FrankCoffieldLearning Styles

Embed Size (px)

Citation preview

  • 8/8/2019 FrankCoffieldLearning Styles

    1/182

  • 8/8/2019 FrankCoffieldLearning Styles

    2/182

  • 8/8/2019 FrankCoffieldLearning Styles

    3/182

  • 8/8/2019 FrankCoffieldLearning Styles

    4/182

  • 8/8/2019 FrankCoffieldLearning Styles

    5/182

  • 8/8/2019 FrankCoffieldLearning Styles

    6/182

  • 8/8/2019 FrankCoffieldLearning Styles

    7/182

  • 8/8/2019 FrankCoffieldLearning Styles

    8/182

  • 8/8/2019 FrankCoffieldLearning Styles

    9/182

  • 8/8/2019 FrankCoffieldLearning Styles

    10/182

  • 8/8/2019 FrankCoffieldLearning Styles

    11/182

  • 8/8/2019 FrankCoffieldLearning Styles

    12/182

  • 8/8/2019 FrankCoffieldLearning Styles

    13/182

  • 8/8/2019 FrankCoffieldLearning Styles

    14/182

  • 8/8/2019 FrankCoffieldLearning Styles

    15/182

  • 8/8/2019 FrankCoffieldLearning Styles

    16/182

  • 8/8/2019 FrankCoffieldLearning Styles

    17/182

  • 8/8/2019 FrankCoffieldLearning Styles

    18/182

  • 8/8/2019 FrankCoffieldLearning Styles

    19/182

  • 8/8/2019 FrankCoffieldLearning Styles

    20/182

  • 8/8/2019 FrankCoffieldLearning Styles

    21/182

  • 8/8/2019 FrankCoffieldLearning Styles

    22/182

  • 8/8/2019 FrankCoffieldLearning Styles

    23/182

  • 8/8/2019 FrankCoffieldLearning Styles

    24/182

  • 8/8/2019 FrankCoffieldLearning Styles

    25/182

  • 8/8/2019 FrankCoffieldLearning Styles

    26/182

  • 8/8/2019 FrankCoffieldLearning Styles

    27/182

  • 8/8/2019 FrankCoffieldLearning Styles

    28/182

  • 8/8/2019 FrankCoffieldLearning Styles

    29/182

  • 8/8/2019 FrankCoffieldLearning Styles

    30/182

  • 8/8/2019 FrankCoffieldLearning Styles

    31/182

  • 8/8/2019 FrankCoffieldLearning Styles

    32/182

  • 8/8/2019 FrankCoffieldLearning Styles

    33/182

  • 8/8/2019 FrankCoffieldLearning Styles

    34/182

  • 8/8/2019 FrankCoffieldLearning Styles

    35/182

  • 8/8/2019 FrankCoffieldLearning Styles

    36/182

  • 8/8/2019 FrankCoffieldLearning Styles

    37/182

  • 8/8/2019 FrankCoffieldLearning Styles

    38/182

  • 8/8/2019 FrankCoffieldLearning Styles

    39/182

  • 8/8/2019 FrankCoffieldLearning Styles

    40/182

  • 8/8/2019 FrankCoffieldLearning Styles

    41/182

  • 8/8/2019 FrankCoffieldLearning Styles

    42/182

  • 8/8/2019 FrankCoffieldLearning Styles

    43/182

  • 8/8/2019 FrankCoffieldLearning Styles

    44/182

  • 8/8/2019 FrankCoffieldLearning Styles

    45/182

  • 8/8/2019 FrankCoffieldLearning Styles

    46/182

  • 8/8/2019 FrankCoffieldLearning Styles

    47/182

  • 8/8/2019 FrankCoffieldLearning Styles

    48/182

  • 8/8/2019 FrankCoffieldLearning Styles

    49/182

  • 8/8/2019 FrankCoffieldLearning Styles

    50/182

  • 8/8/2019 FrankCoffieldLearning Styles

    51/182

  • 8/8/2019 FrankCoffieldLearning Styles

    52/182

  • 8/8/2019 FrankCoffieldLearning Styles

    53/182

  • 8/8/2019 FrankCoffieldLearning Styles

    54/182

  • 8/8/2019 FrankCoffieldLearning Styles

    55/182

  • 8/8/2019 FrankCoffieldLearning Styles

    56/182

  • 8/8/2019 FrankCoffieldLearning Styles

    57/182

  • 8/8/2019 FrankCoffieldLearning Styles

    58/182

  • 8/8/2019 FrankCoffieldLearning Styles

    59/182

  • 8/8/2019 FrankCoffieldLearning Styles

    60/182

  • 8/8/2019 FrankCoffieldLearning Styles

    61/182

  • 8/8/2019 FrankCoffieldLearning Styles

    62/182

  • 8/8/2019 FrankCoffieldLearning Styles

    63/182

  • 8/8/2019 FrankCoffieldLearning Styles

    64/182

  • 8/8/2019 FrankCoffieldLearning Styles

    65/182

  • 8/8/2019 FrankCoffieldLearning Styles

    66/182

  • 8/8/2019 FrankCoffieldLearning Styles

    67/182

  • 8/8/2019 FrankCoffieldLearning Styles

    68/182

  • 8/8/2019 FrankCoffieldLearning Styles

    69/182

  • 8/8/2019 FrankCoffieldLearning Styles

    70/182

  • 8/8/2019 FrankCoffieldLearning Styles

    71/182

  • 8/8/2019 FrankCoffieldLearning Styles

    72/182

  • 8/8/2019 FrankCoffieldLearning Styles

    73/182

  • 8/8/2019 FrankCoffieldLearning Styles

    74/182

  • 8/8/2019 FrankCoffieldLearning Styles

    75/182

  • 8/8/2019 FrankCoffieldLearning Styles

    76/182

  • 8/8/2019 FrankCoffieldLearning Styles

    77/182

  • 8/8/2019 FrankCoffieldLearning Styles

    78/182

  • 8/8/2019 FrankCoffieldLearning Styles

    79/182

  • 8/8/2019 FrankCoffieldLearning Styles

    80/182

  • 8/8/2019 FrankCoffieldLearning Styles

    81/182

  • 8/8/2019 FrankCoffieldLearning Styles

    82/182

  • 8/8/2019 FrankCoffieldLearning Styles

    83/182

  • 8/8/2019 FrankCoffieldLearning Styles

    84/182

  • 8/8/2019 FrankCoffieldLearning Styles

    85/182

  • 8/8/2019 FrankCoffieldLearning Styles

    86/182

  • 8/8/2019 FrankCoffieldLearning Styles

    87/182

  • 8/8/2019 FrankCoffieldLearning Styles

    88/182

  • 8/8/2019 FrankCoffieldLearning Styles

    89/182

  • 8/8/2019 FrankCoffieldLearning Styles

    90/182

  • 8/8/2019 FrankCoffieldLearning Styles

    91/182

  • 8/8/2019 FrankCoffieldLearning Styles

    92/182

  • 8/8/2019 FrankCoffieldLearning Styles

    93/182

  • 8/8/2019 FrankCoffieldLearning Styles

    94/182

  • 8/8/2019 FrankCoffieldLearning Styles

    95/182

  • 8/8/2019 FrankCoffieldLearning Styles

    96/182

  • 8/8/2019 FrankCoffieldLearning Styles

    97/182

  • 8/8/2019 FrankCoffieldLearning Styles

    98/182

  • 8/8/2019 FrankCoffieldLearning Styles

    99/182

  • 8/8/2019 FrankCoffieldLearning Styles

    100/182

  • 8/8/2019 FrankCoffieldLearning Styles

    101/182

  • 8/8/2019 FrankCoffieldLearning Styles

    102/182

  • 8/8/2019 FrankCoffieldLearning Styles

    103/182

  • 8/8/2019 FrankCoffieldLearning Styles

    104/182

  • 8/8/2019 FrankCoffieldLearning Styles

    105/182

  • 8/8/2019 FrankCoffieldLearning Styles

    106/182

  • 8/8/2019 FrankCoffieldLearning Styles

    107/182

  • 8/8/2019 FrankCoffieldLearning Styles

    108/182

  • 8/8/2019 FrankCoffieldLearning Styles

    109/182

  • 8/8/2019 FrankCoffieldLearning Styles

    110/182

    Conclusions

    Entwist le and his colleagues have spent almost30 years refining the validity and reliability of theirinventories to arri ve at it ems that have reasonable

    predict ive validity. They acknowledge the tendencyfor detailed, continuous ref inements to make technicalconstructs less credible and less easy to use byresearchers outside educat ional psychology. They havetherefore supplemented their analysis of approachesto learning with data from qualitat ive studies to explorethe consistency and variability of learning approacheswithin specific contexts (see McCune and Entwistle2000; Entwistle and Walker 2000). In this respect,their methodology and the data their studies haveproduced offer a rich, authentic account of learningin higher education.

    However, one feature of a posi tivist methodology,

    which aims for precise measures of psychometrictraits, is that i tems proliferate in order to try to capturethe nuances of approaches to learning. There are otherlimitat ions to quanti tat ive measures of approachesto learning. For example, apparent ly robustclassifications of meaning and reproductionorientations in a questionnaire are shown to be lessvalid when interviews are used with the same students.Richardson (1997) argued that interviews by Martonand Slj show deep and surface approaches asdifferent categories or forms of understanding, or asa single bipolar dimension along which individualsmay vary. In contrast, questionnaires operationalisethese approaches as separate scales that turn outto be essentially orthogonal to each other; a s tudentmay therefore score high or low on both. Accordingto Richardson, this difference highlights the need forresearchers to differentiate between methods t hataim to reveal average and general disposit ions withina group and those that aim to explain the subtletyof individuals actions and motives.

