Abstract: In this article, we examine student achievement in
College Geometry, Pre-calculus, and Probability and Statistics based upon the
color-coding of the final examination. Multiple class sections taught by the
same teacher were utilized in order to keep the possible confounding variables
to a minimum. This study supports and expands the results of recently published
articles concerning the phenomenon of color bias in testing. It differs from
previously published results as we consider small enrollment classes versus
large enrollment classes, free response test items versus multiple choice test
items, multiple subjects, and the inclusion of a white examination sample to
serve as a control group for the study, thereby making a contribution to the
current body of knowledge. Using an ex-post facto design and the Kruskal-Wallis
Rank Sum Test, the study shows that, based upon the examination color, there is
no significant statistical relationship between the means of the scores This conclusion directly refutes
the common myth of color bias often encountered in the classroom.
Keywords and phrases: color bias, exam color coding, student achievement.
