Fair and Balanced? Experimental Evidence on Partisan Bias in Grading

Abstract

Is grading polarized in political science classrooms? We offer experimental evidence that suggests it is not. Many have argued that instructors’ grading in political science classrooms is skewed by the political characteristics of the instructor, the student, or an interaction between the two. Yet the evaluations of whether such biases exist has been asserted and denied with little evidence—even though prominent theories in political science suggest that the charge is not entirely implausible. Using a set of anonymous essays by undergraduates graded by teaching assistants at a variety of institutions, we test for the presence of bias in a framework that avoids the usual selection bias issues that confound attempts at inference. After evaluating the evidence carefully, we find that the evidence for bias is much weaker than activists claim.