High-Stakes Testing Will This Be

High-Stakes Testing Will this be on the test?' -- The failure of high-stakes testing According to Gregory J. Marchant of the Ohio Journal of Science, high-stakes testing has become a kind of educational 'bandwagon' rather than a component of educational improvement with real empirical evidence to justify its existence (Marchant 2004, p.1). At very least, serious questions remain about the reliability and accuracy of current tests used to assess individual and district-wide achievement.

High-stakes tests are defined as those tests that "carry serious consequences for students or educators," consequences which range from grade retention for school children to rewards or punitive measures for schools and school districts, or in the case of the SAT, even though it is considered an aptitude test, a failure to gain admission to a desired college (Marchant 2004, p.1).

Statement of the problem The grading mechanisms deployed by the various forms of standardized achievement tests are problematic given that "most standardized achievement tests are norm-referenced," which means that "how well an individual does on the test is based on a comparison to a large group of test takers," and an acceptable score is normed purely on how the individual excels in relation to others at the same grade level (Marchant 2004, p.1).

"This is in contrast to a criterion-referenced test that defines how well one does on a test based on the meeting of criteria or mastering a standard" (Marchant 2004, p.1).

Even criteria-based tests usually have criteria that are based upon relative comparisons of norm-referenced information, such as difficulty levels of items selected or even percentile rank of a score, and these 'norms' are often controversial and highly subjective and vary in their rigor from state to state, as in the case of the controversial state testing program known as No Child Left Behind (NCLB) (Marchant 2004, p.1). Under NCLB, states set the standards, choose tests to measure student performance against those standards, and hold schools accountable for the results.

Under these circumstances, high-stakes testing works not only as an intervention but also as an instrument to measure the outcome of the intervention. On one hand, high-stakes testing generates enormous pressure for educators to improve test scores by means of narrowing the curriculum and teaching to the test. On the other hand, any inflated test scores that can result from intensive drilling and coaching under this pressure generate an illusion of real progress and give the false impression that the intervention is working" (Lee 2008).

Students learn to take a particular type of state-specific, norm-referenced test rather than learn critical skills -- such as independent thinking and 'how' to learn. Purpose of the paper The purpose of this paper is to explode a common myth that somehow standardized tests used in high-stakes testing are inherently objective, and are immune to biases or pressure to create compressed assessments of student ability within strict state budgetary limits.

Thesis/Hypothesis Standardized tests, particularly NCLB tests, have disproportionate influence relative to their reliability, and require teachers to spend valuable classroom time teaching the test rather than individualized instruction upon a wide range of skills. The requirements of standardized tests also impede rather than enhance the quality of student education because they place disproportionate emphasis on a few subjects and skills, rather than encourage a holistic approach to learning. Review of Literature Serious problems exist regarding the reliability of the NCLB tests in many states.

Because of budgetary and time constraints, most states attempt to measure large subject domains such as mathematics and the language arts by "using tests with relatively few questions (i.e. 30-40)" and thus are at high risk of creating tests with low reliability thresholds "especially at the subscale level" (Tienken & Wilson, 2007, p. 6). One study of the state-mandated standardized proficiency tests required of New Jersey third and fourth grade classes revealed that they did not meet minimum reliability standards.

"Educators and policy-makers should use a minimum reliability coefficient of at least.85...when making high-stakes decisions about students.. [For example] there are 38 cumulative progress indicators and sub-indicators for the Number Sense / Operations / Estimation cluster of the NJASK4. Only two mathematics clusters...out of ten, 20%, meet the minimum reliability threshold for making group decisions and none meet the minimum desired reliability for making decisions about individual students (Tienken & Wilson, 2007, p. 7).

Although "one would expect higher quality assessment instruments that produce better information to make education decisions given NCLB-imposed penalties for districts associated with poor performance on the test...many states struggle with budget deficits and funding restrictions. They cannot allocate the funds necessary to improve the testing programs.

States are forced to rely on large-scale assessments with too few questions and a narrow focus on skills and knowledge that are easily measured....representatives from the NJDOE [New Jersey Department of Education] have admitted publicly that finances, not technical integrity, drive the state's assessment program..

[Their] current philosophy is 'do the best with what we have'" and is based upon the assumption that some testing, of whatever kind, is better than not testing at all (Tienken & Wilson, 2007, p.16) The irony is palpable -- the high-stakes nature of testing requires teachers, operating under finite limitations of time and money, to focus on teaching a test of questionable value, to preserve funding -- so that the teachers can continue to teach students how to perform well on standardized assessments.

One of the most common frustrations expressed by both students and educators is that the emphasis on high-stakes testing changes the dynamic of the classroom and stifles teacher creativity. Students come to devalue learning and schooling, and shift their emphasis to, "Is this going to be on the test?" (Marchant 2004, p.3).

"Time that previously was devoted to learning skills and knowledge in an appropriate sequential fashion, gets lost in the process of cramming for the tests" and those areas subject to standardized assessment such as the natural sciences, social studies, health, and open-ended writing questions, "are neglected in favor of reading and arithmetic skills that appear on the tests. High stakes testing also seems to encourage the use of instructional approaches and materials that resembles testing" like multiple choice (Marchant 2004, p.4).

Yet NCLB, as it is currently constructed, offers little incentive for improvement. Rather than embark upon quality-improvement projects, because states can set their own standards, states can merely set a new standard or norm, rather than improve instruction. Instead of placing the bar high and working to reach that goal over a number of years, states "fearing public ridicule and potential repercussions...may revisit their definition of proficient" (Marchant 2004, p.6). Given that low-performing districts may have their funding.