pull down to refresh

I was thinking about our common modelling practice of dividing the population into "high types" and "low types".
Yes, what are high types and what are low types? These classifications look suspiciously collectivist.
reply
It's a modeling gimmick to classify people into low skill and high skill groups, with the simple binary being chosen simply for mathematical convenience. It lets the model generate heterogeneity in outcomes such as high types getting degrees and low types not
reply
Another statistical modeling gimmick, huh? How are they working out for the people using the models?
reply
In terms of getting published I guess it's working out just fine
reply
Getting published? How is it doing on modeling reality? Or is it just a model constructed from previous data?
reply
It's not really a statistical technique. It helps work out simplified equilibrium conditions. If it helps, it's no worse of an assumption than the standard one-type assumption, which Austrians make all the time.
I think Austrians should take results from this kind of work as clues or hints as to what might be true. It's a lot easier to engage in deductive reasoning if you have a sense of the destination.