Do learning styles exist?

That is, do some people learn stuff better by hearing it, or seeing it, or rubbing it on their heads? The societal consensus is that it is true. A report in Psychological Science in the Public Interest, a journal of the Association for Psychological Science, and reported in Science Daily says no.

Nearly all of the studies that purport to provide evidence for learning styles fail to satisfy key criteria for scientific validity. Any experiment designed to test the learning-styles hypothesis would need to classify learners into categories and then randomly assign the learners to use one of several different learning methods, and the participants would need to take the same test at the end of the experiment. If there is truth to the idea that learning styles and teaching styles should mesh, then learners with a given style, say visual-spatial, should learn better with instruction that meshes with that style. The authors found that of the very large number of studies claiming to support the learning-styles hypothesis, very few used this type of research design. Of those that did, some provided evidence flatly contradictory to this meshing hypothesis, and the few findings in line with the meshing idea did not assess popular learning-style schemes.

I don't have a dog in this fight. It seems plausible that learning styles would exist, but I haven't actually seen any evidence that it is so. As the paragraph above says, it should be easy enough to test on a large enough scale to believe the results. Let's somebody do so.