2 ms·

Fair point, but even if you can predict paper grades within half a letter grade, there's something much different psychologically about the following: "This pro

by emily37 17y ago

Fair point, but even if you can predict paper grades within half a letter grade, there's something much different psychologically about the following: "This program passes all the automated tests, and the automated tests are 90% of the grade" and "I feel very strongly that this is an A paper, and I'm almost always right about what grades my papers will get." The former is a much more confident statement. Perhaps the quote in the article is meant to say that the grading of CS assignments is almost entirely deterministic; there's always the possibility that your prediction of your paper grade will be way off, but if you pass all the automated tests, you're getting an A or A-.

tsally 17y ago
There's a big psychological difference: you have faith in the determinism grading computer but not in your English professors. The funny thing is that an English professor is more qualified to say whether your paper is a good paper than a grading computer is to say whether your program is a good program. Automated tests are not a production environment for a program. A living, breathing reader is a production environment for a paper. Might I be too bold in suggesting that CS students prefer the determinism of a grading script because they are worried how their code will hold up in the real world? Given the quality of code most CS students produce, that's actually not that bold of an assumption. Seems to me that CS students that do not favor how grading is done in English have that opinion because they aren't used to thorough assessment in CS [1]. [1] The notable exception being theory classes. Go theory! But in those classes you've got a living, breathing grader as well. :-p EDIT: I realize I didn't really address your point. Yes, that is a fair characterization of the quote in the Times. The problem lies in assuming the grading script is doing as good of a job as your English professor in determining correctness. I bet if you increased the accuracy of the grading method for programs you would sacrifice on precision and students wouldn't be as confident about their programs. If you graded on coding style and design decisions, for example. So yes, CS students can be more confident. But confident of what? The fact that your program conforms to a script? That's not the same as being confidant that your program actually works.

by emily37 17y ago

tsally 17y ago

There's a big psychological difference: you have faith in the determinism grading computer but not in your English professors. The funny thing is that an English professor is more qualified to say whether your paper is a good paper than a grading computer is to say whether your program is a good program. Automated tests are not a production environment for a program. A living, breathing reader is a production environment for a paper. Might I be too bold in suggesting that CS students prefer the determinism of a grading script because they are worried how their code will hold up in the real world? Given the quality of code most CS students produce, that's actually not that bold of an assumption. Seems to me that CS students that do not favor how grading is done in English have that opinion because they aren't used to thorough assessment in CS [1]. [1] The notable exception being theory classes. Go theory! But in those classes you've got a living, breathing grader as well. :-p EDIT: I realize I didn't really address your point. Yes, that is a fair characterization of the quote in the Times. The problem lies in assuming the grading script is doing as good of a job as your English professor in determining correctness. I bet if you increased the accuracy of the grading method for programs you would sacrifice on precision and students wouldn't be as confident about their programs. If you graded on coding style and design decisions, for example. So yes, CS students can be more confident. But confident of what? The fact that your program conforms to a script? That's not the same as being confidant that your program actually works.