not worth the time and money. A substantial majority said the testing caused them to teach in ways that contradicted their views of sound instruction.
The board released two studies of teachers’ views of the effects of state-mandated testing on teaching and learning, one from a national survey of 4,200 teachers, the other from in-depth interviews with teachers. The studies compared the effects in low-, medium- and high-stakes states.
In both studies, teachers said higher stakes created more pressure to teach to the test. About 40 percent of survey respondents said students could raise their test scores without improving their real knowledge. As stakes increased, teachers were more likely to narrow classroom curriculum to focus more on tested areas and to engage in more test preparation, including use of items similar to what is on the exams.
The teacher interviews were conducted in Kansas (low stakes), Michigan (medium) and Massachusetts (high). As stakes increased, so did teachers’ reports of test-related effects on their classrooms. Some findings:
-Only one in ten urban Michigan teachers thought the state’s test-based scholarship awards motivated their students, compared to one-third of suburban and rural teachers.
-In Massachusetts, more than half the high school teachers thought that testing demoralized their students. Two-thirds of all teachers thought the tests were unduly stressful and unfair to special populations. Four out of five thought the exam should not be used as a sole hurdle for graduation.
In addition to the over-arching findings, the Board reports detail many of the complex and subtle ways tests with different stakes impact teaching and learning in elementary, middle, and high school levels.
Tests Demotivate
A thorough summary of research on education and motivation by a British
team found that constant testing motivates only some students and
increases the achievement gap between higher and lower achieving
students.The results of the study, titled “A Systematic Review of the Impact of Summative Assessment and Tests on Students’ Motivation for Learning,” rebut the claim that standardized testing motivates low achievers to reap the reward of high scores and avoid the punishment of failure. In fact, researchers Wynne Harlen and Dr. Ruth Deakin-Crick of Bristol University found that the two categories of students particularly discouraged by constant testing are girls and low achievers.
These findings call into question the claims of U.S. high-stakes testing proponents that they have found the key to closing the race-based achievement gap, since the study results suggest that groups such as low-income and many minority students, who traditionally score low on standardized tests, are likely to be among those who are demotivated by consistently poor test results.
The study also found that constant testing encourages even successful students to see the goals of education in terms of passing tests rather than developing an understanding of what they are learning, supporting previous research done in the United States (see Examiner, Winter 1997).
The researchers found firm evidence that achievement of literacy is linked to students’ interest in learning, the degree to which their learning strategies link to existing knowledge rather than just memorizing, and the degree to which they feel in control of their learning. The authors concluded that policymakers must recognize that high-stakes testing is providing information about students’ attainment while reducing their motivation to learn.
A pamphlet summarizing the study, “Testing, Motivation and Learning,” is available from The Assessment Reform Group, http://arg.educ.cam.ac.uk/. See also http://eppi.ioe.ac.uk/EPPIWeb/home.aspx?page=/reel/review_groups/assessment/review_one_abstract.htm.
Counterproductive
John Diamond and James Spillane of Northwestern University found that the response to high stakes accountability in low-performing schools may be counterproductive. They closely examined four Chicago schools—two performing fairly well on mandated tests, two doing poorly. All the schools did considerable test preparation, but the lower-performing schools tended to test more, focus attention on those close to passing, and engage in other activities not likely to help most of their students. Since low-income and minority-group students are concentrated in the lower-performing schools, and those schools use unhelpful methods while schools serving wealthier students use more effective methods to raise scores, the results of high-stakes accountability testing could be to widen test-score gaps.
Dropouts
Using data from the federal government’s National Educational Longitudinal Survey, Sean Reardon and Claudia Galindo of Pennsylvania State found that “the presence of an eighth grade promotion test requirement is strongly associated with an increased probability of dropping out prior to tenth grade.” This particularly affects low-income, lower-achieving students.
• Both Amrein and Berliner studies are at http://www.asu.edu/educ/epsl/EPRU/epru_Research_Writing.htm
• The Board reports are at http://www.bc.edu/research/nbetpp/reports.html
• Diamond & Spillane at http://www.northwestern.edu/ipr/publications/papers/2002/WP-02-22.pdf
• Reardon & Galindo at http://www.pop.psu.edu/general/pubs/working_papers/psu-pri/wp0301.pdf"
