Children and Science Tests Research Team: Josiane Hudicourt-Barnes, Tracy Noble, Mary Catherine O'Connor, Ann S. Rosebery, Catherine Suarez, Beth Warren, Christopher G. Wright
—(2008) Symposium paper presented at the Annual Meeting of the American Educational Research Association. New York, NY, March 24-28, 2008.
Every high-stakes, state level achievement test reports discrepancies—a gap—among the scores of students from various "demographic" groups: students who qualify for free or reduced lunch and those who do not; students who speak English as a first language at home and those who do not; students who identify as White and those who do not. This discrepancy is often called the 'achievement gap', and a number of explanations have been put forward to account for it. Chief among them is the unequal access to resources that exists between low and middle-income school districts and the impact this inequality has on the quality of the teaching and learning that can take place. Within this explanatory framework, it is assumed that if a child in a low-income district answers a science question on the state test incorrectly, it is because the content was not taught, or not taught well.
In this paper, we report on some factors that we hypothesize to be possible causes of unintended item difficulty for low-income students and English language learners; we share some insights into why students choose the answers they do; and we attempt to characterize these preliminary findings in ways that may lead to some concrete suggestions that test-makers and educators can use in evaluating the meaning of students' test performance.