<p>"Alright! Now that is an answer!</p>
<p>What is the “right” level of stratification?</p>
<p>So the education system has been doing it wrong for hundreds of years?"</p>
<p>There is no “right” level of stratification, that’s totally dependent upon what inference you want to draw from the grade. For example, the performance on four totally different exams covering different material + home work assignments that all cover different material + a lab which covers a totally different competency combined with weights that are different in every course at every university spits out a number, like say 88. Now, let’s not even get into the problem with using a single point score from an exam as an exact measure of success, but let’s look at this 88. How is this different from a 92? If you want a system which has +/- that means that you believe that through that whole process you can fundamentally distinguish between someone who earns an 88 versus a 92 and that this accurately reflects a different level of understanding. What if one person took the course with a different professor who included homeworks more and the labs less? That same person, with all of the same scores in the same class on the same exams now could have an 86 or a 92. Does that mean they are demonstrable different from the 88? Should this student get a B, B+, A-? Maybe an A? What if the class average was an 88? What if the distribution of grades in the class fell between an 84-94? How does this change your grade? What if one professor in that situation thinks it means his whole class has learned the material he wanted them to very well and gave everyone an A? Would he be wrong if he believes that earning an 84 would be a sign of superior success? What if he thinks that he should grade on a normal curve and an 84 would fail, yet that 84 with slight adjustments could easily become a 90 which would earn a B+? What if another person thinks that 90 should be a strict cut-off because that represents A level work so everyone from 90-94 is an A and 84-89 is a B? Is 89 different from 90? What justifies that change?</p>
<p>Yes, a lot of education has this whole thing wrong and there are whole books on it written by experts on testing and there have also been many changes at lots of institutions to move away from these systems.</p>
<p>You should look into measurement error, sampling error, bias, reliability, and validity in testing. Read Koretz on testing, or maybe Marzano on classroom grading.</p>
<p>I could get very technical about this rather than stream-of-consciousness examples right now, but the truth is that it’s nearly impossible to make grades a meaningful progress metric at the university-level for comparison across even multiple sections of the same course. That task, at least is possible. It’s essentially impossible to standardized the meaning of a grade across different courses, and even more impossible across institutions.</p>
<p>Brown de-emphasizes the inaccuracy built into these grades by decreasing the stratification, decreasing the emphasis on the grades, and increasing the emphasis on learning through both structures and culture. Measure somebody by what they know, but realize that grades are not particularly reliable measures of that in almost all cases.</p>