When I was an engineering student at Georgia Tech, the first question every student wanted to know after a test was graded was “What was the class average?” The class average was our foremost concern because we knew that the average determined the “curve” – the upward adjustment that would be given to our test score based on this average. If the class average was a 35 (which sadly was not uncommon), you knew you were alright if you made a 55 because the average would be adjusted, to say a 65 (which was a D at the time), giving you a solid 85 score. If you were below the class average, though, you knew you were in trouble because that meant an F. In spite of this practice, failure always seemed like an immanent possibility, and students often “flunked out” of Georgia Tech. I’m sure the class average is still a common obsession at Georgia Tech and among college students throughout the country.
Now as a teacher I realize that when the majority of your students to poorly on a test you should reexamine the test and your teaching strategy carefully to make sure the all the test questions were fair and thus consistent with the curriculum you taught. Sometimes adjustments to the “raw scores” should be made for the sake of fairness. But it is highly problematic when test scores are regularly adjusted for the sole reason of bringing scores up so that the grade distribution looks healthier.
This approach to grading in which student learning is evaluated, and grades given, relative to the performance of their peers is called “norm-referenced assessment.” I see a number of problems with such assessments. First, they do not accurately communicate to the student and to “stakeholders” (parents, prospective employers, future educational institutions) what the student actually has learned and is able to do. The grade only shows that the person performed better than his peers. Second, the meaning of grades changes over time, typically in a downward direction. For example, what if students who took Chemistry at Georgia Tech in 1995 learned more chemistry and made a 65 average on the final exam, which the professor adjusted to a 70. In 2010, the chemistry students at Tech learned less chemistry and made a 50 average on the final exam, which was also adjusted to a 70. In this scenario the student who made a 70 in 2010 knows less Chemistry than the student who made a 70 in 1995 because there is no fixed reference point by which to interpret these grades: they both made a 70 because they were no better or worse than their peers in that particular year.
The widespread practice of norm-reference grading has put our society in a situation where grades cannot be trusted because their meaning is ambiguous. I have heard administrators say that colleges hardly consider GPA anymore as a meaningful criterion for admission and that they even have little regard for performance in “honors” courses. This is why I relate grade inflation to the ethical problem of lying or, using biblical terminology, “bearing false witness against your neighbor.” Grades are a witness or a public, official testimony to others about what a person has accomplished education. When grades to not convey this accurately, their testimony is false and thus the grades lie. So accuracy in grading really is an ethical matter, and not just when it comes to being upright in the administration of standardized tests.
Part of the solution to this problem lies in the practice of “criteria-referenced assessment” in which student learning is evaluated with reference to pre-specified criteria or standards. The push for “Standards-based” classrooms is, fortunately, common in education these days, but it is uncommon to find it consistently and rigorously practiced. It is tempting to not uphold standards when grading when it is in a teacher’s self-interest for the grades to have “pretty” grade distributions.
A professor at Georgia Tech of over 25 years, when I was in graduate school, once confided in me that “we just do not fail students anymore.” I asked why. He explained, “It is just too much trouble for us to fail students; it is much easier for us just to pass them.” That same year another veteran professor told me that the new graduate students in chemical engineering had a very weak understanding of the fundamentals of the discipline. He related this situation to an observation that his colleagues give more A’s and B’s on undergraduate tests than they used to because they did not want to have to deal with struggling students and wanted to be left alone. Unfortunately, it was not in their self-interest that their students were actually learning chemical engineering.