    Despite attempts to ref lect t he complexityof environmental factors affecting students approachesto learning and studying, the model does not discussthe impact of broader factors such as class, raceand gender. Although t he model takes some accountof intensifying polit ical and insti tut ional pressuresin higher education, such as quality assuranceand funding, sociological inf luences on participationand attitudes to learning are not encompassedby Entwistles model.

    There is also confusion over the theoretical basisfor const ructs in the ASI and ASSIST and subsequentinterpretation of t hem in external evaluations.Two contrasting research traditions create theseconstructs: information processing in cognitivepsychology; and qualitat ive interpretat ion of s tudentsapproaches to learning. Outside the work of Entwistleand his colleagues, a proliferation of instrumentsand scales, based on the original measure (the ASI),has led to the merging of constructs from both researchtradit ions. Unless there is discussion of the originaltradit ions from which the constructs came, t he resultis a growing lack of theoretical clarit y in the fieldas a whole (Biggs 1993). Entwistle and his colleagueshave themselves warned of this problem and providedan overview of the conceptions of learning, theirhistory within the approaches to learning modeland how different inventories such as those of

    Entwistle and Vermunt relat e to each other (Entwist leand McCune 2003).

    There are a number of st rengths in Entwistles work.For example, he has shown that ecological validit y isessential to prevent a tendency to label and stereotypestudents when psychological t heory is t ranslated intothe pract ice of non-specialists. The issue of ecologicalvalidity illuminates an important point for our reviewas a whole, namely that the expert ise and knowledgeof non-specialists are both context-specific andidiosyncratic and this affects t heir ability to evaluateclaims and ideas about a part icular model of learningstyles. High ecological validit y makes a model

    or instrument much more accessible to non-specialists.Entwistles work has also aimed to simplify the diverseand sometimes contradictory factors in studentsapproaches to studying and learning, and to offera theoretical rat ionale for them. He has attemptedto reconcile ideas about the stability of learningstyles with t he idea that approaches are idiosyncraticand fluctuating and affected by complex learningenvironments. His work highlights the need forresearchers to relate analysis and t heoreticalconstructs to the everyday experience of teachersand students, and to make their constructs accessible(see also Laurillard 1979).

  • 8/8/2019 FrankCoffieldLearning Styles

    111/182

    page 102/ 103LSRC reference Sect ion 7

    Table 3 4Entwistles Approachesand Study SkillsInventory for Students(ASSIST)

    General

    Design of the model

    Reliability

    Validity

    Implicationsfor pedagogy

    Eviden ce ofpedagogical impact

    Overall assessment

    Key source

    WeaknessesComplexity of the developing modeland instruments is not easy fornon-specialists to access.

    There are dangers if the modelis used by teachers without in-depthunderstanding of its underlyingimplications.

    Many of t he sub-scales are l ess reliable.Testretest reliabilit y not shown.

    Construct and predictive validit y havebeen challenged by external s tudies.Unquestioned preference for deepapproaches, but strategic and evensurface approaches may be effectivein some contexts.Rather weak relat ionships betweenapproaches and att ainment.

    The scope for manoeuvre incourse design is variable outsidethe relat ive autonomy of highereducation, especially in relationto assessment regimes.There is a large gap between usingthe instrument and transforming thepedagogic environment.As t he terms deep and sur facebecome popular, they become attachedto individuals rather than behaviours,against the authors intention.

    Not tested directl y as a basisfor pedagogical interventions.

    StrengthsModel aims t o encompass approachesto learning, study strategies, intellectualdevelopment skills and atti tudes inhigher education.

    Assesses study/ learning orientati ons,approaches to study and preferencesfor course organisation and instruction.

    Internal and external evaluationssuggest satisfactory reliability andinternal consistency.

    Extensive test ing by authorsof construct validity.Validit y of deep, surface andstrategic approaches confirmedby external analysis.

    Teachers and learners can share ideasabout effective and ineffectivestrategies for learning.Course teams and managers can useapproaches as a basis for redesigninginstruction and assessment.Model can inform the redesign

    of learning milieux within departmentsand courses.

    Has been influential in trainingcourses and staff development inBritish universities.

    Potentially useful model and inst rument for some post-16 contexts outs ide thesuccess it has had in higher education, but signif icant development and testingwill be needed.

    Entwistle 1998

    These features and the high output of work byEntwistle and his colleagues have made it crediblewith practit ioners and staff developers withinUK higher education. It has provided a model of learningwith which academics who wish to be good teacherscan engage: this is absent in teacher training for thefurther and adult education sectors, and for work-basedtrainers, where there is no influential t heory of learningthat could improve professional understanding andskills. Nevertheless, it is perhaps worth reiteratingHaggiss warning (2003) that the model runs the riskof becoming a rigid framework that excludes socialmodels of learning.

    Finally, al though Entwistle and his colleagues arguethat researchers need to build up case studiesby observing students studying and interviewing themabout their approaches, it is not clear how far ASSISTis usable by universit y lecturers. Entwist les concernto safeguard ideas about learning approaches fromoversimplificat ion in general use might be a reasonfor this. Nevertheless, notions such as deep, surfaceand strategic approaches to learning are now partof the everyday vocabulary of many HE teachersand the wealth of books on teaching techniques thatdraw directly on many of the concepts reviewed hereis testimony to Entwistles continuing influence onpedagogy in higher education. To use a term coinedby Entwistle himself, t he model has proved to bepedagogically fer tile in generating new ideas aboutteaching and learning in higher education.

  • 8/8/2019 FrankCoffieldLearning Styles

    112/182

    7.2Vermunt s fr amework f or classifying learningstyles and his Inventory of Learning Styles (ILS)

    Introduction

    Jan Vermunt is an associate professor in t he GraduateSchool of Education at Leiden Universit y. He alsohas a part-time role as professor of educationalinnovation in higher education at Limburg Universit y.His main areas of research and publicat ion have beenhigher education, teaching and teacher education.He began his research on the regulation of learning(ie the direction, monitoring and control of learning)and on process-oriented instruction in the psychologydepartment at Tilburg Universit y in the late 1980s.Vermunt has published extensively in English andin Dutch, and his Inventory of Learning Styles (ILS)is available in both languages.

    Definit ions, descript ion and scope

    For Vermunt, the terms approach to learning andlearning style are synonymous. He has tried to findout how far individuals maintain a degree of consistencyacross learning situations. He defines learning st yle(1996, 29) as a coherent whole of learning activitiesthat students usually employ, t heir learning orientationand their mental model of learning. He adds thatLearning style is not conceived of as an unchangeablepersonalit y attribute, but as t he result of the temporalinterplay between personal and contextual inf luences.

    This definit ion of learning style seeks to be f lexibleand integrative and, in comparison with earlierapproaches, strongly emphasises metacognitiveknowledge and self-regulation. It is concerned withboth declarat ive and procedural knowledge, includingself-knowledge. It deals not only with cogniti veprocessing, but also with motivation, effort and feelings(and their regulation). However its formulat ion wasnot directly influenced by personality theory.

    Within Vermunts framework, four learning stylesare defined: meaning-directed, application-directed,reproduction-directed and undirected. Each is said(1996) to have distinguishing features in f ive areas:

    the way in which students cognitively process learningcontents (what st udents do)

    the learning orientations of students (why they do it)

    the affect ive processes t hat occur during studying(how they feel about it)

    the mental learning models of students(how they see learning)

    the way in which students regulate their learning(how they plan and monitor learning).

    The resulting 4x5 matrix is shown in Table 35 andsuggests linked sets of behavioural, cognitive, affective,conative and metacognitive characteris tics. However,it should be noted that the framework is conceivedas a f lexible one. Vermunt does not claim that hislearning styles are mutually exclusive, nor that for alllearners, the links between areas are always consistentwith his theory. The case illustrations and quotationsprovided by Vermunt (1996) are captured in summaryform as learner characterist ics in Table 35. His fourprototypical learning styles are set out in columnsfrom left (high) to right (low) in terms of their presumedvalue as regards engagement with, and success in,academic s tudies.

    Origins

    Developed through his doctoral research project (1992),Vermunts framework has clearly been influencedby several lines of research about deep, surface andstrategic approaches to learning that date back tothe 1970s, and by Flavells ideas about metacognition(eg Flavell 1979). The work began with t he qualit ati veanalysis of interviews and later added a quantitativedimension through the development and use of the ILS(Vermunt 1994 ).

    The Inventory of Learning Styles

    Description of the m easure

    When the ILS was published, the original frameworkwas simplified in t hat affective processes didnot appear as a separate area. However, t he areaof learning orientations remains, encompassinglong-term motivation and goals, and (to a lesser extent)dimensions of interest and confidence. The ILS isa 120-item self-rating instrument, using 5-point Likertscales. Its composition in t erms of areas is shownin Table 36.

    Reliability and validit y

    Statist ical evidence to support the grouping of itemsinto sub-scales has been provided. In two large-scalestudies, Vermunt (1998) found that alpha valuesfor t he sub-scales were generally higher than 0.70.Confirmatory second-order factor analysis supportedin almost every detail the grouping of sub-scales intoVermunts hypothesised four learning styles, althoughthere was some overlap between styles.

