Yan Zhan

University of Colorado Boulder · Economics · 2026–27 job market

Fields

Labor, Education, Applied Micro, Econometrics

Job market paper

Teacher Grading Heterogeneity and Luck in High School GPA

Students taking the same course in the same high school are often graded by different teachers. I first ask whether these teachers meaningfully differ in their grading standards within the same track and year, and if so, by how much. I then ask how much a student's course grades, cumulative GPA, and class rank could vary depending on which teachers they are assigned, holding fixed their relative performance. Using administrative data from North Carolina public high schools, I estimate, for each subject, a mixed-effects model of classroom grading moments (mean, standard deviation, and skewness) with school × year × track fixed effects, a teacher random effect, and controls for students' performance in their other courses. On the first question, teachers differ meaningfully in how they grade: a teacher one standard deviation above the average gives classroom grades about 2.3 percentage points higher. On the second, simulations based on these estimates show that a student's grade in a single course can differ by 4 to 7 percentage points depending on which teacher they are assigned. Averaging across courses shrinks this gap to about 1.9 points in cumulative percentage score, but rounding course grades to letter bands before averaging brings some of it back: the cumulative-GPA gap is 0.21 points on a fine-letter scale and 0.28 points on a coarse five-level scale. Teacher assignment moves class rank by 6.4 percentile points on average when rank is computed from exact percentage scores, and by up to 10.6 points when it is computed from coarse letter grades.