Differential Grading Standards and Student Incentives

We present data on grades from three Canadian universities. These data suggest that grading standards differ significantly across disciplines within universities. To the extent that grading standards are not uniform across disciplines, the GPAs of students with different course mixes cannot be meaningfully compared, and therefore their GPAs cannot legitimately be used to assess their relative achievement. Yet GPAs are used in precisely this way-to award scholarships, honours and degrees, and to ration access to courses, academic programs, and jobs. Hence, we think differential standards raise a fundamental issue of integrity for universities. We develop a simple human capital model to assess some of the distortions that differential standards give rise to. We suggest some non-intrusive ways to rectify the problem.