Learning styles

Learning styles has become a very fashionable trend in education, with different systems claiming to classify students into different types of learning style categories. Many have claimed that teachers can and should adapt to their students’ learning styles. However, there are significant problems with learning styles theories and popular LS applications. Many learning styles classifications exist, and no agreement on which system teachers should follow, or even on how to scientifically define learning styles for such purposes. While students may have certain learning preferences, there is no clear scientific evidence that students’ learning styles can be accurately assessed, or that such learning styles affect students’ performance. Also, it is not clear how teachers could adapt their teaching to students’ learning styles, and that teaching to particular learning styles affects students’ learning outcomes. Finally, no solid scientific evidence exists that teaching according to learning styles affects students' learning outcomes, as discussed below.

1 Overview of learning styles classifications

Much has been written about differences in individual learning styles and individual differences in learning. This has become a popular topic among educators, with many different classification systems, published studies, popular books, popular articles, and diagnostic quizzes available for teachers. This has become a veritable cottage industry, and in fact, there are over 70 different theories of learning styles (Coffield et al., 2004), including theories like MI that were originally developed for other purposes but are commonly used as learning style models. You may have heard of these or others.

1. Multiple intelligences (MI). Educational psychologist Howard Gardner has successfully popularized his MI theory among educators, according to which, students are talented in one or more specific types of intelligence – logical-mathematical, spatial, linguistic, kinesthetic, musical, interpersonal, intrapersonal, naturalistic, and existential. However, solid scientific evidence or diagnostic methods are lacking (how does one develop a feasible IQ test for eight or more intelligences?), and it can difficult, if not unreasonable, for teachers to develop different activities to accommodate students’ supposedly different styles.
2. Hierarchic intelligences. Psychologist Robert Sternberg proposed a model of three types of intelligence (analytical, creative, and practical intelligences), each of which in turn consists of several subtypes. The model is complex, is difficult to test, lacks scientific evidence, and is difficult for educators to practically apply.
3. Kolb’s learning styles. This model and other similar models posit several different cognitive styles in learning. Learners are classified along two dimensions: (1) a preference for concrete experience versus abstract conceptualization, and (2) a preference for reflective observation versus active experimentation. These then combine to form four possible learning styles: (1) Converger, (2) Diverger, (3) Assimilator, and (4) Accommodator. While the two dimensions seem sensible, as classifying persons according to whether they like concrete experience or abstract thinking, or observing versus experimenting, seems clear and useful. A famous test known as the Learning Style Inventory (LSI) was developed based on the Kolb system. However, the four categories of learning styles seem somewhat unclear and difficult to define. Others have created modified versions of the Kolb inventory, e.g., the Honey-Mumford model.
4. Auditory cf. visual learning styles. This seems very intuitive – some of us feel that we learn better with visual information (reading, visual aids, real objects), and some of us feel that we learn better from auditory information (listening to lectures or such). Another common approach is a three-way distinction between auditory, visual, and tactile / kinesthetic (namely, the Fleming VARK model). This models is partly based on a pop psychology fad known as neuro-linguistic programming (NLP), which lacks scientific credibility and is not accepted by cognitive scientists, who in fact regard it as a pseudo-science.
5. Information perception and processing models. These models classify learners according to how they take in information (concrete versus abstract) and how they process information (holistically or intuitively, versus analytically and sequentially), leading to a four-way distinction such as (1) concrete sequential, (2) concrete random, (3) abstract sequential, and (4)] concrete sequential in the Gregor-Butler model. The Felder-Silverman model posits a multi-way distinction of active vs. reflective, sensing vs. intuitive, visual vs. verbal, and sequential vs. global learning styles.
6. MBTI (Myers-Briggs Personality Type Indicator). This is a well known psychological inventory. It was designed as a measure of personality and cognitive preferences, and was never intended to diagnose or categorize cognitive abilities, work-related abilities, learning abilities, or learning styles. It is somewhat controversial in psychology due to difficulties in scientifically defining and validating the constructs that comprise the MBTI typology, but it can be useful for personal self-understanding of one's cognitive preferences.
7. Left brain / right brain pop psychology. This is largely pop psychology, and is thus largely untrue. While there are some differences in cognitive process between the left and right brain hemispheres, the differences are complex and nuanced. The claims that very general abilities (language, art, math, music) are assigned to one hemisphere only, and the claims that people can be classified as left or right brained, are completely untrue, if not nonsensical. For more, see the article on brain hemisphericity.

