I have learned about some inappropriate and wasteful research efforts. Please don’t add to my list!
Researchers have tried to find differences between individuals’ preferences when they are learning some specific skills. That may be appropriate when the skill is carefully chosen. Teaching children to swim is unlikely to be effective when the teaching methods are predominantly Read/write or Visual. There has been VARK research done on swimming, knitting, tying knots and sign language. Often the researcher has failed to note that such tasks have a built-in bias and differences between individuals may be meaningless. Of course, people who have a relatively high score for the Kinesthetic mode will learn better (and faster) if the task is a Kinesthetic one! And lawyers may have a bigger proportion who have Read/write preferences than body-shop workers. But beware! Most of the VARK database (60%) is composed of those who have no particular single preference. They are multimodal and use two, three or four modes to help them learn. They will dominate the statistics and make conclusions invalid unless the researcher can use very sophisticated statistical measures to unravel their multiple characteristics.
Some have tried to categorize teaching methods as being more Kinesthetic or more Visual, but again, beware! We are multimodal beings and there may be variety even when somebody appears to be using only one mode to teach or train. An instructor using PowerPoint is probably using mainly Read/write methods with print on the screen and some Aural output. If that same instructor chose to add many examples, case studies and applications it would better suit those who prefer to learn using their Kinesthetic preferences. The addition of graphs, diagrams, charts, maps and an overview schematic would assist those who would have Visual preferences. Because it is on a screen does not make it a Visual!