i think the article is partially missing the point. i'm pretty solidly convinced that i don't learn well socially, and pretty convinced that i do learn well when given a lot of space and time to explore things independently. does that mean that some abstract type exists that describes me? that it can be quantified? only very weakly.
i never interpreted the idea underlying this as a strict separation of people into types. this isn't some absurd exercise in platonism. it was always a question of individuality v. conformity. and, on that level, i don't think it would be hard to find examples of students increasing their score by trying a different approach than the standard learning model. whether those individual shifts are then applicable to anybody else is a different question.
http://www.scientificamerican.com/article.cfm?id=is-teaching-to-a-students-learning-style-a-bogus-idea&WT.mc_id=SA_MindFacebook