0.7 Ward, s. (february 2009). what every educator should know about  (Page 5/7)

To further complicate matters, by their very nature standards-based assessments have limitations. The characteristics of the tests themselves can make the process murky with variations in the difficulty of items and the mix of item formats (K. P. Boudett, et. al., 2007). Levels at which performance standards are set depend on multiple factors, including the judgment of the panels assembled to set them and the particular method used to do so. (K. P. Boudett, et. al., 2007). For example in Georgia, student performance standards for the Criterion Referenced Competency Tests (CRCT) are established through a standard setting process in which educators from around the state participate. Educators make recommendations on what scores define categories of student performance. As a result of this process, student scores on the CRCT are reported in the following scale scores and performance levels: Below 800 Does Not Meet Expectations, 800-850 Meets Expectations, 850 and higher Exceeds Expectation, and scores above 900 generally indicate exceptional performance (Georgia Department of Education, 2008).

Similarly, in California the performance levels or cut scaled scores for the state content standards assessments were determined by a group of teachers, school and district level administrators, academics and county office educators in the kindergarten through university arenas. According to members of this group, cut-scores for proficient were determined based on an agreement reached by group consensus. If every state developed proficiency levels like Georgia and California, not only would there be a minimum 50 different definitions of proficient, but the proficiency levels in some states could be deemed irrelevant.

Focusing on California again, part of the problem with the definition may be that cut scaled scores for proficiency are too low. Children in California need only score 350 on state tests to be considered proficient. Since the range of the test is from 150-600, it could be argued that proficient amounts to surprisingly basic levels of performance! According to Fuller et.el. (2006), how cut points are set may increase the mastery of basic domains by low performing students. Additionally, the cut points for determining which students are deemed proficient are set at varying levels across states. Within a given state, cut points also shift over time (Linn, 2001).

Returning to our opening quote on New Jersey state exams, until most recently New Jersey youngsters only had to answer 33% of the state tests correctly to be categorized as proficient. On July 15, 2008, the New Jersey State Board of Education voted to raise the cut scores in grades 5 - 8 for proficiency in reading and math. Now students in those four grades must answer at least 50% of the questions right to be deemed proficient. The state Education Commissioner Lucille Davey was quoted as saying, “What we don’t want to do is mask our weaknesses….The incentive may be to have the lowest standard we can, but that won’t help our kids” (Mooney, 2008, July 16). The New Jersey State Department of Education is recalculating student scores on the 2008 exams based on the new proficiency cut scores. Up until this point, 76 percent of New Jersey 6th grade students statewide passed the language arts examination (New Jersey Department of Education, 2007; Mooney, 2008, July 16). This number is expected to drop to 54% after 2008 scores are recalculated (Mooney, 2008, July 21).