A Critical Analysis Of Learning Styles and Pedagogy In Post-16 Learning

A critical analysis of Learning Styles and Pedagogy in post-16 learning: A systematic and critical review by Coffield F., Moseley D., Hall E., & Ecclestone K.


In the adult education field there exist many educational and instructional theories to guide and help educators to provide appropriate education to their students’ needs, and within all the theories, one can find several referring to learning styles. This paper explores a joint analytical project about thousands of learning styles’ models, titled: A Critical Analysis of Learning Styles and Pedagogy in post-16 learning. A systematic and critical review published in 2004 by Coffield F., Moseley D., Hall E., & Ecclestone K.

Type of study

According to the authors, their investigation is systematic and critical, based on the analysis of  “three linked areas of activity: theoretical, pedagogical and commercial (Coffield et al., 2004),” within several learning styles’ models. The authors also state four additional purposes for the study to:

  1. Discard the weakest Papers,
  2. Summarize the large numbers of high-quality research papers,
  3. Simplify complex statistical arguments and,
  4. Impose some order on a field which is marked by debate and constructive critique as well as by disunity, dissension and conceptual confusion (2004).

The investigators express a positive preconception about the subject, although they indicate some dissatisfaction with the way some professionals of the field have threated the topic.

Stated purpose of the study

The aims of the investigation were “to carry out an extensive review of research on post-16 learning styles, to evaluate the main models of learning styles, and to discuss the implications of learning styles for post-16 learners (Coffield et al., 2004)”. The investigators were trying to catalog the best models in the field.

Study objectives

The following are the research objectives that Coffield et al., (2004) explicitly stated in the research.

  • identify the range of models that are:
    • available
    • influential or potentially influential in research and practice
  • locate these models within identifiable ‘families’ of ideas about learning styles
  • evaluate the theories, claims and applications of these models, with a particular focus on evaluating the authors’ claims for reliability and validity
  • evaluate the claims made for the pedagogical implications of the selected models of learning styles
  • identify what gaps there are in current knowledge and what future research is needed in this area
  • make recommendations and draw conclusions about the research field as a whole.

Study questions

The following are the questions postulated by Coffield et al., (2004) and answers we gave based on the review.

  1. What models of learning styles are influential and potentially influential?
    They found 3,800 papers and identified 13 potentially influential major key models.
  2. What empirical evidence is there to support the claims made for these models?
    They found little empirical evidence for those focused on an academic setting; while they found that commercial models tend to unwelcome “critical engagement with the theoretical and empirical bases of their claims (Coffield et al., 2004)”. They found that this happens due to the economical gains these models offer to their creators.
  3. What are the broad implications for pedagogy of these models?
    These models provide awareness about preferences of the learner’s own learning abilities and the opportunity to develop others, however, there is little empirical evidence about that.
  4. What empirical evidence is there that models of learning styles have an impact on students’ learning?
    In page 51, they revealed some confusion they found when creators of these models try to justify their investigations by linking them to other empirical research, not necessarily directly related or beneficial for their purposes. As Coffield et al., noted, this illustrates the confusion that can result from linking style labels with ‘brainedness’ in the absence of empirical evidence. The absence of hard evidence does not, however, prevent McCarthy from making " 'a commonsense decision to alternate right- and left-mode techniques’ (1990, 33) in each of the four quadrants of her learning cycle (2004)".

Study Methodology

The research followed a qualitative methodology. The researchers aimed to “carry out an extensive review of research on post-16 learning styles, to evaluate the main models of learning styles, and to discuss the implications of learning styles for post-16 teaching and learning (Coffield et al., 2004)”. They measured the data neither quantitatively, nor statistically, as a quantitative research would suggest. The investigators selected the models they thought were the best ones. Resembling the typical tendency of qualitative research, the investigation was subjective and led to establish which were the best studies and academics of the topic, suggesting a norm or a standard.

They also performed the following methodology associated with qualitative research.

  • Observation and analysis of the models’ description and structure.
  • In depth interviews with the researchers that created some of the analyzed models, as Coffield et al., stated, “an important part of our evaluation of each of the 13 models was to send the authors a copy of our report on their model and to ask for comment (2004)”.
  • Focus groups with interest groups, as Coffield et al., stated, “The project team also sought advice from a local advisory group whose members read out drafts reports from a mainly practitioner perspective (2004)”.

Were the measurements valid and reliable?

The procedures to select the literature, which was the sample of the research, were the following. The investigators reviewed 3800 general references related to the topic of the investigation. Afterwards, they “accumulated a database containing 838 references and papers related to the topic […] and logged them into their database (Coffield et al., 2004)”. Then, they reduced the literature to 631 texts and from there; they finally selected 351 texts directly referring to the 13 major models. The 3449 papers that were discarded were considered to be the “weakest papers (Coffield et al., 2004).

