Predicting Middle Level State Standardized Test Results Using Family and Community Demographic Data

摘要

The use of standardized test results to drive school administrator evaluations pervades education policymaking in more than 40 states. However, the results of state standardized tests are strongly influenced by non-school factors. The models of best fit (n = 18) from this correlational, explanatory, longitudinal study predicted accurately the percentage of middle school students scoring proficient or above on the New Jersey state-mandated standardized tests in mathematics and language arts for grades 6–8 during the years 2010, 2011, and 2012 for 70% to 78% of the schools in the statewide samples (n = 292 to 311), using only family and community demographic variables from the U.S. Census. Just three demographic variables, (a) percentage of families in a community with income over $200,000 a year, (b) percentage of people in a community in poverty, and (c) percentage of people in a community with bachelor’s degrees, predicted results accurately in 14/18 of the models. The findings suggest that state standardized test results are not as objective and transparent as advertised by state and federal department of education officials. Some middle level school administrators might be getting rewarded or punished based on factors that they do not influence.