  • 8/8/2019 FrankCoffieldLearning Styles

    113/182

    page 104/ 105LSRC reference Sect ion 7

    Table 3 5Vermunts learning styleswith illustrations of theircomponentsSource:Vermunt (199 0)

    Cognitive processing

    Learning orientation

    Affective processes

    Mental modelof learning

    Regulation of lear ning

    Meaning-directedLook for relationshipsbetween keyconcepts/ theories: buildan overview

    Self-improvement andenrichment

    Intrinsic interest andpleasure

    Dialogue with expert sstimulates thinking and

    engagement with subjectthrough exchange ofviews

    Self-guided by interestand their own questions;diagnose and correctpoor understanding

    Application-directedRelate topics to everydayexperience: look forconcrete examples anduses

    Vocational or real worldoutcomes

    Interested in practicaldetails

    Learn in order to useknowledge

    Think of problems andexamples to t estunderstanding,especially of abstractconcepts

    Reproduction-directedSelect main points toretain

    Prove competence bygetting good marks

    Put in t ime and effort;afraid of forgetting

    Look for structure inteaching and texts to

    help take in knowledgeand pass examinations.Do not value criticalprocessing or peerdiscussion

    Use objectives to checkunderstanding; self-test;rehearse

    UndirectedFind study difficult ; readand re-read

    Ambivalent; insecure

    Lack confidence; fear offailure

    Want teachers to domore; seek peer support

    Not adaptive

    Table 3 6Areas and sub-scalesof the ILS

    Area

    Cognitive processing

    Learning orientation

    Ment al modelof learning

    Regulation of learning

    Sub-scaleDeep processing:relating and structuringcritical processingStepwise processing:memorising and rehearsinganalysingConcrete processing

    Personally int erestedCertificate-orientedSelf-test-orientedVocation-orientedAmbivalent

    Construct ion of knowledgeIntake of knowledgeUse of knowledgeStimulating educationCooperative l earning

    Self-regulation:learning process and resultslearning contentExternal regulation:learning processlearning resultsLack of regulation

  • 8/8/2019 FrankCoffieldLearning Styles

    114/182

    The fit between theory and empirical findings seemsalmost too good to be true. In Table 37, exemplarsof each learning style are shown, const ructed by takingthe f irst item of each sub-scale with high factor loadingson each style factor. These exemplars certainly havea high degree of face validity as representing differentapproaches to study. It will be seen that there is somedegree of overlap between st yles, as well as twosignif icant gaps which are consistent with Vermuntstheory. As applicat ion-directed learners are t houghtto use a mixture of self-regulat ion and externalregulation, it is not surprising that there is no statementbased on the sub-scale loadings for regulation for suchlearners. The second gap is that there is no statementabout processing st rategies for undirected learners,which is consistent with Vermunts qualitative findingthat such learners hardly ever engage in study-relatedcognitive processing.

    The relevance of the ILS for use in the UK HE contexthas been established by Boyle, Duffy and Dunleavy(2003). The authors administered the 100-item (shortform) version of the ILS to 273 students. They foundthat three of the four main scales have good internalconsistency, while the fourth (learning orientation) hada borderline alpha value of 0.67. However, the reliabilit yof the 20 sub-scales was rather less satisfactory thanin Vermunts 1998 study, with only 11 sub-scales havingalpha values of 0.70 or above. Confirmator y factoranalysis suppor ted Vermunts model of four learningstyles, alt hough the application-directed and undirectedstyle measures showed less integration across

    components than the other two.Despite it s face and factorial validity andmultidimensional st ructure, i t has not been confirmedthrough independent research that the ILS is a goodpredictor of examination performance. With a sampleof 409 psychology undergraduates, Busato et al. (2000)found that only the undirected style predicted academicsuccess (negatively), and even then accounted forless than 4% of the variance over the first academicyear. Both t he meaning-directed st yle and openness(between which there was a Pearson r measureof 0.36 ) had vir tually zero correlations with fouroutcome measures. Achievement motivation and the

    personality variable of conscientiousness were slightlybetter predictors in this study, but not nearly as goodas performance on the fi rst course examination ona introductory module.

    In their UK study, Boyle, Duffy and Dunleavy (2003)also found that a factor measure of undirectedlearning st yle was a negative predictor of academicoutcomes for 273 social science students, but i taccounted for a mere 7% of t he variance. On thisoccasion, meaning-directed style was a posit ivepredictor, accounting for 5% of the variance, butneither reproduction-directed nor application-directedstyle yielded a signif icant correlation.

    Evaluation

    Vermunts framework was not designed to apply in allpost-16 learning contexts, but specifically to universitystudents. However, he and his students are, at t hetime of writing, developing a new instrument toassess learning at work and a new version of the ILSfor t he 1618-year-old group (Vermunt 2003). The new1618 instrument will take account of current t eachingpractices and will include an affective component.The ILS asks about:

    how students attempt to master a particular pieceof subject matter

    why they have taken up their present course of st udy

    their conceptions of learning, good education andcooperation with others.

    By limiting his focus to higher education, Vermunthas been able to produce a reliable self-assessmenttool, but this means that it s relevance is largelyunknown in other contexts, such as problem-basedlearning, vocational education, adult basicskills learning or work-based training. When aninst rument modelled on the ILS was applied by Slaats,Lodewijks and Van der Sanden (1999) in secondaryvocational education, only the meaning-directedand reproduction-directed patterns were found.Moreover, Vermunts framework does not map wellonto the categories empirically established in Canadianadult education sett ings by Kolody, Conti and Lockwood(1997). Cross-cultural differences in the factorst ructure of the ILS were reported by Ajisuksmoand Vermunt (1999).

    The structure of the framework consistsof Entwistle-like learning styles on the horizontal axis(which represent different levels of understanding)and a mixture of content and process categorieson the vert ical axis. This is clearly a framework ratherthan a taxonomy, as the vertical axis cannot be saidto represent a dimension.

  • 8/8/2019 FrankCoffieldLearning Styles

    115/182

    page 106/ 107LSRC reference Sect ion 7

    Table 3 7Exemplar vignettes ofVermunts four learningstyles using ILS items

    Meaning-directed exemplarWhat I do

    Why I do it

    How I see learning

    How I plan and monitor my learning

    I try to combine the subjects that are dealt with separately ina course into one whole.I compare my view of a course topic with t he views of t he authorsof the textbook used in that course.I use what I learn from a course in my activities outside my studies.

    I do these studies out of sheer interest in t he topics thatare dealt with.

    To me, learning means trying to approach a problem frommany different angles, including aspects that were previouslyunknown to me.

    To test my learning progress when I have studied a textbook,I tr y to formulate the main points in my own words.In addition to the syllabus, I study other literature related to thecontent of t he course.

    Application-directed exemplarWhat I do

    Why I do it

    How I see learning

    How I plan and monitor my learning

    I use what I learn from a course in my activities outside my studies.

    I do not do these studies out of sheer interest in the topics that aredealt with.I aim at attaining high levels of s tudy achievement.When I have a choice, I opt for courses t hat seem useful to me for mypresent or future profession.

    The things I l earn have to be useful for solving practical problems.

    Reproduction-directed exemplarWhat I do

    Why I do it

    How I see learning

    How I plan and monitor my learning

    I repeat the main parts of t he subject matter until I know themby heart .I work through a chapter in a textbook item by it em and I study eachpart separately.

    I aim at attaining high levels of s tudy achievement.

    I like to be given precise instruct ions as to how to go about solvinga task or doing an assignment.

    If a textbook contains questions or assignments, I work them outcompletely as soon as I come across them while studying.I experience the introductions, objectives, inst ructions, assignments

    and test items given by the teacher as indispensable guidelines formy studies.

    Undirected exemplarWhat I do

    Why I do it

    How I see learning

    How I plan and monitor my learning

    I doubt whether this i s the right subject area for me.

    I like to be given precise instruct ions as to how to go about solvinga task or doing an assignment.The teacher should motivate and encourage me.When I prepare myself for an examination, I prefer to do so togetherwith other students.

    I realise t hat it is not clear t o me what I have to remember and whatI do not have to remember.

  • 8/8/2019 FrankCoffieldLearning Styles

    116/182

    Definit ions of the four styles are reasonably clear.Meaning-directed cognitive processing has anemphasis on synthesis and critical thinking, whereasreproduction-directed processing emphasises analysisand to some extent, the unthinking studying of parts.However, this contrast is not without problems, as itcan be argued that mastery of a subject requires bothsynthesis and analysis in other words, a full anddetailed understanding of whole-part relationships.Vermunt acknowledges that learning styles can overlapand one example of this is t hat an interest in practicalapplications can be found alongside an interest inabst ract ideas and subject mastery. Indeed Vermunthimself found that meaning-directed learners tended togive themselves higher ratings for concrete processingthan did application-directed learners (Vermunt 1998).The undirected style seems to apply to less successfullearners. These may be people who study in haphazard

    or inconsistent ways or who simply do not study at all.In two studies where cluster analysis rather thanfactor analysis was used (Wierst ra and Beerends 1996;Vermetten, Lodewijks and Vermunt 2002), three, ratherthan four, groups were identif ied. In both cases,groups were found in which meaning-oriented deepprocessing was associated with self-regulation andin which reproduction-oriented surface processingwas associated with external regulation. The studiesdiffered, however, in finding rather different thirdclusters, called f lexible learners in one case andinactive learners in the other. This may reflect thefact that students in different faculties differ in learning

    style and clearly illustrates t he context dependencyof the framework.

