Physicists tend to have very good modeling chops. A fellow in the early 1900s (Ising) was trying to come up with a model for ferromagnetism and came up with Markov random fields, basically. That is amazing to me.
Meanwhile, in psychometrics: “I know, let’s model intelligence by one number!”
edit: There is some controversy about how much of this was Ising vs Ising’s advisor. This does not affect my point about physicists, though.
Right, I kind of swept that under the rug as part of approximations—as in, ‘try making a seemingly overly-simple model of the individual components and see if the relevant behavior emerges’. Could have been clearer on that.
Physicists tend to have very good modeling chops. A fellow in the early 1900s (Ising) was trying to come up with a model for ferromagnetism and came up with Markov random fields, basically. That is amazing to me.
Meanwhile, in psychometrics: “I know, let’s model intelligence by one number!”
edit: There is some controversy about how much of this was Ising vs Ising’s advisor. This does not affect my point about physicists, though.
Right, I kind of swept that under the rug as part of approximations—as in, ‘try making a seemingly overly-simple model of the individual components and see if the relevant behavior emerges’. Could have been clearer on that.