Nice post!
Re: Knewton, I think you are making a false dichotomy. Do you actually know what kind of analysis they are doing? My understanding is that they do what might be called "statistical bug detection" (among other types of analyses). I would say that this is one of the holy grails of psychometrics, which few people can approach because there is a high hurdle in terms of knowledge engineering. e.g. you need a graph to tell you which production rules (whether good or buggy) lead to which kinds of behavior.
Re: Knewton, I think you are making a false dichotomy. Do you actually know what kind of analysis they are doing? My understanding is that they do what might be called "statistical bug detection" (among other types of analyses). I would say that this is one of the holy grails of psychometrics, which few people can approach because there is a high hurdle in terms of knowledge engineering. e.g. you need a graph to tell you which production rules (whether good or buggy) lead to which kinds of behavior.
Comment Form