    In some ways, Vermunts t reatment of regulationresembles the model of cognit ive engagement putforward by Corno and Mandinach (1983). Self-regulat ionappears in both models and Vermunts conceptof external regulation (meaning relying on externallyimposed learning objectives, questions and tests)resembles Corno and Mandinachs concept of passivelearning or recipience. However, unlike Corno andMandinach, Vermunt does not make full use of Kuhlstheory of action control (1983), since in t he ILS,he emphasises the cognitive rather than the affect ive

    aspects of metacognitive control. There are no itemsin the ILS relating to the control of motivation, emotionsor even attent ion. This may well limit t he predictivepower of the instrument.

    Vermunts framework is compatible with more thanone theory of learning, as one would expect froman approach which seeks to integrate cognit ive,affective and metacognit ive processes. His valuingof meaning-directed and application-directed waysof learning as well as process-based instruction(Vermunt 1995) reflects mainly cognitive andmetacognit ive theorising. He accepts t hat learnersconstruct meanings, but has de-emphasised theinterpersonal context of learning, as only undirected(largely unsuccessful) students tend to see learningin terms of opportunities for social st imulation/ entertainment and cooperation (possibly in orderto compensate for their fear of failure). He makes useof behavioural discourse when he speaks of t he needfor teachers to model, provide feedback and test.However, as argued above, his t reatment of the affectivedomain and of personality factors is rather incomplete.

    So far as conation is concerned, this is not neglected,as the word t ry appears in 20 different ILS items.

    The empirical basis for the framework as presentedin 1998 is very much stronger than in the 1996 paper.The 1996 qualitative data was based on interviewswith only 24 first-year Open University students takingdifferent courses and 11 psychology students ata tradit ional universit y; nor did t he paper includea full audit trail for the categorisation of statements.However, t he psychometric suppor t for the ILS isreasonably robust, even though we are not told exactlyhow the choice of items for the sub-scales was made.A number of researchers have found testretest

    correlations for each of the four areas in the range0.4 to 0.8 over periods of between 3 and 6 months.This suggests that there can be as much variabilityand change as stabili ty in approaches to study. Indeed,Vermetten, Lodewijks and Vermunt (1999) found t hatlaw students were using different learning strategiesat t he same time on four different courses.

    It would be inappropriate to regard Vermuntsframework as definit ive. It may not be applicableto all t ypes and stages of learning. If i t is t o beused in post-16 contexts outside higher education,further theory development and validation willbe needed, possibly allowing personality, affect ive,

    social-collaborative and study-skill componentsto feature more prominently. The well-supportedtheoretical models of Demetriou (Demetriou andKazi 2001) and Marzano (1998) suggest promisingways forward. At the same time, it will be importantto evaluate and seek to improve teaching andstudy environments as much as learning styles,since learning takes place where person and situationinteract. In recent work, Vermunt has addressedthis area using the ILS and the Inventory of PerceivedStudy Environments (IPSE) (Wierstra et al. 2002).

  • 8/8/2019 FrankCoffieldLearning Styles

    117/182

    Implications for pedagogy

    Vermunt developed his framework for use wit hpost-16 learners and although its main use has been asa research tool, it is likely to be seen as meaningful and

    helpful by both learners and teachers. Technical termssuch as metacognition , regulation and affective do notappear in the ILS itself, but will need clear defini tionand explanation for teachers who use it. The vocabularydemand of the ILS is around 1213 years accordingto the Flesch-Kincaid readabilit y index. The frameworkis not too complex for everyday use and its emphasison the importance of motivation and metacognit ionduring adolescence and beyond is well suppor ted byresearch (Marzano 1998; Demetriou and Kazi 2001).It cer tainly provides a common language for teachersand learners to discuss how people try to learn, whythey do it, how different people see learning, how theyplan and monitor it and how teachers can facilit ate i t.

    Vermunt believes that meaning-directed approacheswill prove superior t he more courses move away fromtraditional teaching programmes (with a high focuson teacher control and the transmission of knowledge)towards process-oriented study programmes whichfocus on knowledge construction and ut ilisat ionby learners and are characterised by a gradual andsystematic transfer of control over learning processesfrom instruct ion to learners (Vermunt 1996, 49).He believes that this process will be facilit atedif teachers become more aware of individual differencesin learning st yle and address weaknesses by teachingdomain-specific thinking and learning strategies.Research by Schatteman et al. (1997) into the effectof interactive working groups is consistent withthese ideas, but is far from definiti ve, as t he groupswere not well attended and data was available foronly 15 part icipants.

    In addition to this, Vermunt sees considerable potentialin the use of the ILS to reveal dissonant approaches tolearning; for example, by students who combine externalregulation with deep processing or self-regulation withstepwise processing. So far, there are a few studieswhich suggest that such combinations are maladaptive(eg Beishuizen, Stout jesdijk and Van Putten 1994).

    Recognising that teachers themselves have learningstyles which may well affect their practice, Vermunthas been involved in a number of studies in whichhis model has been applied in work with teachersand student teachers (eg Zanting, Verloop and Vermunt2001; Oosterheert, Vermunt and Denissen 2002).In these contexts, he has again used qualitat iveapproaches to assessing learning orientation,affective processes, mental models of learning andself-regulation as a basis for developing more objective,contextually appropriate methods. This work showsgreat promise for teacher education and professionaldevelopment in all sectors, including post-16 educationand training.

    In a theoretical paper on congruence and frictionbetween learning and teaching, Vermunt and Verloop(1999) suggest that both congruence and constructivefriction between student and teacher regulationof learning are likely to prove beneficial. They claimthat congruence is to be found:when teacher regulat ion is high and studentregulation is low

    when student regulation is high and teacherregulation is low.

    Constructive friction occurs in si tuations wherethe teacher expects s tudents to perform withgreater self-regulation, whereas destructive frictionis experienced when students are capable of moreautonomy than their teachers allow or when theyare incapable of taking responsibility for their ownlearning in a loosely structured learning environment.These ideas imply that t eachers need to understandtheir students better than at present and to becomemore versat ile in the roles they adopt. Common sensewould support these notions, at least on the basisof extreme case scenarios, but their practical util it yacross higher education and for lifelong learningis as yet largely untested.

    Vermunts research into the learning of undergraduatestudents and others has had significant impactin northern Europe. Its main thrust has beento encourage learners to undertake voluntarily verydemanding activities such as relating and structuringideas, crit ical processing, reading outside thesyllabus, summarising and answering self-generatedquestions. This kind of approach requires strongmotivation, intellectual openness, a conscientiousattit ude, a sense of self-efficacy and self-confidenceplus well-established and eff icient metacognitiveand cognitive strategies. These qualities havefor many years been seen as desi rable outcomesof higher education. However, al though they canbe acquired and developed, there is no easy wayin which this can be achieved in the diverse areasof post-16 lifelong learning.

    Vermunt has performed a valuable service in showingthat, if progress is to be made, attention needsto be given not only to individual differences in learners,but to the whole teachinglearning environment.While the motivations, self-representations,metacognit ive and cognitive st rengths and weaknessesof learners are of concern to all involved in education,it is clear that t hese are also a function of the systemsin which learners find themselves. Vermunts conceptualframework and t he ILS can usefully help to developa better understanding of these complexities.His approach can cert ainly be adapted for usein all contexts of li felong learning.

    Empirical evidence of pedagogical impact

    As yet, there is little evidence of this kind, apar tfrom the studies mentioned in the previous sub-section.The ILS has not been widely used in pos t-16intervention studies.

    page 108/ 109LSRC reference Sect ion 7

  • 8/8/2019 FrankCoffieldLearning Styles

    118/182

    Table 3 8Vermunts Inventory ofLearning Styles (ILS)

    General

    Design of the model

    Reliability and validit y

    Implicationsfor pedagogy

    Eviden ce ofpedagogical impact

    Overall assessment

    Key source

    WeaknessesIt has littl e to say about how personalityinteracts wit h learning style.

    It excludes preferences for representi nginformation.It is not comprehensive: there are noitems on the control of motivation,emotions or attention.The interpersonal context of learning isunderemphasised.Not applicable to all t ypes and stages oflearning.Notions of constructive anddestructive friction are largelyuntested.

    Little evidence so far of impact onpedagogy.It is not a st rong predictor of learningoutcomes.

    StrengthsIt applies to t he thinking and learning ofuniversity students.New versions in preparation f or 1618age group and for learning at work.

    Used for studying the learning styles ofteachers and student teachers.

    It is experientially grounded ininterviews with students.It seeks to integrate cogniti ve, affecti ve,metacognitive and conative processes.It includes learning strategies,motivation for learning and preferencesfor organising information.

    It can be used to assess approaches tolearning reliably and validly.

    It is dependent on context, ie a learningstyle is the interplay between personaland contextual influences.It provides a common language forteachers and learners to discuss andpromote changes in learning andteaching.Emphasis not on individual diff erences,but on t he whole teachinglearningenvironment.

    A rich model, validated for use in UK HE contexts, with potential for more general usein post-16 education where text-based learning is import ant. Reflect ive use of t heILS may help learners and teachers develop more productive approaches tolearning.

    Vermunt 1998

  • 8/8/2019 FrankCoffieldLearning Styles

    119/182

    7.3Sternbergs theor y of think ing styles and hisThinking Styles Inventory (TSI)

    Introduction

    Robert Sternberg is a major figure in cognitivepsychology; he is IBM professor of psychology andeducation at Yale University and was presidentof the American Psychological Association in 2003/ 04.His theory of mental self-government and modelof thinking styles (1999) are becoming well knownand are highly developed into functions, forms, levels,scope and leanings. He deals explicitly with therelationship between thinking styles and methodsof instruction, as well as the relationship betweenthinking styles and methods of assessment. He alsomakes major claims for improving student performancevia improved pedagogy.

