A methodology for measuring sociotype
Quote:
Originally Posted by
InkStrider
True. The question is, how could this be addressed? People use different 'type indicators' to gauge a person's sociotype, which is probably the root of typing discrepancies.
I have a few ideas. IMO the only real way is a more form of information analysis that does not rely on questionnaires and self-reporting but based on choices(information preference) analysis.
More akin to Google +1, Facebook Like and rating movies. I was engaged in the Netflix prize prior to doing socionics seriously and started studying socionics again to see if it could help my analytics, one of the most common methods and at the core of all analysis like this is http://en.wikipedia.org/wiki/Singula..._decomposition which generates 2^x categories of indicating and contraindicating values(significant choices will fall into one or the other category). Think of SVD as a categorization prior to a prediction.
Of course this could produce a type prediction, 2^x categories with choices that would indicate one type or a contratype(i.e the conflictor).
This is a common techniques in information analysis, and is used in many schools of application, any number of systems will generate 2^x categories for prediction. Including choosing what sort of advertising to display, whether or not something is a security threat or whether or not you would like a movie. I would like to do something like this for socionics, but this is the sort of thing that takes thousands upon thousands of hours of work to implement.
This isn't really anything that is related to this thread. So I will move it.