I try to live my life as if grades aren’t important to me, but the Thayer School has some simply outrageous practices. I attribute them almost entirely to the homework graders, who are mostly students.
Today I got back a problem set.
In part 1 my Modified Newton’s Method had an error, and didn’t converge as rapidly as it ought to have. The number of iterations is circled, and 10% is taken off. Fair enough.
I noticed the bug right before the assignment was due, and typed up a paragraph documenting it, and stating what I thought was the probable cause. This cost me another 20%. Academics frequently provides powerful incentives for unethical behavior, and equally powerful disincentives for ethical behavior.
One of the problems said “Specify the asymptotic oder of convergence, α, and write the asymptotic error constant λ.
I started the problem by writing out the text book definition of lambda, with the variable α correctly filled in as 3. Six pages of algebra later, I had a correct expression for λ. I then derived the general definition of λ itself, and again wrote in the correct value for α, and I included written explanation of the technique used in the for the derivation.
I’ll admit that nowhere did I use the precise phrase ”α = 3”.
Literally directly under my second restatement of the definition, not even 2 cm from the value for alpha, is written “where is α? – 22%”.
Finally, we were asked to comment, qualitatively, on the rate observed rate of convergence on an application that required 2 iterations. My answer of “3 data points is not enough to determine a rate of convergence” was, apparently, incorrect. The correct answer was “quadratic”.