Learning styles in medical education: a scoping review

Objectives The literature on learning styles is conflicting. Systematic reviews over the past 10 years have found no benefit to their use in education, but positive mentions of learning styles continue to be prevalent in the literature, and the clinician looking to expand their understanding of learning and teaching can easily miss the few critical articles that exist. We aimed to find out exactly how learning styles are interpreted and discussed in the medical education literature. Methods We conducted a scoping review, using the Arksey and O’Malley framework, to "map" the literature; producing an overview of the research area and its key concepts, without restrictions about the type or quality of study included Results 176 studies were included in the final analysis, of which only 28 expressed any reservations about learning styles. Publications in 2015 were double that of 2008 and came from every continent of the world. 122 studies stated as fact that different students have different learning styles, and 67 studies, that teaching should be matched to learning style. In only approximately half of these was this assertion supported by reference to another source. This positive view was maintained in many cases despite researchers’ own findings to the contrary. Conclusions Learning styles persist in the medical education literature, despite a lack of evidence to support their use and are often discussed as a scientific fact. The volume of work that mentions learning styles as an accepted fact inspires further work, which then plays its part in reinforcing acceptance of the theory. This uses time and money that could more usefully be spent researching other topics. Sometimes learning styles are used as a reason to advocate for a move to a more varied, student-centred curriculum, which results in good outcomes for the students despite the questionable theoretical argument for the change. This scoping review presents a critical review of the use of Davies-Kabir M, Aitken G MedEdPublish https://doi.org/10.15694/mep.2021.000169.1 Page | 2 learning styles in the medical education literature, and questions, despite increasing evidence to the contrary, the near ubiquity of their acceptance.


Introduction
The evidence for different learning styles The term "learning styles" suggests different preferences and aptitudes for different study styles or modes of instruction. An extensive evaluation (Coffield et al., 2004) identified 71 different models of learning styles, categorizing 13 as major models. Learning styles research occurs in many fields: psychology, sociology, education, management and business studies. Some theories describe learning styles as a relatively fixed state: as a component of a stable personality type such as Myers-Briggs (see Murray, 1990), as features of the cognitive structure e.g. (Gardner, 2011) or as a constitutional preference for information input in different modalities e.g. (Gregorc, 1984;Dunn, 1990). Others view learning styles as more flexible learning preferences e.g. (Kolb, 1981;Felder, 1993;Honey and Mumford, 2006) and others look at learning approaches or strategies e.g. (Samarakoon, Fernando, Rodrigo and Rajapakse, 2013;Woolf et al., 2013) determined partly by student characteristics, but also by course content and delivery, workload and assessment demands (Newble and Entwistle, 1986). (Pashler et al., 2008) were commissioned by Psychological Science in the Public Interest to assess the scientific evidence for the practical application of learning style assessment in schools. In their literature review they found no high-quality evidence to validate the use of a learning-style assessment in educational settings. They did, however, find several high-quality studies that contradicted the hypothesis that instruction is best provided in a format that matches the learning style of the student.
Despite the resounding lack of evidence for the usefulness of learning styles in education, a recent analysis of the current research in higher education found that 89% of recent research papers implicitly or directly endorse the use of learning styles (Newton, 2015). The author describes learning styles as "the educational equivalent of homeopathy: a medical concept for which no evidence exists, yet in which belief and use persists."

Research questions
This review sought to answer the following questions. In the medical education literature: How are learning styles interpreted and discussed?
What assumptions are made about learning styles theory?
What are the motivations behind the use of learning styles?
How does the use of learning styles theory influence pedagogic thinking?

Stage 3 Study selection.
Eliminate studies that do not address research question. This can be an iterative process based on increasing familiarity with the literature.

Stage 4 Chart the data.
Use a data charting form to study specific information about each study.

Stage 5
Collate, summarize and report the results. A scoping study does not seek to assess quality of evidence so there is no attempt made to present a view regarding the weight of evidence. Use a narrative account including a basic numerical analysis of the extent, nature and distribution of studies, followed by a thematic review. Maintain clarity of reporting strategy so that the position and any potential bias of the reporting can be made clear.

Stage 6
Optional stage -Consult with stakeholders. Adapted by the authors from Arksey and O'Malley (2005) framework for a scoping review.

