The STAR Math test as a predictor of Arkansas BenchmarkTest scores

Abstract

With the failure of the legislative branch to reauthorize the No Child Left Behind law in 2007 and recent reports that more schools than ever are failing to achieve Adequate Yearly Progress, educators are reviewing practices and curriculum. As a result of federal and state laws, it is necessary to identify an accurate predictor of student achievement prior to the administration of the state-mandated test. For this study, student samples were drawn from sixth, seventh, and eighth grade populations of a Northwest Arkansas Middle School. Samples were separated by grade level and ranked according to the grade equivalency on the fall STAR Math pre-test and the scores on the spring Arkansas Benchmark Test. A quasi-experimental design was implemented to test both the magnitude and reliability of the independent variable, the STAR Math test, on the dependent variable, the Arkansas Benchmark Test. A Pearson r correlation was calculated in each grade level over a three-year period for the relationship between the STAR Math and Arkansas Benchmark. A strong positive correlation was found between the ordinal ranks of grade equivalence on the STAR Math pre-test and the ordinal ranks of the averaged raw score percent on the Arkansas Benchmark Test. Furthermore, a coefficient of determination, a line of best fit, an analysis of variance (ANOVA), and an Omega-squared were used to determine the statistical significance and develop a triangulation of data. Further study is recommended to predict a specific benchmark score based on a STAR Math grade equivalency.