creator: Rhoades, Kathleen
description- – The goal of this 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. On-site interviews were conducted with 360 educators (elementary, middle, and high school teachers) in 3 states (120 in each state) attaching different stakes to the test results. In Kansas, state test results were used to determine school accreditation but had no stakes for students. In Michigan, school accreditation was determined by student participation in and performance on the state test and students received an endorsed diploma and were eligible for college tuition credit if they scored above a certain level on the 11th grade tests. In Massachusetts, school ratings were based on the percentage of students in different performance categories and students, starting in 2003, had to pass the 10th grade test to graduate. No clear relationship was found between the level of the stakes attached to the state test and the influence of the state standards on classroom practice. Findings suggest that other factors are at least as important, if not more so, in terms of encouraging educators to align classroom curricula with these standards. At the same time, as the stakes attached to the test results increased, the test seemed to become the medium through which the standards were interpreted. Taken together, findings suggest that stakes are a powerful level for effecting change, but one whose effects are uncertain. A one-size-fits-all model of standards, tests, and accountability in not likely to bring about the greatest motivation and learning for all students. Three appendixes contain a grid describing state testing programs, the interview protocol, and the methodology.
subjectcollectiondatepublishercreator description- – The nature and extent of human error in educational testing over the past 25 years were studied. In contrast to the random measurement error expected in all tests, the presence of human error is unexpected and brings unknown, often harmful, consequences for students and their schools. Using data from a variety of sources, researchers found 103 errors that occurred in the past 25 years. The incidence of such errors has risen dramatically since 1998. Errors were categorized as"active"(those arising from individual mistakes) and"latent"(those arising form poor managerial decisions). Active causes are attributed to most errors, but latent causes were found for 24 of the errors, with 2 cases, both in the United Kingdom, confirmed. Latent error has been characterized as more problematic, since its existence is then connected to the production of active error. The existence of latent error has only been confirmed in the United Kingdom, but confirmation requires a systemic review of the sort that is not often done. Human error has been most likely to affect public school students in grades 4 through 12, and it is estimated that well over 1.5 million students and 4,000 schools have been affected. Most errors are not detected by testing contractors and personnel; they are generally found at local school districts, state departments of education, and those intimately involved with the testing. To reduce the incidence and impact of errors, the use of multiple measures is recommended, and the use of monitoring procedures that minimize error is also suggested. Four appendixes list errors not detected by contractors, errors detected by contractors, errors in school rankings, and gray areas. (Contains 31 endnotes, and 314 references.) (SLD)
subjectcollectiondatepublishercreator