2 Research findings

However, research psychologists and educational psychologists have sounded important warnings about this whole area of work. Many published studies have claimed to show evidence for different learning styles. However, more recent meta-analyses or reviews of such studies by psychologists have shown serious shortcomings in these studies, in that they lack proper scientific controls or the kinds of experimental designs that would provide solid scientific support for the supposed findings; i.e., all these experiments were lacking or poorly designed, and cannot reliably demonstrate the effects of learning styles (Hargreaves et al., 2005; Reynolds, 1997; Pashler et al., 2009; Curry, 1990). Other problems exist with studies claiming to provide empirical validation of learning style inventories (e.g, Freedman & Stumpf, 1978).

Many of the purported studies on learning styles have not been published in peer reviewed journals, which is of course problematic. Of those that have appeared in reviewed journals, most have not been controlled, empirical studies. The very few controlled scientific studies that have been conducted have found no effect of learning styles on students’ learning outcomes. Maybe some of these are valid, but the scientific evidence so far is lacking. The various studies claiming evidence for learning styles also do not consider important variables, such as the learning context, types of contents being learned, and learner-specific variables that need to be controlled for.

Even the seemingly intuitive and likely distinction between auditory and visual learners is suspect. One study (Kratzig & Arbuthnott, 2006) found no correlation between the VARK styles (auditory, visual or kinesthetic) with memory and learning performance on auditory, visual or kinesthetic versions of a learning task. The reliable studies on this do not show strong learning effects for these as learning styles, but rather, the visual-auditory distinction is more of an individual learning preference. Clearly, we tend to prefer one or the other, but this preference may not necessarily translate to or correlate with superior learning by one means or the other. Perhaps such an effect exists, but again, currently the evidence is weak or controversial; and it it exists, it might depend on the type of learning task and context, but carefully controlled studies on this are lacking. What studies do show, however, is that people in general – those with auditory or visual preferences – learn better when speakers or lecturers use both of these sensory “channels”. That is, when a speaker or teacher enhances his/her speech with visual aids, listeners or students learn better – whether they consider themselves to be visual or auditory learners (e.g., Constantinidou & Baker, 2002).

3 Applicability

The various LS inventories ask students rather general questions, e.g., whether one likes to observe and then try something, or immediately and directly work on something to figure it out. Such context-free questions are problematic, because learners have a repertoire of learning strategies that they would bring to bear on different situations. These studies and inventories do not take into account the particular learning context, the type of learning task, task complexity, and other characteristics of the learning situation, nor do they consider individual variables beyond learning styles that might affect learning outcomes. So if we are to believe in any kind of learning styles theory, it would require controlled scientific experiments that account for particular learning contexts and other variables.

It would be quite burdensome and unfeasible for teachers to accommodate multiple learning styles in a single classroom, and no properly designed scientific studies have found any learning benefits from teachers attempting to do so. Some systems classify students according to several cognitive dimensions (e.g., sequential, holistic, or analytical modes of information processing), others classify students according to preferred sensory modality (e.g., auditory, visual, tactile / kinesthetic), and others classify students according to other criteria. It would be impossible for teachers to try to differentiate their teaching according to four, eight or twelve learning styles. Again, no evidence from proper scientific studies exists that show that such teaching adaptations have any effect.

4 See also

For more on pop psychology myths, one may refer to the well written and readable book by Lilienfeld et al. (2010).

• Lilienfeld, S. O., Jynn, S. J. Ruscio, J., & Beyerstein, B. L. (2010). 50 Great Myths of Popular Psychology. West Sussex, UK: Wiley Blackwell.

4.1 Published review & research articles

1. Dembo, Myron H. & Howard, Keith. (2007). Advice about the Use of Learning Styles: A Major Myth in Education. Journal Of College Reading and Learning, 37, 2, 101-109.
2. Hall, E. & Moseley, D. (2005). Is there a role for learning styles in personalised education and training? International Journal of Lifelong Education, 24,(3), 243-255.
3. Pashler, H., McDaniel, M., Rohrer, D. & Bjork, R. (2008). Learning Styles: Concepts and Evidence. Psychological Science in the Public Interest, 9, 105-119. doi:10.1111/j.1539-6053.2009.01038.x
4. Rayner, S. (2007). A teaching elixir, learning chimera or just fool’s gold? Do learning styles matter? Support for Learning, 22(1), 24-30.
5. Scott, C. (2010). The enduring appeal of ‘learning styles.’ Australian Journal of Education 2010 54(5). DOI: 10.1177/000494411005400102 Online at http://aed.sagepub.com/content/54/1/5.