> HOW do you measure knowledge? And when you decide how, how do you scale it?
When I was in school, many moons ago (in France) there were no quizzes. Zero. "Tests" were either dissertations (for topics such as literature, history, etc.) or problems. Everything was done in class, in longhand.
There were no good or bad answers, even in math class, because what was evaluated was the ability to describe the problem, the approach, and the solution, and you got points for that even if the ultimate result was wrong.
"Cheating" was very difficult; copying what another student was writing was hard and not very effective, because unless you could reproduce their whole argument, just taking a sentence or two would not make sense.
This system didn't "scale" very well; in fact it didn't scale at all.
If you build a system that let one person "teach" classes of hundred of students and generate quizzes that can be instantly rated by a machine, then some (most?) students are going to try to game that system.
This is inevitable, and I'm not even sure it's a bad thing.
> "Cheating" was very difficult; copying what another student was writing was hard and not very effective, because unless you could reproduce their whole argument, just taking a sentence or two would not make sense.
In high school we often were given two (or more) sets of problems so we can't copy off each other because people sitting one sit away from each other have different sets.
I remember at least one test where I wrote down problems from both sets (they were verbally dictated by the teacher at the beginning of the test). Then I just solved both and passed the solution to his problems to the classmate sitting behind me (I was asked for this ahead of time by him).
In Poland cheating is frowned upon by teachers and they tried to catch the cheaters but there were no formal systems in place to report or excessively punished cheating (like in USA).
Yes, although many of the students in the story weren’t interested enough in learning, had “low morals”, “no honor” and some apparently were scumbags, as a group they were somewhat efficiently solving the problem of passing the class… that’s not nothing!
In the real world the solutions to your problems can’t be found online, or if they can it’s valid to search them there (and lawyers will charge you a lot to do that). Collectively searching and distributing a solution is something young people are quite adept at (e.g. gaming wikis).
When I was in school, many moons ago (in France) there were no quizzes. Zero. "Tests" were either dissertations (for topics such as literature, history, etc.) or problems. Everything was done in class, in longhand.
There were no good or bad answers, even in math class, because what was evaluated was the ability to describe the problem, the approach, and the solution, and you got points for that even if the ultimate result was wrong.
"Cheating" was very difficult; copying what another student was writing was hard and not very effective, because unless you could reproduce their whole argument, just taking a sentence or two would not make sense.
This system didn't "scale" very well; in fact it didn't scale at all.
If you build a system that let one person "teach" classes of hundred of students and generate quizzes that can be instantly rated by a machine, then some (most?) students are going to try to game that system.
This is inevitable, and I'm not even sure it's a bad thing.