Proponents of teaching to students’ “learning styles” argue that people learn better when instructional materials are presented in a format that meshes with their learning style. This is called the “matching hypothesis,” or the “meshing hypothesis.” Unfortunately for its supporters, and despite many attempts to prove its validity, it’s not supported by evidence-based research.
The eLearning Guild’s director of research, Jane Bozarth, conducted a thorough review of research on learning styles and found, overwhelmingly, that attempts to match instruction style to learners’ preferred approach failed to improve outcomes. “The evidence needed to support the matching hypothesis is just not there,” she wrote.
It’s not hard to understand why the matching hypothesis is attractive: “Frankly, the idea of matching instruction to preferences is simple. It seems intuitive. It makes it easy to categorize people and ideas, and it’s easy to create materials that give the impression of some kind of customization,” Bozarth wrote. “Finally, unlike much talk in academic fields, it’s easily accessible to laypeople.”
Despite the appeal, there are better ways to drive learner outcomes. “In the aggregate, researchers have recommended that time and energy would be better spent matching instructional approach to content and type of material being taught rather than to any perceived individual preference or ‘style,’ ” Bozarth wrote.
Testing the matching hypothesis
Harold Pashler et al. published a comprehensive attempt to determine whether the matching hypothesis was valid. In it, they argue that for the hypothesis to be valid, they should be able to find evidence that students given instruction according to their “learning styles” must not only achieve better results, but that instructional approach must also produce worse results with students whose learning styles differ. “In other words, the instructional method that proves most effective for students with one learning style is not the most effective method for students with a different learning style,” the authors wrote.
Laura Massa and Richard Mayer tested the hypothesis as well, using visual and verbal “learning styles” in a test of eLearning materials. They found no difference in performance when test subjects received materials in their preferred format, when compared with learners who received materials in the opposite format (that is, visual learners given verbal content performed no worse than visual learners given visual content, and vice versa). In fact, in the particular lesson studied, all learners benefited more from a visual presentation of the learning materials, regardless of their learning preference.
These results support what many excellent teachers already know: Tailoring the instructional approach to the material produces the best results for all students, regardless of those students’ stated learning preferences.
In addition, Pashler et al. said that differences in “educational backgrounds” are a factor in tailoring instruction. “New learning builds on old learning, for example, so an individual student’s prior knowledge is bound to determine what level and type of instructional activities are optimal for this student.”
What instructors can glean from this is that providing scaffolding to guide learners from basic concepts to more advanced information is essential. Personalizing lessons to the extent that learners can skip material they know or access additional materials if they are new to the topic is ideal. This level of personalization is increasingly feasible in many forms of eLearning. Emerging technology is increasingly facilitating adaptive eLearning as well; this uses an AI-based engine to target content to learners according to their performance, reviewing weak areas and de-emphasizing material the learners have mastered.