    Definit ion, descript ion and scope of t he model

    Sternberg is keen to distinguish between styleand abili ty. An abilit y refers to how well someonecan do something. A style refers to how someonelikes to do something. A style therefore is a preferredway of using the abil it ies one has (1999, 8). We donot have a style, but rat her a profile of st yles(1999, 19; original emphasis).

    In his book on Thinking styles (1999), Sternberg usedthe two terms thinking st yles and learning stylesas synonyms; for example (1999, 17): Teachers

    fail to recognise the variety of t hinking and learningstyles that students bring to the classroom andso teach them in ways that do not f it these st yleswell. However, by 2001, Sternberg was making cleardistinct ions between learning, thinking and cognitivestyles. In more detail, he conceptualised learningstyles as how an individual prefers to learn by reading,for instance, or by attending lectures. Thinking st ylesare characterised as how one prefers to think aboutmaterial as one is learning it or after one alreadyknows it (Sternberg and Zhang 2001, vii). Cognitivestyles are described as the ways of cognizing (sic)the information (Sternberg and Zhang 2001, vii)by being impulsive and jumping to conclusions,or by being reflecti ve. Cognitive st yles are consideredby Sternberg to be closer to personali ty than ei therthinking or learning styles.

    Sternbergs theory of thinking/ learning styles isderived from his theory of mental self-government,which is based on the metaphorical assumption(for which no evidence is offered) that the kindsof government we have in the world are not merelyarbitrary or random constructions, but ratherin a certain sense are mirrors of the mind on thisview, t hen, governments are very much extensionsof individuals (1999, 148). Sternberg chooses fourforms of government: monarchic, hierarchic, oligarchicand anarchic, but not democratic or dictatorial.No explanation is given as to why these four formsof government have been chosen and others excluded.

    His theory is constructed from three functionsof government (legislat ive, executive and judicial);four forms (monarchical, hierarchical, oligarchic andanarchic); two levels (global and local); t he scopeof government which is divided into internal andexternal; and leanings (liberal and conservative).Each of these aspects of government is considerednecessary for the management of the self in everydaylife. Sternberg provides a diagrammatic summaryof his styles; he does not call it a taxonomy, but thatis what it amounts to (see Table 39).

    A brief descript ion of the 13 styles is given below.

    1Legislative people like to come up with their own waysof doing things and prefer to decide for themselveswhat t hey will do and how they will do i t. This st yle isparticularly conducive to creativity: In schools as wellas at work, legislative people are often viewed as notfi tting in, or perhaps as annoying. (1999, 33)

    2Executive people like to follow rules and preferproblems that are pre-structured or prefabricated executive stylists do what they are told and often doit cheerfully (1999, 21). They are implementers who

    like to follow as well as to enforce rules. They can oftentolerate t he kinds of bureaucracies that drive morelegislati ve people batty (1999, 35).

    page 110/ 111LSRC reference Sect ion 7

    Table 3 9Summary of stylesof thinkingSource: Sternberg (1999 )

    FunctionsLegislative

    Executive

    Judicial

    FormsMonarchic

    Hierarchic

    Oligarchic

    Anarchic

    LevelsGlobal

    Local

    ScopeInternal

    External

    LeaningsLiberal

    Conservative

  • 8/8/2019 FrankCoffieldLearning Styles

    120/182

    3Judicial people like activities such as writ ingcrit iques, giving opinions, judging people and theirwork, and evaluating programs (1999, 21). They like toevaluate rules and procedures; they prefer problems inwhich they can analyse and evaluate t hings and ideas(1999, 39).

    Sternberg makes three general points about this style.every organisat ion needs judicial people as well aslegislative and executive ones (1999, 40).t he same person can and typically will perform allthree of these functions in greater or lesser degree.But people often feel more comfortable in one roleor another (1999, 40).Any number of people who might be legislat ive inschool might be executive in their choice of clothingor vice-versa. We thus need to underst and st yles

    in the contexts in which they are expressed (1999, 43).The signif icance of context is explicitly acknowledged,but not explored in any detail.

    4Monarchic people are single-minded and drivenby whatever they are single-minded about, and do notlet anything get in t he way of them solving a problem.They tend to be mot ivated by a single goal or needat a time (1999, 46).

    5Hierarchic people recognise the need to setpriorities, accept complexity and tend to fi t well into

    organisations because they recognise the need forpriorities (1999, 23). They tend to be systematic andorganised in their solutions to problems and in theirdecision making (1999, 51 ).

    6Oligarchic people tend to be motivated by several,often competing goals of equal perceived importance(1999, 23). The oligarchic person is a cross betweena monarchic person and a hierarchic one (1999, 54).

    7Anarchic people seem to be motivated by a potpourriof needs and goals t hat can be difficult for them,as well as for others, to sor t out (1999, 23). They areat risk for anti-social behaviour they are the studentswho challenge teachers, not necessari ly on principledgrounds, but rather for the sake of challenging theteachers or any other authority figures (1999, 58 ).They can challenge the system and have a potent ialfor creativity.

    Sternberg argues appropriately that these stylesare not in and of themselves good or bad (1999, 51),but it is important to point out that the titles(eg monarchic, anarchic) he employs are evaluativeand normative.

    8Global individuals prefer to deal with relat ively largeand abstract issues. They ignore or dont like details,and prefer to see the forest rather than the trees(1999, 24).

    9Local individuals like concrete problems requiringworking with details. The danger is they may lose theforest for the trees (1999, 24).

    Sternberg argues that: Most people tend to beeit her more global or more local: they focus more onthe big picture or more on the small details. But somepeople are both: they are equally attentive to the bigpicture and to the little details (1999, 64).

    10Internal individuals tend to be introvert ed,task-oriented, aloof and sometimes socially lessaware. They like to work alone (1999, 25).

    11External individuals tend to be extrovert ed, outgoingand people-oriented. Often, they are socially sensitiveand like working with other people wherever possible(1999, 25). According to Sternberg, In management,a distinction is sometimes made between task-orientedand people-oriented managers. This distinction isroughly comparable to that between internalists andexternalists (1999, 70).

    12Liberal individuals like to go beyond existing rules and

    procedures, to maximise change, and to seek situationsthat are somewhat ambiguous (1999, 26).

    13Conservative individuals like to adhere to existing rulesand procedures, minimise change, avoid ambiguoussituat ions where possible, and st ick with familiarsituations in work and professional life (1999, 26).

    In general, Sternberg wishes to distinguishbetween st ylist ic leanings and polit ical ones(1999, 75). Sternberg argues that the two are probablyonly weakly correlated, if at al l, and he gives theexample of the US polit ician, Newt Gingrich, who has

    a conservative political philosophy, but a decidedlyliberal personal style.

  • 8/8/2019 FrankCoffieldLearning Styles

    121/182

    The 15 principles of t hinking styles

    Sternberg makes 15 general points about this theorywhich he feels are essential to its understanding andthese are listed briefly below.

    1Styles are preferences in the use of abili ties,not abilit ies themselves.

    2A match between styles and abilities creates a synergythat is more than the sum of its par ts.

    3Life choices need to f it styles as well as abilities;for example, careers and choice of spouse.

    4People have profiles (or patterns) of styles, not justa single style.

    5Styles are variable across t asks and si tuat ions;for example, inf luence of weather, company, etc.

    6People differ in the strength of their preferences.

    7People differ in t heir st ylistic f lexibilit y.

    8Styles are socialised that is, they are learned;for inst ance, by children observing role models.

    9Styles can vary across the lifespan that is, styles,like abilit ies, are fluid rather than fixed, and dynamicrather than static enti ties; for example, t he style neededby a new recruit is very different from that neededby a senior partner in a law firm.

    10Styles are measurable.

    11Styles are teachable.

    12Styles valued at one t ime may not be valued atanother. (His claim is that different styles are requiredfor different levels or kinds of responsibilit y in anorganisation, which seems remarkably similar to theninth principle.)

    13Styles valued in one place may not be valued in another.

    14Styles are not, on average, good or bad it is a quest ionof fit. A style may fit well in one context, but poorly ornot at all in another.

    15We confuse stylistic fi t with levels of abili ty.The consequence is t hat people and inst itut ions tendto value other people and insti tut ions that are likethemselves. (But the question needs to be asked:do we not at t imes also value people precisely becausetheir style is very different from our own?)

    Origins and influence

    One of the att ractions of Sternbergs approach is thathe ends his book (1999) by raising 10 of the mostfrequently mentioned problems with theories of learning

    styles and claims to deal with t hem all satisfactorily.As will become clear, however, some of the problemsare just as applicable to Sternbergs own work as theyare to the research of those he criticises. He beginsby asking: Why do we need another theory? Whatare the problems with theories of learning styles?The 10 problems he tackles are listed below,together with a brief account of his response, plussome comment from this research team (material inbrackets), where appropriate.

    1There is no unifying model or metaphor that integratesthe various st yles, not only between theories, but even

    within theories. Sternbergs contention is t hat his t heoryof mental self-government provides a clear organisingmetaphor, namely that of government.

    2Some of the st yles seem too much like abilities;for example, the f ield dependence/ independencetheory of Witkin.

    3Some of the learning styles seem too much likepersonality traits; for example, Myers-Briggs.Sternberg argues that s tyles differ from personalitytraits in being more cognitive.