Identifying relevant studies
ERIC, Medline, Embase and the Research and Development Resource base were searched using the search terms "learning" near "styles" or "preferences" AND "medic" OR "doctor" OR "physician" OR "surgeon". Terms were combined using Boolean operators to ensure that no relevant terms were missed, modifying terms as necessary to suit the slightly different subject headings of the different databases.
Studies were excluded if they did not mention learning styles, or have either medical students or doctors as participants. The first phase of selection was based on title and abstract, or the full paper if no abstract was available. The second selection phase applied the inclusion and exclusion criteria to the full texts. Included studies were read in detail and a data capture form was used to extract relevant information. Thematic review was undertaken in three stages: at the study selection phase, the data charting phase, and a final review once all the data was charted. At each stage the codes were reviewed and new codes added as themes emerged from the data.

Results
The numbers of studies at each stage of the search process are detailed in Figure 1.  Only one study out of 850 had to be excluded on the basis of abstract and paper being unobtainable. All 201 studies selected for inclusion in phase one were obtained as full text articles; 5 of these were only available in Spanish and the translated abstract was used for phase two and data extraction.

How are learning styles interpreted and discussed?
The majority of the papers (82) were original experimental studies, 23 were opinion pieces and the remainder review articles or the use of learning styles in a particular context. The number of publications mentioning learning styles or preferences has increased from 14 in 2008 to 29 in 2015, with a general trend of increase for the years in between (see Figure 2).

Figure 2. Breakdown of included studies by year of publication
VARK/VAK was the learning style most commonly discussed (42 studies): VAK, (Barbe, Swassing and Milone, 1979) and VARK (Fleming and Mills, 1992) classify students by their preferred modality of learning: visual, auditory or kinaesthetic, and sometimes "read/write". Kolb's learning system (Kolb, 2014), which originated in his work on experiential learning theory, came second (33 studies). 16 studies referred to the tripartite model. Based on the original work of Marton and Säljö (Marton and Säljö, 1976), this describes three different approaches to learning: surface (student aims to learn the bare essentials and reproduce through rote learning), deep (meaningful learning integrated with previous knowledge), and achieving or strategic (student aims towards whichever goals will achieve the highest grades). 11 mentioned Honey and Mumford's (Honey and Mumford, 2006) theory, and 8 Felder's (Felder, 1993) index of learning styles.

Confusion of terminology
Learning styles, learning approaches, learning strategies and learning preferences are terms which do not seem to have universally accepted definitions. For example, (Mitchell et al., 2009) use the phrase learning styles and strategies repeatedly in their introduction but do not define the terms or explain the difference between styles and strategies. They go on to use a questionnaire that tests 'approaches' to learning, and report the results of this in a section entitled 'learning style'. In their discussion learning styles are described as "cognitive differences in the way in which individuals learn" and strategies as "methods individuals choose to use in their learning", but then the students approach to learning anatomy (study a small piece first or tackle a whole area) is discussed under the heading "Learning styles and cultures of learning in anatomy" (2009:56). In another anatomy paper (Mustafa et al., 2013) use the terms learning styles, strategies, approaches and methods seemingly interchangeably. What assumptions are made about learning styles theory?
Many of the studies contained the assertion that different students or populations have different learning styles, and that teaching style should be matched to learning style. Unfortunately, only about half of these felt the need to back up the statement with a reference to another source suggesting that learning styles have become part of the tacit knowledge of medical education, in taken for granted and largely unquestioned ways.
In only 16% of studies did authors express any reservations about learning styles. This overwhelming acceptance of learning styles in the medical education literature is very similar to that found in the higher education literature in a recent review (Newton, 2015).
Several authors (Adesunloye et al., 2008;Rubinstein, Dhoble and Ferenchick, 2009;Wilkinson and Boohan, 2012;Al-Qahtani, 2015;Liew et al., 2015) found nothing significant in the results of their learning styles experiments, but nonetheless went on to extol the usefulness of learning styles theory and recommend further research in the area.
Some authors maintained a fixed view of learning styles or of matching theory that did not follow, or went beyond, the findings of the study. (Ankad et al., 2015) tested VARK learning styles in medical students and found no significant difference between learning style in different demographic groups, and no effect of learning style on outcomes, but went on to conclude that knowledge of a particular students learning style would help teachers to develop more effective teaching strategies. (McGee, 2015) found that neurology residents preference for different styles of learning depended on topic, but concluded that the results suggested the need to assess individual residents learning style to ensure the most efficacious learning. More alarmingly, (Isiglu-Alkac and Gulpinar, 2010) investigated cognitive responses of medical students by using Event Related Brain Potentials (ERPs) measured by EEG. They showed that ERPs were stronger in specific regions for specific tasks, and they suggest that with further development this technique could be used to determine the learning styles of medical students.
A small number of papers were critical of learning styles. Some supported the idea of learning styles but disputed matching theory (Vollers, 2008;Bhagat, 2013;Feeley and Biggerstaff, 2015;Roberts et al., 2015) and some were critical of the whole field (Norman, 2009;Royal and Stockdale, 2015), with (Newton, 2015) describing learning styles as a "neuromyth" and suggesting that their persistence could be due to confirmation bias as educators "might reasonably approach the literature with an expectation that learning styles are a useful tool," a view likely to be confirmed by a casual search of the literature. (Royal and Stockdale, 2015) raise concerns about the impact of learning styles on educator's workload with the potential for teachers to feel pressure to alter successful approaches to include a consideration of learning styles in their teaching design. They attribute their ongoing popularity to commercial interests, the large amount of positive literature, and the intuitive face validity of matching theory.

