“Norm-referenced grading practices widen the achievement gap for college students underrepresented in STEM”
Abstract: In the study, we explored grading practices at UCLA including data from the 2010-2011 to 2015-2016 academic years. A k-means cluster analysis was utilized to categorize the courses into 16 clusters, which were assigned to either norm- or criterion-referenced grading schemes. The assigned grading schemes were confirmed through a course syllabus analysis and a department chair questionnaire. Our results demonstrate that clusters assigned to norm-referenced grading practice tend to be associated with lower average grades and higher fail rates. When analyzing the data disaggregated by URM, SES and gender status, we observed that norm-referenced grading selectively disadvantages URM and low-SES students. Furthermore, departmental grading policies and expectations are correlated with observed grade distribution outcomes and could therefore be a good starting point to improve the success of URM and low-SES students if changed to represent more evidence-based and inclusive practices.
Authors: Nadia Sellami, Kelly Wahl, and Erin Sanders
This study extends work described in the Enhancing Student Success and Building Inclusive Classrooms Report to the EVC/P by Victoria Sork and Sylvia Hurtado, December 2015. Please click here to access the PDF files and to download the report and relevant appendices.