It’s difficult to grade student’s assignments

The four apocalyptic horsemen

I have been reading hundreds of student assignments this autumn. In many of these assignments, the students are expected to answer questions. Grading these assignments means deciding whether the answer is "right" or "wrong". This is sometimes difficult. Most answers are somewhere between "right" and "wrong". Therefore, the need for grading exposes a deeper problem: Before I can decide where on the spectrum from "right" to "wrong" the students' answers fall, there needs to be an objective "right" and "wrong". Otherwise, objective grades are an illusion to begin with (not even considering the problem of how to determine them reliably).

However, all our scientific knowledge is preliminary. It is subject to modification or even reversal when new, better data becomes available. So, how can one grade any assignment with any certainty? Luckily, not all of our knowledge is equal. One important hallmark of a scientific statement is falsifiability, a concept that was perhaps first popularised by the science philosopher Karl Popper in his book The Logic of Scientific Discovery. What are examples of scientifically falsifiable theories?

  • That the movement of objects obeys laws (approximated at slow speeds by Newton's laws of motion)
  • That the world we experience is made up of particles and their interaction with each other (quantum physics)
  • That energy cannot be created or destroyed (laws of thermodynamics)
  • That living organisms on earth are evolving (evolution)
  • That time and mass are not absolute (relativity)

All these theories have withstood the test of time. Despite countless attempts, every single experiment has shown that these theories hold up to scrutiny. And that is a high barrier to pass. After all, the observation of a single black swan overthrew the theory that all swans are white. The above-mentioned scientific theories are so thoroughly supported by our data that most scientists would call them in everyday speech as "true" because their probability of being correct is extremely close to 1.

Let’s assume that a superior alien being arrived on Earth with a spaceship that travelled faster than light. Let’s further assume that this alien explained that our universe is a cosmic experiment. And that all our evidence for historic evolution had been faked using hyper-advanced technology and that the human race started 7000 years ago with Adam and Eve, created by the very same alien designer. If that superior alien life form unequivocally demonstrated that such technology existed, that would constitute perhaps sufficient proof to question our current view of how life evolved on planet Earth. Note that the above-imagined overturning of reality as we know it has been predicted by many young-earth creationists for roughly the last 2000 years. This hypothetical event is called the "rupture" or the "second coming of Christ". To convince anybody, the superior being should not stick with a few parlour tricks that can be replicated by any good human magician. Purportedly, all humans that have ever lived will be resurrected on that occasion. While this may convince every last critic, it won't happen immediately but only after a thousand years...

Absent such a scenario, evolution is a far better explanation for what we can observe and measure. In terms of a possible superior being, I remain agnostic, which is the only possible scientific opinion. That still leaves me with the problem of grading students' assignments. The most exciting questions scientists ask are those that do not have a "right" or "wrong" answer and where there is a significantly larger-than-zero probability of any answer being overturned by better data. But to deliver an outstanding discussion of these questions, you need to be proficient in the type of scientific knowledge that is unlikely to be overturned anytime soon.