Coping with change
A changing clinical educational environment is mentioned in many of the papers. The president of the Triological society, Jesus Medina, in an editorial in The Laryngoscope (Medina, 2013), discusses the speed of change in the educational environment, attributing this to working time restrictions, additional demands on trainees (domestic and childcare responsibilities), and the expansion of knowledge and technology. In an editorial in Colorectal Disease, Engel (Engel, 2014) writes: "Surgical educators are grappling with resident working hour limits, changing social demands and merging partner roles, to name but a few game changers…Time is limited and information seems to evolve at lightning speed".
Concerns over changes in the clinical working environment may also have inadvertently contributed to the popularity of learning styles in the literature. Jack et al. (2010) describe how the current surgical teaching environment in the USA has been redefined by the influence of the 80-hour work week, and report a number of studies that show that the 80-hour work week has had a negative impact on medical education, going on to suggest pairing teachers and learners based on preferred learning styles. A conference abstract in The Heart Surgery Forum (Anon., 2010) describes the European working time directive as one of the "huge threats for the future of our profession" and suggests designing a curriculum that considers individual student learning styles as a response to this. Medina (2013) and Engel (2014) suggest adapting teaching for different learning styles. Several other authors also mention learning styles as helpful in managing changes to surgical teaching caused by recent legal restriction of work hours (Habib et al., 2012;Kim et al., 2013;Richard et al., 2014;Quillin et al., 2016). In fact, of the 18 studies written by general, orthopaedic or ENT surgeons, nine (50%) mentioned anxiety about the challenges to learning caused by a change in professional culture and legal restriction on hours.
Some studies focus on the challenges of teaching a new generation with potentially different learning styles (Moreno-Walton, Brunett and Akhtar, 2009;Kurup, 2013). Lehmann et al. (2015) posit that the generation gap between students and faculty may hinder learning, controversially claiming that millennial medical students are very different to preceding generations.
Sarmah et al. (Sarmah, Ryzynski and Sousa, 2012) see the current challenge of medical education as being to meet the needs of generation Y students with "low attention spans, their learning styles are visual or kinaesthetic with a preference for multimedia technology". And Chu et al. (2012) state that "Millennial learners have uniquely different learning styles than previous generations", explaining that this is because their neuroanatomy has been affected by exposure to digital input while growing.

