Working on the usual stuff - with homework now on top of the independent research. And daily life is constantly intruding too of course.

My main work lately has been trying to develop typologies of educational pathways (specifically for science and engineering majors) with some great transcript data. This sort of exploratory data analysis is relatively new to me. We're trying a variety of clustering strategies, but if anybody knows good ways of looking at pathways over time I'd be interested. I'm familiar with group based trajectory analysis and plan on trying that, and we're also going to set up some kind of Markov chain model. The problem with that is that coursetaking isn't dichotomous. You can be taking multiple courses at once. So something like a continuous Markov chain model would be great but I don't know of such a thing.

Also potentially of interest - I picked up

*Pure Theory of Capital*from the library yesterday for the

*Critical Review*article. It's not going to get technical enough to really differentiate between Hayek's work in detail, but I want to familiarize myself enough with it so that I don't miss anything critical.

"My main work lately has been trying to develop typologies of educational pathways (specifically for science and engineering majors) with some great transcript data. This sort of exploratory data analysis is relatively new to me. We're trying a variety of clustering strategies, but if anybody knows good ways of looking at pathways over time I'd be interested. I'm familiar with group based trajectory analysis and plan on trying that, and we're also going to set up some kind of Markov chain model. The problem with that is that coursetaking isn't dichotomous. You can be taking multiple courses at once. So something like a continuous Markov chain model would be great but I don't know of such a thing."

ReplyDeleteI think that it's great you're doing exploratory data analysis. Better than assuming the standard normal distribution without doing a goodness of fit test! Perhaps you might want to consult Dr. Michael Emmett Brady for advice regarding statistical testing, as I think he's fairly familiar with the literature regarding this issue.

How good are you making Markov chains, anyway? I thought you said somewhere before that you didn't have a very good teacher when it came to that stuff.

Regarding

The Pure Theory of Capital...is the library you're borrowing it from the American University library, or a different academic library?I'm not sure I understand your normal distribution point. I just mean that I am generating categories with cluster methods for the time being and presumably other methods in the future. I'm not sure you're using the word in the same way unless you simply mean that I'm recognizing there are distinct sub-groups I'd like to talk about.

DeleteHey Daniel,

DeleteI heard your toilet is broken. Perhaps you might want to consult Dr. Michael Emmett Brady for advice. He has some great writings on heterodox plumbing!

Daniel,

DeleteMises would reject the notion that your constantly flushing toilet is somehow a sign of "toilet failure" because Mises rejected the idea that the water in your toilet rests between the flushings in an equilibrium-like state. Enterpreneurs will break into your house and flush the toilet because there are profit opportunities. You should read something about his evenly rotating toilet.

Daniel,

Deleteyou are young and smart but the dynamic modelling of toilets is a deep and subtle issue. Read Alchian 1990.

There is absolutly no merit in studying Hayek's Austrian theory of plumbing. As Sraffa has shown in 30s there is more than one "natural rate of flushing".

ReplyDelete@Lord Keynes

ReplyDeleteThis is bullshit. You can deduce logically from basic axioms that the problem is actually too much toilet paper issued by the Fed.

We should go back to gold instead of toilet paper out of thin air.