The Pernicious Predictability of State-Mandated Tests of Academic Achievement in the United States

The purpose of this study was to determine the predictiveness of community and family demographic variables related to the development of student academic background knowledge on the percentage of students who pass a state-mandated, commercially prepared, standardized Algebra 1 test in the state of New Jersey, USA. This explanatory, cross-sectional study utilized quantitative methods through hierarchical regression analysis. The results suggest that family demographic variables found in the United States Census data related to the development of student academic background knowledge predicted 75 percent of schools in which students achieved a passing score on a state standardized high school assessment of Algebra 1. We can conclude that construct-irrelevant variance, influenced in part by student background knowledge, can be used to predict standardized test results. The results call into question the use of standardized tests as tools for policy makers and educational leaders to accurately judge student learning or school quality.