If you want to trigger an L&D professional, bring up learning styles. The industry has been arguing about them for decades, and at this point, the conventional wisdom is that matching instruction to learning styles doesn’t actually improve learning. But that hasn’t stopped people from designing around them, advocating for them, or even researching them. As Selena Gomez would say, the heart wants what it wants when it comes to learning styles.
Much of the argument against learning styles traces back to a 2008 paper by Pashler and colleagues. That paper has been cited approximately a bazillion times. But in 2024, Clinton-Lisell and Litzinger published a new meta-analysis that complicates things a bit. Guess what? They actually found an effect. The researchers analyzed 21 studies involving 1,712 participants and found a small but statistically significant benefit to matching instruction to learning styles (Hedges’ g = 0.31). Holy crap! Does that mean learning styles are “a thing?”
Not so fast. There are some pretty important caveats to this finding. Only 5 of the 21 studies actually met the authors’ quality standards. Further, most of these studies didn’t report reliability measures for their outcomes. Every single study only involved a single learning session, which says nothing about long-term retention. All of that aside, only 26% of the learning outcome measures demonstrated the kind of crossover interaction you’d need to support the matching hypothesis. So, despite finding a statistically significant effect, the authors concluded that the evidence wasn’t strong enough to recommend widespread use of learning-style matching.
Here’s why their findings are important: just because something might work doesn’t mean it’s worth doing. In the real world, we don’t have unlimited time and money to develop training. Matching instruction to learning styles means assessing each learner’s style, developing multiple versions of the same content, and delivering the right version to the right person. That’s a whole lot of work for a small and not particularly reliable improvement. Could we do it? Sure. But is it work the bang for the buck? Nope.