The 'learning styles' theory — that people learn best when material is presented in their preferred format (visual, auditory, kinesthetic/reading-writing) — is extremely popular in education but has been directly tested and consistently fails to hold up. A comprehensive 2008 review by Pashler, McDaniel, Rohrer, and Bjork examined the specific claim the theory requires — that matching teaching style to a person's preferred learning style improves outcomes compared to mismatching it — and found almost no methodologically sound studies supporting it, despite searching extensively for evidence given how widely the theory was already being applied in schools.
What is real: people do have genuine preferences for how they like information presented, and different subject matter genuinely suits different formats (diagrams for spatial relationships, audio for music) — that's not in dispute. What the research specifically doesn't support is the stronger claim that matching instruction to a self-identified 'visual learner' or 'auditory learner' actually improves how much that person retains, compared to just using whichever format best suits the material itself. The stronger predictor of successful learning, based on the same body of research, is using techniques like retrieval practice and spacing — regardless of anyone's stated preference.
Commissioned specifically to evaluate whether the widely-applied learning-styles theory held up, the four researchers set a clear bar: a study would only count as real support if it tested people classified by learning style, taught them with either a matching or mismatching format, and found the matched group actually learned more. Despite the theory already being applied in thousands of schools by the time of the review, the team found only a small number of studies using that design, and almost none of those found the predicted matching advantage. Some found no learning-style effect at all; a few found people performed better with a format that didn't match their stated preference.