    4There is no compelling demonstrat ion of therelevance of the styles in real world settings.(This is so, but it is also true of Sternbergs own theory.)

    5There is insufficient connection between the theoriesof st yles and psychological theory in general. Sternbergargues that styles cannot be considered independentlyof the environment in which they occur. People act ivelyrespond in varied ways to the environment, dependingin large part upon their st yles of responding.

    6The styles specified by the theories are sometimessimply not compelling. Sternberg lists five criteriafor a successful t heory is it elegant, reasonablyparsimonious, internally coherent, empirically validand heuristically useful? He then claims that his theorymeets all five criteria. (We would argue that thereare serious questions to be asked about the validityand reliability of his theory.)

    7There is insuff icient use of converging operationsor multiple methods of measurement.

    page 112/ 113LSRC reference Sect ion 7

  • 8/8/2019 FrankCoffieldLearning Styles

    122/182

    8There is litt le or no serious research to show theusefulness of the styles. In Sternbergs own words(1999, 155): Theories and research on styles areat t he fringes of the psychological world. In this areaof psychology, there is a high ratio of theor y to data in everyday terms, that means big talk, no show Many schools are buying into systems for assessingstudents learning styles and for teaching the studentsthat have no solid research base at all (1999, 155).(This is our central crit icism of Sternbergs own work.)

    9The theories do not seem to be theories of styles at all,but rather of the variables that affect s tyles. Sternbergis right to claim that this criticism applies most clearlyto t he theory of Dunn and Dunn, who concentrateon environmental variables which may affect learningstyles.

    10The styles specified by the theories do not sat isfysome or even most of the 15 principles l isted above.

    Measurement by the author

    Description

    Sternberg has administered his inventoryof thinking/ learning styles in schools and elsewhere.In all, four measures have been used and theseare described brief ly below.

    1The Thinking Styles Inventory: 13 inventories witheight statements rated on a 17 scale.

    2The Thinking Styles Tasks for Students which, Sternbergclaims, measure styles via performance rather than viaan inventory; for example, When Im studying literature,I prefer. The student chooses from a legislat ive,executi ve or judicial response or some other response.(The response, however, does not comprise observedperformance, but self-reports of likely performance.)

    3The Thinking Styles Questionnaire for Teachers which

    assesses the styles teachers use when they teach(1999, 124) or rather the st yles which teachers reportthat they use.

    4Students Thinking Styles Evaluated by Teachers.

    Very little information is provided on the second, thirdor fourth of t hese instruments and yet Sternberg claimsthat these four measures meet the criteria for beinggood tests (1999, 125).

    Reliability and validit y

    There are few details given about the reliabilityand validity of these inventories. What data is providedis summarised below. In The MSG Thinking Styles

    Inventory by Sternberg and Wagner (1991), whichis unpublished; the learner completes each of the13 inventories on a 7-point scale from the statementf its me not at all well to f its me extremely well.Each style may vary across tasks, situat ions and yourtime of life (1999, 30).

    With regard to the TSI, Sternberg (1999, 125) claimsthat the 13 scales had internal-consistency reliabilitiesranging from .57 to .88 with a median of .82. Factoranalysis was employed and identified five factors,three of which were predicted and consistent with thetheory; one was not predicted, but was consistent;while the last was neither predicted nor consistent.

    Sternberg concludes: Thus the stat istical analysisgenerally supported the theory, although the secondfactor remains unexplained (1999, 12 6).

    Sternberg also claims t hat his scales correlatewith scores on other test s, thus demonstrating goodexternal validit y . With the Myers-Briggs Type Indicator,for example, 30 out of 128 correlations werestatistically significant; and 22 correlations outof 52 were signif icant wit h the Gregorc Style Delineator(see Zhang and Sternberg 2001 for further details).In general, the posit ion of Sternberg and his associatesis t hat The TSI has been shown to be reliable andvalid for US samples (Zhang and Sternberg 2001, 204).

    External evaluation

    Reliability and validit y

    Porter (2003) tested the reliability and validit yof the TSI in a study of 150 fi rst-year psychologyundergraduates at Westminster Universit y. Accordingto Porter, t he theory of mental self-government (MSG)and the TSI inst rument have been presented inthe lit erature as potentially powerful tools for usein higher education (2002, 296) and so need to beindependently evaluated. Porter describes otherstudies (eg Zhang and Sternberg 2001), which

    concluded that thinking st yles contribute to academicachievement and that this contribution is differentiallyrelated to culture and gender. Porters study, however,offers only limited support for the theory of MSGand the reliabilit y and validity of the TSI (2002 , 301);he argues, therefore, that both will have to be improvedbefore the TSI can be used in educational pract ice.Port ers students found the MSG theory bothplausible and interesting, but they considered the13 inventories to be both too long and boring. Porteralso questioned whether f irst-year students understandtheir own learning well enough to complete theinventories satisfactorily.

  • 8/8/2019 FrankCoffieldLearning Styles

    123/182

    Sternbergs theory and the TSI were part of the batteryof tests used by Demetriou and Kazi (2001) in theirattempt to build and test a theory of the mind and itsdevelopment from childhood to adolescence. The scaleof the project is impressive, with a sample of 840participants from 10 to 15 years of age in Thessaloniki,Greece and a follow-up study of 322 students fromthe University of Cyprus. It is, however, impor tantto realise that only the fi rst t wo of the f ive dimensionsof Sternbergs theory were test ed (ie function andform were tested, but not level , scope and learning ).Moreover, t he test of thinking styles consti tuted onlya very small par t of the data collect ion which involvedthree test ing periods of 2 hours; the battery consistedof six tests of cognit ive ability (quantitat ive, causal,spatial, social understanding, drawing and creativity)and self-evaluation questionnaires on cognitive ability,personality, cognitive and problem-solving strategies

    and occupational preferences, as well as thinkingstyles. It is, nevertheless, important to note that thealphas for the three styles: executive (0.56), legislative(0.51), and evaluative (0.59) were considerably lowerthan t hose which Sternberg claimed for them.

    Demetriou and Kazi (2001, 196) conclude thatSternbergs thinking styles

    are derivatives of the more fundamental dimensions involved in the realms of personality and cognition.In a sense, this f inding is in line with Sternbergs conception of thinking styles as the interface between personality, intelligence and actual performance.One can live without them

    No conclusions were drawn by these authors in relat ionto thinking st yles and pedagogy.

    General

    Each of the 13 styles is based on a shor tself-assessment inventory of no more than eightquestions, some of which may strike some respondentsas unanswerable; for example, Question 1 in theExternal Style Inventory reads: When starting a task,I like to brainstorm ideas with friends or peers.

    This statement is likely to raise the following questionsin the minds of respondents: does this refer to everytask? Is brainstorming appropriate for all tasks?Without a detailed description of the kind of task thepsychologist has in mind, some respondents mayfind themselves unable and unwilling to answer thisquestion. It does not matter how sophisticated thestat istical analysis of responses to such questions is,if the responses do not accurately reflect the behaviourof the respondents. Each of the 13 inventories hasa similar vague statement; for example, the MonarchicStyle Inventory contains the following statement:When trying to finish a task, I t end to ignore problemsthat come up. We argue that i t depends on the taskand on the type of problem that comes up.

    The statements in the 13 inventories are ratherobvious, so i t is relatively easy to guess the intentionsof the psychologist who wrote the item. It wouldtherefore be simple to fake a response, for instance,to a Conservati ve Style statement such as When facedwith a problem, I like to solve it in a t radit ional way.Respondents could decide whether they wish to appearas left- or right-wing or somewhere in between.

    Implications for pedagogy

    The signif icance for pedagogy of Sternbergs researchon thinking styles can be summarised in five briefproposit ions which are of a very general nature.

    Teachers should use a variety of teaching methods(eg lectures, group discussions).

    Teachers should use a variety of assessment methods(eg multiple-choice questions, essays, projects).

    Teachers should provide students with anunderstanding of different thinking styles and shouldthemselves be aware of the styles they eitherencourage or punish.

    Teachers should know about gender and cross-culturaldifferences in thinking st yles.

    Teachers should use extracurricular activitiesto enhance the quality of teaching and learning(see Zhang and Sternberg 2001).

    The fifth recommendation does not appear to stem fromSternbergs own research, but from the work of others

    on creative thinking.Sternberg is convinced that his theory is importantfor pedagogy and has carried out a series of studiesof thinking/ learning styles in both secondary andhigher education, and cross-cultural studies in China,Hong Kong and the US. In his own words (1999, 115):The key principle [of the theory] is that in orderfor students to benefi t maximally from instructionand assessment, at least some of each should matchtheir styles of thinking. He is convinced that differentmethods of instruction work best for different stylesof thought and produces a table (reproduced hereas Table 40 ) to show the various types of compatibilit y.

    His argument is that teachers need the f lexibilit y tovary their teaching style to suit s tudents different stylesof thought and that few methods of instruct ion arelikely to be optimal for everyone.

