Having recently suffered through a corporate software training class facilitated by someone with an extensive miseducation in how people learn, I appreciated this article in the Chronicle of Higher Education critical of the received wisdom on learning styles.
The gist of the journal article referenced in the Chronicle is that learning styles don't have much to do with how well you learn from a given presentation, but rather how much you like learning it. The critical issue is that kinds of presentation are better suited to certain subjects. You will never learn much real science if you never spend any time in a lab for example.
The main criticism advanced against this article is that the authors failed to adequately reference the field, a scientific faux pas. However, the rebutters failed to answer the charge that most education research is poorly done, and thus cannot adequately answer the question. Unless the uncited papers are vastly superior, we will be just seeing more of the same. I noted that the authors specifically said they were not attempting to survey the entire field, which is quite large at this point.
There was quite a bit of interesting tidbits near the end of the article. I had not known that the Myers-Briggs Type Indicator and learning styles have a common lineage, and a common either/or type of categorization. I have been a bit critical of the MBTI, so my dislike of learning styles is not surprising given this common root. There was a brief review of a number of interesting attempts to sort students by ability and then check to see whether different teaching styles have an effect on students of high or low levels of ability. There is a bit on personality types and learning as well, but even the authors call this research fragmentary. I'll have to ponder the results of this some more.
h/t Jerry Pournelle