Think about it. I'm a student taking an Electric Circuits course because assumably the course is going to be of practical or theoretical use to me at some point in the future. Ignoring the question of whether or not Cricuits analysis will ever actually be useful to someone who would love to call himself "a code-monkey", it seems fairly intuitive that I should leave this course with an understanding of the material being covered in the course.
Makes sense, right?
Right. Moving on. I need to learn the material being covered in the course. The University wishes to make sure that I learn the material being covered in the course, and by extension, so does the professor. How do they determine if I've learned the course's material? Tests.
Awesome. Tests suck, but whatever, they're a good way to figure out if a large group of people knows something that they need to know; give them some problems that can only be answered with the information or concepts that they need to know.
And from there we, for whatever reason, created this thing called a "grading scale" to group students into arbitrarily predefined levels of comprehension. And here, things start to fall apart. In order to "pass" the class, you must, typically, attain above a 60% or so, which corresponds to an F; a failing grade. This implies that the mandated amount of "stuff" you need to learn in a class using such a grading scale is just over half, right?
First of all, that is the most backwards requirement imaginable. In courses where you are learning concepts and ideas that will see application in later courses (a-la Calculus 1, 2 and 3, or Intro. to Programming and Linear Data Structures), knowing just over half of the material is academic suicide! But Universities typically require certain courses be taken prior to other courses (called pre-requisite courses), which very strongly implies that the material in the pre-requisite course will see moderate application in the higher level course. But even these pre-requisites operate under this "Just learn more than half of the material and you're good" mindset.
The fuck? The University requires that course A is taken before course B, because the students need to know the material from course A to succeed in or to understand course B. Great. Fine. But their general standard for "knowing the material" in a course is just over 50%? In what world does that make the slightest bit of sense?
That's like saying "You understand addition and subtraction well enough, so despite not understanding multiplication and division, I think we can safely move on to Algebra". You need to understand all of the material completely in order to move on, at least that's how pre-requisites typically (and effectively) work.
Once you get to High School though? That is gone. As are all pretenses of education as the focus of the educational system, but that's a rant for another day.
More importantly, however, is the fact that generally speaking a student can pull off a barely passing (or even convincingly passing) grade while comprehending close to nothing from the course. I could give anecdote after anecdote where I've done exactly that, but I'd rather just point out a widespread systemic cause for this issue, because that just seems more convincing to me.
Curves. We've all experienced them, and most of us have gotten a benefit of some kind from them. Theoretically, a curve is nothing more than the arbitrary raising of an entire class's grade due to perceived difficulty in an assignment or test being greater than the capabilities the class should ideally posess. If a prof or a teacher gives a test that is way too difficult for his/her students, or that could not be reasonably completed by a student that comprehends the material in the time allotted, that is when a curve should be implemented.
Practically speaking is another story entirely. In practice, curves are implemented whenever the graph of the class scores do not produce a curve that orients itself over the arbitrarily chosen "normal curve" which clusters it's scores around the 70%, or the C-grade portion of the graph. In other words, whenever the class as a whole scores lower than is arbitrarily and universally predicted across the board for all courses and students, it is expected that the scores will be arbitrarily raised to meet this arbitrary and universal standard.
Unfortunately, the moment we alter grades arbitrarily to fit a pre-defined standard, grades cease to be a trustworthy indication of student comprehension of concepts in the worst possible way. We aren't accidentily generating false negatives here; we're generating-en masse-false positives.
We're incorrectly and inappropriately flagging individuals with little to no comprehension of a course's materials as having a higher level of comprehension than they really do. Oftentimes this can actually lead to individuals that comprehend far less than 50% of the course's material receiving grades of 70%, or even 80%, which allows them to more or less pass a course that they comprehend next to nothing from.
And then, we send them on to the next course, because they've met the "more than 50% comprehension" requirement thanks to the arbitrary raising of their grades in order to conform to an arbitrarily chosen "standard comprehension level". Students with little to no comprehension are sent to Calc 2, feeling quite proud of themselves for understanding enough to pass Calc 1, and suddenly the topics are more complicated, and they build off of all the topics they never learned or understood from Calc 1, and they wonder why they fail!
This current system, with it's arbitrary grading standards and it's universal grading guidelines is completely insustainable, and eventually it's going to break under it's own failings. The already worthless grades are going to become that much more worthless, and the education system will have to be scrapped and remade from the ground up at that point.
But it's already a disservice to the students and companies that look towards grades as the only feasible indication of student comprehension, as they are constantly disappointed by the reliability of this data, and we really need to start looking at a way to fix this very broken system.