    Again, Sternberg argues, without any supportingevidence, t hat different methods of assessment tendto benefit different thinking styles and producesa table to exemplify the connections (see Table 41).

    page 114/ 115LSRC reference Sect ion 7

  • 8/8/2019 FrankCoffieldLearning Styles

    124/182

    Table 4 0Thinking st yles andmethods of instructionSource: Sternberg (1999)

    Met hod of instructionLecture

    Thought-based questioning

    Cooperative (group) learning

    Problem solving of given problems

    Projects

    Small group:students answering factual questions

    Small group:students discussing ideas

    Reading

    Style(s) most compatiblewith method of instructionExecutive, hierarchical

    Judicial, legislat ive

    External

    Executive

    Legislative

    External, executive

    External, judicial

    Internal, hierarchical

    Table 4 1Thinking stylesand methodsof assessmentSource:Sternberg (1999)

    Met hod of assessmentShort-answer andmultiple-choice tests

    Essay test s

    Projects and port folios

    Interview

    Most compatible stylesExecutive, localJudicial, localHierarchicalInternal

    Executive, localJudicial, globalJudicial, globalLegislativeHierarchicalHierarchicalConservative

    Internal

    JudicialLegislativeExternalInternalHierarchicalMonarchic

    External

    Main skills t appedMemoryAnalysisTime allocat ionWorking by self

    MemoryMacro analysisMicro analysisCreativityOrganisationTime allocat ionAcceptance of teacher

    viewpointWorking by self

    AnalysisCreativityTeamworkWorking by selfOrganisationHigh commitment

    Social ease

  • 8/8/2019 FrankCoffieldLearning Styles

    125/182

    Empirical evidence for im pact on pedagogy

    Sternberg and his associates (eg Grigorenko and Zhang)have carried out many studies exploring particularaspects of the theory of mental self-government and

    the TSI: for inst ance, t he ability of thinking styles topredict academic achievement over and above abilit y;the relationships between thinking styles and learningapproaches, student characteristics (such as age,gender and socio-economic status) and self-esteem.The significance for pedagogy of the findings of thesestudies tends to be inferred by the authors rather thandirectly studied. The results most relevant to pedagogyinclude the f indings from a study of four US schools thatstudents performed better when they were more liketheir t eachers st ylistically, independent of actual levelof achievement and that different school[s] rewardeddifferent styles (Sternberg 1999, 130). In general,it can be said that the earlier studies with Grigorenkowere carried out with relatively small samples(eg 124 students from four schools), but the latercross-cultural studies with Zhang involve substantialnumbers of participants (eg 646 students fromHong Kong, 215 from China and 67 from the US):see Zhang and Sternberg (2001) for more details.

    Conclusions

    Sternberg has produced an original theory of mentalself-government (MSG) and has derived his TSI from it;this is beginning to be used and tested, part icularlyin China. It is important to realise that t his new theoryhas not been developed from the thinking or empiricalstudies of other researchers, so it may be better toconsider it not as a theory of learning or thinking st yles,but as an int riguing metaphor which may or may notprove to be productive in stimulating research andin changing practice. It is, at present, too early to offera comprehensive evaluation.

    A series of research projects in universities andsecondary schools in the US, Hong Kong and mainlandChina are now enhancing our understanding of t hinkingstyles. The claims made for the implications of thetheory for pedagogy are extensive, but the numberof empirical studies which have tested these claimsremains low. Moreover, the implicat ions for pedagogythat Sternberg lists are of a very general natureand some of them have only a tenuous connectionwith his research.

    One possible (but highly unrealistic) outcome fromthis theory, which describes no less than 13 differentthinking styles, is that teachers and tutors could beinvited to produce lessons which cater for all 13 styles.Sternberg avoids such diff icult ies by couching his advicein very general terms; for example, that t eachers shoulduse a variety of teaching and assessment methods andshould provide their s tudents with an understandingof different s tyles. In other words, the implicationsfor pedagogy are based on common-sense inferencesfrom the theory rather than on the f indings of anyexperimental studies.

    Grigorenko and Sternberg (1995) have suggestedtwo main reasons for the sudden flowering of researchinterest in learning styles in the late 1960s and early1970s. First , the notion was attract ive to many theoristsbecause of their disappointment with intelligencetests and the need for new measures of individualdifferences (1995 , 218). Second, researchers frompsychology and business s tudies began to explorethe concept of learning styles because it was so flexibleand ill defined.

    More recently, Sternberg has assessed thelearning/ thinking/ cognitive styles field and addressedthe mystery of why such research, so active and unifiedunder the cognitive styles banner in the middle of the[20th] century, seems to be so much less unifiedand active by the end of the century (2001 , 249 ).He attributed the current lack of unity and activit yto four main reasons: the early theories were notdistinguishable from abilities or personality traits;the main theorists remained isolated from each otherand from the psychological li terature more generally;the quali ty of early empirical research was poor;and no common conceptual framework or languagehas emerged in its place, different languagesand labels have proliferated. Sternberg concluded(2001, 250) as follows: The result is a kindof balkanisation of research groups, and balkanisationhas always led to division and, arguably, death bya thousand cuts. It is also arguable that Sternberghas himself contributed to such balkanisation and thatthe answer to his own question do we need anothertheory of learning styles? is probably best answered

    in the negati ve.

    page 116/ 117LSRC reference Sect ion 7

  • 8/8/2019 FrankCoffieldLearning Styles

    126/182

    Table 4 2Sternbergs ThinkingStyles Inventory (TSI)

    General

    Design of the model

    Reliability and validit y

    Implicationsfor pedagogy

    Eviden ce ofpedagogical impact

    Overall assessment

    Key source

    WeaknessesWhy these 13? 13 are too many.Learners self-assess their l ikelybehaviour by responding to s tatementswhich are context-free.

    Sternberg offers a metaphor ratherthan a theory.No explanation is given as to why someforms of government (eg monarchic) arechosen and not others (eg democratic).

    Only limited empirical support for thereliabilit y and validit y of the TSI.Scores for reliabili ty considerably lowerthan t hose found by author.Little or no support for validity of the TSI.

    No solid research base for thesesuggestions, which are logicaldeductions from the t heory.

    Fifth suggestion stems f rom researchon creativity, rather than learningstyles. The advice is of a very general,common-sense nature, most of it knownto teachers before any research doneon learning styles.

    There is a need for independentevaluation.

    Strengths13 thinking styles are proposed,based on the functions, forms, levels,scope and leanings of government.

    Based on a new theoryof mental self-government.

    Claimed by author to be both reliableand valid.

    Teachers to use a variety of teachingand assessment methods.Teachers to be aware of the learning

    styles they encourage or punish.Teachers to let students know aboutthe range of styles.Teachers to know about gender andcross-cultural differences in styles.Teachers to use extra-curricularactivities to enhance qualityof teaching and learning.

    A series of st udies in t he US and Chinahave so far produced mixed results.

    An unnecessary addition to the proliferation of learning styles models.

    Sternberg 1999

  • 8/8/2019 FrankCoffieldLearning Styles

    127/182

    This section begins by discussing the variousteaching strategies that the developers and advocatesof learning st yle inst ruments have suggested, witha brief evaluation of t he strengths and weaknessesof each. This entry into the world of course developers,insti tut ional managers and front-line practit ionersnecessarily involves us in a much wider literature thanthat consulted for the 13 major models evaluatedearlier in this report .

    The sub-sect ions which follow attempt to answer twoquestions which are crucial for educational practice.

    Why do some people find learning st yles so appealing?

    Why do others f ind them unacceptable?

    We then discuss the lack of research into pedagogyin the UK, part icularly compared with Germany; andwe offer a brief overview of the different definitions

    of, and approaches to, pedagogy which have been takenby psychologists, sociologists and adult educators.This section ends with the crucial distinction, drawn byAlexander (2000), between teaching and pedagogy;we argue that the learning styles literature is in themain concerned with the former rather than the latt er.

    What advice for pr actit ioners?

    In the current state of research-based knowledgeabout learning styles, there are real dangersin commending detailed strategies to practi tioners,because the t heories and inst ruments are not equallyuseful and because there is no consensus aboutthe recommendations for pract ice. There is a needto be highly select ive. As we have seen, for example,with regard to Dunn and Dunn (Section 3.2), Gregorc(Section 3.1) and Riding (Section 4.1), our examinationof the reliabilit y and validity of t heir learning styleinstruments s trongly suggests that they should notbe used in education or business. On the other hand,the research of Entwistle (Section 7.1) and Vermunt(Section 7.2), which is both more guarded in its claimsand built on more solid theoretical foundations,offers thoughtful advice that might, after careful t rialsand revisions, be extended to post-16 learning outsidehigher education.

    A significant proportion of the literature on thepractical uses of learning styles is not, however,so circumspect. Fielding, for instance, goes so far asto argue that an understanding of learning styles shouldbe a student enti tlement and an institutional necessity(1994, 3 93). A thriving commercial industr y has alsobeen built to offer advice to teachers, tutors andmanagers on learning st yles, and much of it consistsof inf lated claims and sweeping conclusions whichgo beyond the current knowledge base and the specif icrecommendations of particular theorists. For example,McCarthy (1990) developed what she calls the 4MATcycle of learning from Kolbs model, and a US website

    (www.volcano.und.nodak.edu/ vwdocs/ msh/ llc/ is/ 4mat.html) devoted to her approach claims thatIt represents graphically the teacher behaviorsappropriate to each stage and st yle, and providesa framework for planning any lesson or unit , for anyage level or content area.

    Some of t he leading learning theorist s, moreover,make extravagant claims for their model, which reflectbadly on the whole field of learning st yles research.Rita Dunn, for example, whose approach was evaluatedin Section 3.2, is quoted by ONeil (1990 , 7 ) as claimingthat Within six weeks, I promise you, kids who youthink cant learn will be learning well and easily The research shows that every single time you uselearning st yles, children learn better, they achievebetter, they like school better.

