Notes on: Pask, G.  (1976) 'Styles and Strategies of Learning'.  In British Journal of Educational Psychology, 46: 128-48.

Dave Harris

Conversations require conversational domains with entailment structures and behaviour graphs [see earlier article].  Ideally, each topic should have its own behaviour graph, describing what is to be done to model or explain a topic.  Learning a topic is an essential requirement for the sort of depth processing advocated by Marton and Saljo (M and S).  We can define a concept as 'the procedure for realizing a topic relation', and memory as 'a procedure for reproducing a concept' (128).

A number of experiments are reported.  Some measured learning styles according to retention tests, and different teaching strategies and learning strategies were compared [see some possible examples here].  Understanding takes place when a lecturer gives an explanation and a derivation [from other concepts].  This means that retention is a good test of understanding.  There are different learning styles [holist and serialist ones, described in the earlier article], and some more detail is given of how the strategies produced different sorts of descriptions of networks [the task is learning the taxonomy of Martian fauna again].  Matching different teaching styles to particular learning strategies is desirable, and 'mismatched students acquire hardly any relevant knowledge' (132).  However, sometimes students feel that they ought to use a particular approach, and 'there is a strong institutional bias to structure material [in a serialist way], and most examinations favour serial recall'.  However, the holist approach resembles M and S's deep approach (132).

There is a hard core of evidence on the importance of learning styles. Bruner, Guildford, and Kagan have all developed work on the difference between impulsive and reflective styles, for example, while Within et al. have talked about the differences between field dependent and field independent approaches [others are summarized on page 132]. The holist or serialist dimension reflects a strategy rather than a style, however, and can be exhibited only where there is a 'strict [logically structured] conversation'.  In normal conditions, holists appear as 'comprehension learners', while serialists appear as 'operation learners' (133), and some students can be both—'versatile'.  Comprehension learners easily pick up the overall picture of the subject matter, and can see, for example, redundancies in a taxonomic scheme.  Operation learners pick up on rules, details and methods, but are 'often unaware of how they fit together', which can produce arbitrary or accidental learning procedures.  Both need explicit procedures.  Both are necessary for understanding.  So everyone needs expert diagnosis and teaching.

Further experiments on the taxonomies are reported 133f, including more details of TEACHBACK  (see earlier piece) , where students are required to explain and give an account of their learning procedures.  These are then content analyzed.'Falsehoods', inventions, requests for information, deductions, irrelevant statements and redundant statements are all identified.  Those groups that featured matched instruction did better.

Other experiments have described using other procedures.  One tested 'real life' comprehension or operational styles.  Students were required to recall the history of a spy ring [another taxonomy?].  65 students from Henley Grammar school, 40 students from the Architectural Association School, and another 50 students from various schools and colleges were recruited.  It's possible to describe some of their understandings as 'deep' [actually, for six of them], and this approach can also be described as involving description and procedure building [with this lingering division between logical and empirical truths].  Insisting on both of these building procedures avoided the 'pathologies' of either approach [not surprising because understanding has already been defined as including both, with holists good at description, and serialists good at procedure].  By contrast, if explanations are to be tested by multi choice exams, even a 'surface' understanding will do (143).

Versatiles performed well, but if they were placed in self study conditions, with little conversation [ on the basis of the structured conversational domain explained earlier] , explanation was often 'replaced by a correct response condition' [that is, they try to work out what the right answers were?].  In self study conditions, holists get the right answer but can 'globe trot', adding in all sorts of irrelevancies.  Serialists also get right answers, but these can be 'improvident', that is 'no longer supported by an adequate descriptive scheme' (143).

There is an additional benefit to properly organized conversational learning, since it permits a transfer from one subject matter to another.  This is because the learner learns about their own mental processes.  However, conversational learning is not like ordinary learning, which has already been structured [by the subjective conventions of academics?].  Learning to learn involves both versatility and the ability to structure material.  This was tested out with a sample of 24 6th form  and college of education students: they were given some information and then tested on list of concepts, by being asked to pick any that they felt they could explain, and then asked to join any connections.  Their decisions were then discussed through 'a non technical presentation of principles derived from CASTE' (144).  Then they were retested.  Then they moved on to the task involving the spy ring.  The results do show improvements in performance, although the versatiles were good anyway.  'Important practical implications for education are beginning to emerge' (146).

Back to education studies