Take No Prisoners

Getting the BS out of Grading

In my previous two posts, I established first that the grades assigned to students are subjective, if not entirely arbitrary; and second, that grades are pragmatically important because admissions, scholarships and jobs hinge on them. In summary, the system of grading is horribly screwed up, and we can’t just ignore it because careers, self-esteem and wheelbarrows of money depend on grades. This drew an interesting objection.

Jo writes: “So it is all bullshit. We know this. Give them what they want and move on to something better…”

I call this the “quityerbitchin” argument: yeah, you have a point, but unless you can tell us how to fix it, shut up already. With this I must agree, so let’s have at it then, shall we?

Two Alternatives to get the Bullshit out of Grading

If grades are subjective, but society pretends they’re objective, and this causes problems, we have two logical possibilities:

1. Embrace the Subjectivity of Evaluation

The first option is to simply accept that grading is subjective – to embrace the subjectivity of evaluation. While this may satisfy the social constructivists and interpretivists, it does cause some societal problems. If grades are just someone’s far-from-impartial opinion of a student, then basing job offers, admissions to schools and millions of dollars of scholarship money wholly or partially on grades smacks of incompetence and irresponsibility. Morality thus compels us to stop using grades as a primary selection criterion. However, it is presently unclear what could then serve as criteria on which to compare students.

While this option remains a possibility, it feels unsatisfying because one of the main functions of grading is to discriminate the good students from the bad.

2. Maximize Grade Objectivity

The other logical alternative is to try to make grades as objective as possible, with the understanding that a measure will always have some degree of error. This can be achieved by applying the same rigorous standard for instrument development that social scientists employ. Since a complete elucidation of the research surrounding the theory of measurement would occupy several volumes, I’ll stick to a few of the major points:

  • The instructor must know what construct (e.g. knowledge of geography, or arithmetic proficiency) he or she is trying to measure. Said construct must be clearly and specifically defined.
  • The constructs instructors are supposed to measure must be standardized nationally, or better, globally.
  • Essays, projects, reports, papers, presentations, etc. must always be graded by several graders. Lack of agreement among graders indicates a problem with the assignment.
  • “Objective” tests (multiple choice, etc.) must be pretested and validated. You have to test the test to make sure it measures what it’s supposed to measure and has no confusing questions.
  • Students should be graded against a standard (not on a bell curve) and the standard must be a bona fide national, or preferably global, standard, designed by an objective process.
  • The 1 to 100 grading scale must be replaced by a coarser scale (no larger than 1 to 5) wherein differences are meaningful enough to achieve reliable scoring.
  • Due to the difficulty of creating good measures, schools should share tests that work.

Please note that I am not advocating an academic dystopia of endless public exams filled with countless multiple choice and fill-ins. I suspect that those sorts of tests are incapable of measuring most of the variables that grades should reflect. To me, this list screams out for problem-based learning and a revolution toward educational post-modernism, but that is a topic for another day.

And so dies the “quityerbitchin” argument.