    In a similar vein, Felder has written art icles on therelevance of learning styles to the teaching of scienceto adults. After examining four different models theMyers-Briggs Type Indicator, Kolbs Learning StyleInventory, Herrmanns Brain Dominance Instrumentand his own Felder-Silverman instrument he concludes(1996, 23): Which model educators choose is almostimmaterial, since t he instruct ional approaches t hatteach around the cycle for each of the models areessent ially identical. We disagree strongly: it matterswhich model is used and we have serious reservationsabout t he learning cycle.

    For other commentators, t he absence of soundevidence provides no barrier to basing their argumentson either anecdotal evidence or implicit suggestionsin the research. Lawrence (1997, 161), for instance,does exactly that when discussing the detrimentaleffects of mismatching teaching and learning styles.More generally, the advice offered to practi tionersis too vague and unspecific to be helpful; for example,restructure the classroom environment to make it moreinclusive rather than exclusive. The quality of advicegiven to new post-16 teachers can be gauged byexamining one of t he leading textbooks (Gray, Griffinand Nasta 2000), where the topic of learning stylesis dealt with in t hree pages. The authors advocate,without justification, Honey and Mumfords fourlearning styles (see Section 6.2 ) and then refer theirreaders to the practical manual on learning stylesproduced by the Further Education DevelopmentAgency (FEDA 1995). Typical of their unproblematicapproach to learning styles is t he claim that a criticalpart of a carefully-planned induction is to makean accurate assessment of each students unique

    learning styles (Gray, Griffin and Nasta 2000, 197).In sum, clear, simple, but unfounded messagesfor practitioners and managers have too often beendistilled from a highly contested field of research.

    Section 8

    Implications for pedagogy

    page 118/ 119LSRC reference

  • 8/8/2019 FrankCoffieldLearning Styles

    128/182

    Yet even among critics of research on learning st yles,there is a tendency to write as if there was only onemonolithic movement which was united in its thinking;in contradistinction, this review has presented a widespectrum of theoretical and practical posit ions ona continuum, consisting of five main families orschools of t hought (see Figure 4, Section 2). Bloomerand Hodkinson (2000, 5 84), for instance, argue thatt his literature proposes that learners possess relativelyfixed preferences and capacities for learning [and] itseldom explores the extent to which, and the conditionsunder which, preferences change. This cri ticism appliesonly to those theorists who emphasise deep-seatedpersonal t raits at t he extreme left-hand side of thecontinuum, but is not relevant to t he clear majorityof learning st yle theorists who are concerned to improvestyles of both learning and teaching. Bloomer andHodkinson are simply wrong in claiming that most

    theorists treat learning styles as fixed.Bloomer and Hodkinson (2000) make, however, a moreserious criticism of the learning styles literature to theeffect t hat, even if they are prepared to accept thatlearning styles exist, they consti tute only a minor partof individual dispositions which influence the reactionsof learners to t heir learning opportunities, whichinclude the teaching style of their teachers. Are thesedispositions anything more than Entwistles (1998)orientat ions and approaches to learning; or arethey a broader concept? To Bloomer and Hodkinson,dispositions are both psychological and social; by thelatter term, they mean that disposit ions are constructed

    by the contexts in which people live and are not simplypersonal reactions to those contexts. Moreover, thesedispositions are said to be wide-ranging in coverage,interrelated in scope and help to explain the strongreactions which many students have to the cultureof different educational instit utions. (See Ball, Reayand David 2002 for more research on this issue.)Dispositions would appear to be tapping contextual,cultural and relational issues which are not pickedup by the learning style instruments of Entwistle (1998)or Vermunt (1998).

    The strategies which follow are treated separately,but in practice, t hey tend to overlap and theorists often

    advocate a judicious selection of approaches ratherthan an exclusive focus on just one. Furt hermore,because we have adopted the stance of treatingteaching, learning and assessment as one interactivesystem, we avoid the temptation to deal with strategiesfor students separately from strategies for t eachers,tutors or managers.

    Increase self-awareness and metacognition

    A knowledge of learning st yles can be used to increasethe self-awareness of students and tutors abouttheir s trengths and weaknesses as learners. In other

    words, all the advantages claimed for metacognition(ie being aware of ones own thought and learningprocesses) can be gained by encouraging all learnersto become knowledgeable about their own learningand that of others. According to Sadler-Smith(2001, 300), the potential of such awareness liesin enabling individuals to see and to question theirlong-held habitual behaviours; individuals can be taughtto monitor their selection and use of various learningstyles and st rategies.

    Moreover, as Apter (2001, 306) suggests, anunderstanding of the various elements which producedifferent states of motivation in different contexts

    can allow people to come more in control of t heirmotivation and hence of their learning. Learners canbecome more effect ive as learners if they are madeaware of t he important qualities which they and otherlearners possess. Such knowledge is likely to improvetheir self-confidence, to give them more control overtheir learning, and to prevent them attributing learningdifficulties to their own inadequacies. The upshot couldbe that s tudents and teachers choose the strategymost appropriate for the task from a toolbox ofstrategies (Adey, Fairbrother and Wiliam 1999, 30).Kolb (1999, 5) neatly summarises the advantages ofthis fi rst st rategy as follows: Understanding yourlearning style type, and the strengths and weaknessesinherent in that type, is a major step toward increasingyour learning power and getting the most from yourlearning experiences.

    One option is to leave students to diagnose their ownlearning style so that t he responsibility for learningis passed to t he learner. But Merrill (2000) argues thatmost students are unaware of their learning stylesand so, if t hey are left t o their own devices, they aremost unlikely to st ar t learning in new ways. Herrmann(1989) places some emphasis on the understandingof individual learning styles as a start ing place fordevelopment, and as a flexible response to life changesand needs, but the popularity of a model can leadto oversimplist ic generalisat ions. For example, theMyers-Briggs Type Indicator, which was intended toenable individuals to explore the interactions of theelements which make up personality type dynamics has so far entered popular consciousness t hat si tesexist on the internet advising (for example) ENTP(extrovert, intuit ive, t hinking and perceptive) individualsas to which other t ypes would make their idealmarriage partners. Hence, the need for dialogue witha knowledgeable tutor who understands t he learningstyles lit erature as a whole and has a critical feel for i tspotential and pit falls. Such a tutor is likely to pour coldwater on, for example, the ext ravagant claims made by

    Gregorc (1985) that serious, individual study of learningstyles will reduce naivete [sic], increase personalresponsibilit y for thoughts and actions, and improveyour relationships.

  • 8/8/2019 FrankCoffieldLearning Styles

    129/182

    Serious in-depth study of such matters is notadvocated in guidance for new teachers. For example,Huddleston and Unwin (1997, 72) define learningstyles as study skills and transit ion from one styleof teaching/ learning to another; and advocate, withoutany explicit rationale (like Gray cited earlier), theuse of both Kolbs LSI (Section 6.1) and Honeyand Mumfords LSQ (Section 6.2), neither of whichare unproblematic, as our earlier evaluat ions showed.

    In these debates, t he research of Entwistle (Section 7.1)and Vermunt (Section 7.2) is valuable because, asdiscussed earlier, they have shown that attent ion needsto be given not only to individual differences in learners,but to the whole teachinglearning environment.Both have demonstrated that while the motivations,self-representat ions, metacognitive and cognit ivestrengths and weaknesses of learners are all keyfeatures of their learning style, t hese are also a functionof the systems in which learners operate. A central goalof their research is to ensure that lecturers can relateconcepts of learning to the specific conditions in whichthey and their students work that is, i t is the wholelearning milieu that needs to be changed and not justthe learning preferences of individuals.

    A lexicon of learning for dialogue

    Learning styles can provide learners with a muchneeded lexicon of learning a language withwhich to discuss, for instance, their own learningpreferences and those of others, how people learn andfail to learn, why they tr y to learn, how different peoplesee learning, how they plan and monitor it , and howteachers can facili tate or hinder these processes.Through dialogue with a tutor knowledgeable about therelevant li terature, t he students repertoire of learningstyles can be enhanced in the hope of raising theirexpectat ions and aspirations.

    Students can be taught, for instance, which of the71 learning styles are well founded and which arenot, and when and how to choose the most appropriatestyle. Similarly, tutors can be helped to understandthat what they may have been categorising as lazy,unmotivated or t ruculent behaviour may be causedby a clash in learning styles between themselvesand students/ colleagues. Even some of the f iercestcritics of learning styles concede that a par ticulartest can be safely used as a means of facilit atingdiscussion about learning (Reynolds 1997, 126).As a result, some practitioners use the topic of learningstyles simply as a motivational ice-breaker, as a meansof warming up t he class, or as an acti vity-basedintroduction to t he topic of learning.

    For students, particularly those who are less confidentabout t heir learning, t he acquisit ion of a new vocabularywhich they can use to describe and explore their ownbehaviour can be an immensely motivating and positiveexperience and has the potential to help them to reflectand develop their crit ical thinking. However, this isdependent both on the quality of the experience of usingthe learning styles instrument and on the nature of thefeedback. In this respect, Jacksons LSP (Section 5.3)emerged from our review as a particularly good exampleof feedback in which traits are described but individualsare not labelled, and the caveat t hat st yles arecontext-dependent is frequently repeated. Respondentsare given areas of st rength and weakness to focuson, but are urged overall to consider the goal of thetask to be accomplished and to be strategic in their useof their talents.

    One of the values of Honey and Mumfords workis that it is primarily aimed not so much at st udentsin education as at managers and trainers who wishto improve the learning of their st aff by mea