Could others replicate the research procedures if they wished?

According to the paper, “the team who carried out the research have combined expertise in cognitive psychology, education, professional development of post-16 practitioners, sociology and policy studies (Coffield et al., 2004)”. They also pursued advice from other academics and groups interested in the topic of discussion. Therefore, to replicate the procedures one would have to assemble a similar team of professionals and ask the same or comparable questions to the ones presented in the investigation.

Research Results

According to Coffield et al., the technique used in the research was the following: “Each model was examined for evidence, provided by independent researchers, that the instrument could demonstrate both internal consistency and test–retest reliability and construct and predictive validity. These are the minimum standards for any instrument, which is to be used to redesign pedagogy (2004)”.
In accordance to the qualitative methodology implemented in the research, there were no indications of statistical techniques. Researchers made their decisions based on how the models paired with the criterions (questions) established at the beginning of the study. In addition to that, the researchers were not completely detached from the subject, due to their involvement in the academia and the use of the same methods they were investigating.

Did clear questions emerge from the data that were collected?

At the conclusion of the research, Coffield et al., postulate the following question: Finally, we want to ask: why should politicians, policy-makers, senior managers and practitioners in post-16 learning concern themselves with learning styles, when the really big issues concern the large percentages of students within the sector who either drop out or end up without any qualifications? (2004).

Discussion of Results

The results of the research were the selection of 13 models, and according to Coffield et al., “one of the most obvious conclusions is the marked variability in quality among them; they are not alike nor of equal worth and it matters fundamentally which instrument is chosen (2004)”.
Out of the 13 recommended models, there is an A list of models that Coffield et al., considered and “appraised as promising”:

  • The Cognitive Style Index of Allinson and Hayes has the best psychometric credentials.
  • Apter reversal theory is a theory of personality, not of learning style. It was included because the concepts of motivation and reversal (e.g., change from work to play) are important for understanding learning styles.
  • Entwistle: his Approaches and Study Skills Inventory for Students (ASSIST) is useful as a sound basis for discussing effective and ineffective strategies for learning and for diagnosing students’ existing approaches, orientations and strategies.
  • Herrmann: his ‘whole brain’ model is suitable for use with learners as well as with teachers and managers, since it is intended to shed light on group dynamics as well as to encourage awareness and understanding of self and others.
  • Jackson: the Learning Styles Profiler (LSP) is a relatively new, but sophisticated, instrument, which has yet to be tested by independent researchers.
  • Vermunt: his Inventory of Learning Styles (ILS) can be safely used in higher education, both to assess approaches to learning reliably and validly, and to discuss with students changes in learning and teaching (2004).

Apart from the A list, the following are the accepted models with validated instruments that according to Coffield et al., “proved to be the most psychometrically sound and ecologically valid (2004)”.

  • Allinson and Hayes’ Cognitive Styles Index (CSI)
  • Apter’s Motivational Style Profile (MSP)
  • Dunn and Dunn model and instruments of learning styles
  • Entwistle’s Approaches and Study Skills Inventory for Students (ASSIST)
  • Gregorc’s Mind Styles Model and Style Delineator (GSD)
  • Herrmann’s Brain Dominance Instrument (HBDI)
  • Honey and Mumford’s Learning Styles Questionnaire (LSQ)
  • Jackson’s Learning Styles Profiler (LSP)
  • Kolb’s Learning Style Inventory (LSI)
  • Myers-Briggs Type Indicator (MBTI)
  • Riding’s Cognitive Styles Analysis (CSA)
  • Sternberg’s Thinking Styles Inventory (TSI)
  • Vermunt’s Inventory of Learning Styles (ILS)

Reasonable explanations of the findings

They concluded that some of the most used models have serious issues and weaknesses regarding “low reliability, poor validity and negligible impact on pedagogy (Coffield et al., 2004).” Based on those conclusions, they recommended discontinuing the use of those models.

Coffield et at. also found the following ongoing complications in the field of learning styles:

  • Theoretical incoherence and conceptual confusion
  • Learning styles in practice: labeling, vested interests and overblown claims
  • A tendency among some of the researchers ‘to rush prematurely into print and marketing with very early and preliminary indications of factor loadings based on one dataset (Curry 1990, 51, in Coffield et al., 2004)’.

Reasonable implications for practice of the findings

To conclude, the researchers recommend the use of their findings since they consider that those models would help to develop self-awareness and metacognition since pedagogy on its own is not enough. It is beneficial for educators who are in the field to have instruments scrutinized and analyzed by experts in order to obtain accurate tools for their field.


Coffield, F., Moseley, D., Hall, E., & Ecclestone, K. (2004). Learning styles and pedagogy in post-16 learning: a systematic and critical review. LSRC reference, Learning & Skills Research Centre, London. Retrieved on February 26, 2014