Arguing for change
Many authors uncritically accept learning styles as fact, but then use their results to argue for a more creative approach to curriculum design. Bagi, Raichur and Mantur, (2012) call for the instigation of a wide variety of teaching methods in their medical school, based only on their finding that students who score highly for "read/write" on a VARK test have slightly better outcomes in final exams, and Prithishkumar and Michael (2014) use the finding that 82% of first year medical students are VARK multimodal to argue that "multiple modalities of information presentation are necessary to keep the attention and motivation of our students". Anbarasi et al. (2015) developed some innovative, exciting and very creative teaching methods to provide VARK tailored physiology teaching to students, with the control group receiving lectures with no pictures. While one may question the equivalence of experience of both groups, the outcome was a plan to introduce the new teaching methods to all students, so despite the methodological flaws the outcome was positive a curriculum innovation for all students. Bhalli, Khan and Sattar, (2015) tested the learning styles of 99 medical students. They found no significant correlation between learning style and preferred teaching style or academic outcomes. They concluded: "Aligning our instructional strategies with learning styles of the medical students will improve learning and academic performance." This confusing non sequitur can perhaps be explained by looking at the other results reported: "Majority preferred interactive lectures (28.57%), problem-based learning (25.98%) and small group discussion (20.78%) as their teaching methodologies while one-way lectures, student presentations and guest speakers were preferred least." The recommendation to consider "the diversity of learning styles while developing the curriculum" could therefore be interpreted as a call for a move towards a more active student focused style of learning.
How does the use of learning styles theory influence pedagogic thinking? Only three of the 176 studies showed evidence supportive of a change in teaching practice. All three concerned the tripartite theory of learning approaches: A study of attitudes to e-learning (Svirko and Mellanby, 2008): maximizing the use of clinically relevant 1.
material appeared to increase the interest and enjoyableness of the course and thereby aid deep learning and information retention. A quasi-experimental study (Kegels et al., 2008) in primary care: the intervention group experienced practice-2.
based learning in small groups, with a reported shift from superficial to deep learning. A literature review (Feeley and Biggerstaff, 2015) of VARK and the tripartite model concluded that medical 3.
school exam success had no relation to learning styles, but students' learning approaches were important. Students who developed a "growth mind-set" (viewing academic success as dependent on their own efforts, rather than on any intrinsic intelligence or ability) showed a positive and resilient approach to learning and a reduced fear of failure.

Conclusions
These findings echo the lamentations of other authors. Elaine Hall was one of the authors on a seminal paper (Coffield et al., 2004) which found no evidence for the use of learning styles in post-16 education. 12 years later, having been asked to write about learning styles again, she wrote: "I'm sure many academics have days when they feel their life resembles a horror movie, it's just that in relation to Learning Styles I am one of those weary, gore-spattered actors. No matter what we've hit it with, the thing won't die." (Hall, 2016) So why do learning styles theories persist? Some authors have suggested it is because of the commercial interests at play (Rohrer and Pashler, 2012). We did not find evidence for this in the medical education literature: and found only one study (Rubinstein, Dhoble and Ferenchick, 2009), where the authors tested a puzzle which they had patented, appeared to be complicated by potential commercial gain.
We found that some authors see the diagnosis of individual learning styles as a way to manage their anxiety about change. Some feel out of touch with the generation coming into their profession, or are worried about the impact of shorter working hours. Doctors are acculturated into the scientific method. They are used to being able to use scientifically tested protocols to manage uncertainty, and they perhaps want to be able to measure and target their teaching in the same way.
Some authors see learning styles as a way to argue for a change to traditional educational practices. They use data on learning styles found to argue for change: whether reduced class sizes, constructive alignment of the curriculum, an increase in interactive and small group teaching, a decrease in large expert led didactic lectures, an innovation in teaching methods, or more consideration given to how to motivate and interest students. They perhaps realise that they have a variety of alternative curriculum design options available to them and when making sense of these options are beguiled by a seemingly scientific way to decide (and demonstrate) which teaching styles will suit their students.

Recommendations
Learning styles persist in the medical education literature, despite a lack of evidence to support their use and are often discussed as a scientific fact. The volume of work that mentions learning styles as an accepted fact inspires further work, which then plays its part in reinforcing acceptance of the theory. This uses time and money that could more usefully be spent researching other topics.
Unlike learning styles, learning approaches are worthy of consideration by medical educators, and the evidence suggests that it may be possible to design curricula that encourage deep learning and so improve understanding and knowledge retention.
There is no evidence to support the consideration of other learning styles in educational practice, and medical educators should be aware of this. In planning their teaching educators should think about using whichever methods will help their students understand and engage deeply with the material they are trying to communicate. Rather than concentrating on diagnosing and accommodating the learning styles of their students, time will be better spent developing teaching resources to facilitate learning.

Take Home Messages
Learning styles continue to be commonly used in medical education There is no evidence base supporting this The notion that learning styles are fixed and measurable obsures the complexity of the learning process We encourge a more critical approach when considering student learning

Meg Davies-Kabir
Meg is an ST4 in child and adolescent psychiatry and a graduate of the MSc Clinical Education programme at the University of Edinburgh, UK.

Gillian Aitken
Gill Aitken is the Programme Director of the MSc Clinical Education programme at the University of Edinburgh, UK. ORCID: http://orcid.org/0000-0002-5492-1943