Socionics hasn't graduated beyond a classification system so it can't possibly 'work'; as such, it's relatively accurate but the models need work (excuse the pun). Success in school has more to do with motivation, aptitude for academics and intellect (in this order); type likely contributes to the attraction toward certain fields but not success in them. And, I've met so many who could ace the tests but not succeed in the associated profession - and there's no correlation with type. Many seem to erroneously try to associate IEs with specific outcomes when they only represent information preferences and processing structure. Preference plus structure may affect behaviour but it doesn't equal success......

a.k.a. I/O