The Debunker Club, organized by Will Thalheimer, is focusing on debunking Learning Styles this month and have devoted a page to resources to this end. An excerpt from that page:

Probably today’s most ubiquitous learning myth is that people have different learning styles and that these learning styles can be diagnosed and used in learning design to create more effective learning interventions. This myth has resonated and spread throughout the world’s learning-professional community probably because it hints at an idea that seems sensible — that people learn differently. Unfortunately, there are dozens and dozens of ways to separate people by type, so it’s hard to know which distinctions to use for which learner, for which topics, for which situations. More importantly, the research evidence shows clearly that using learning styles in designing/deploying learning does not reliably improve learning results.