-,

•,
,
roN , ',-
, '.
•
.
/lJ/g.
,
and --,
:•
,
1(1('(.'1/

-hol-

" N.
I
,
•
CHAPTER 4
.
L' miC

a nd
~c. In
I
,
,,
Learning Strategies,
(p p,

lbur-
•

•
Teaching Strategies, and
.
HU1:=> .
•,
,
Conceptual or Learning Style
'nsls.
, ,
Eds.) ,I GORDON PASK "
I
re nee J,
,

j
~ nsis.
.) Till'
I
, INTRODUCTION AND OVERVIEW
){ ion,
I, 73-
1
,
J From the mid- to late 19605, Brian Lewis, Bernard Scott, and I conjec-
'liO/wl
rI tured that learning strategies, teaching strategies, and even plans of
,,
, action have characteris tic types which can be differentiated (Lewis &
\a th.'t Posk, 1964, 1965; Posk, 1961 , 1970, 1972; Pask & Lewis, 1968; Pil sk &
Scott, 1971, 1972, 1973), Individual difference psychologies hilve main- . ~:

silatis
tained a similar stance and with greater precision regarding the nature of
. 2(J~-
f s trategies, An overview of the approach take n by my own group in the
19605 is described in the rema inder of this sec tion . Learning and tcnch- ••

ing strategies can, under appropriate circumstances, be substantially

f exteriorized or externalized for observation. Protocols can serve this
purpose, but we used maplike representations of what may be known or
1
,,
learned. These represen tations were open to continuous evolution as
further topics and relati ons between them were added by learners,
I
Later, these representations were seen to be manipulable systemat- ,
ically and without the imposition of rules that insult freedom of thought
or creativity (about mid -1970s) . The maps and representations of topics
f (communicable, shared, or public concepts, rather than personal con-
i,,
GORDON PASK • OOC Program , Faculty of Education, Univcf!'ity of Amsterdam, Grote
t
I
l3ickers traat 72, Amsterdam, The Nctherl.:mds 1013 KS.

I
•
83
l
I
t
"
!
 •

I,
, 1
,'.• ••
84 I STRIi7TGI ES iii'
- I
• GORDON I'IiSK

,
•
• Cl'pts) were called "entailment structures." Later still, Kallikourdis,
I consist of sin
Scott. and I discerned that entailment structures are special and re- • mayor may I
•

I s trictive cases of "entZlilment meshes." in which shared, public concept creation of al

relations constitute expressions in a "protologic" (Lp) (Pask, 1975a, b; , A slyle is
1976 n, b, c; Pask & Scott, 1973; Pask, Kallikourdis, & Scott, 1975). class of teacJ
An entaillllellt stru cture consists of topics and connections among school, this is
I, verbal comm
topics which show how they may be derived or understood from other •

to pics . The entailment lIIesh recognizes that the entailment of one topic to whom do.
frol11 others is a momentary situation that occurs during action or exp la- psychologists
,
nation; in fact, the relationships between topics are /lot static and hier- mOllS unit "0
1, archical. " Lp" refers to a prolologic developed out of Conversation Theo- of strategy is .
ry which is a model of what underlies cognitive activity. It is not a model tive is needec
i
of logic or language itself but rather a "substrate" for them; hence the offers one SlII
vidual (a PS)
qualifier, "proto," meaning "underlying." Its rules involve the pro-
cesses by which, for example, the normally heterarchical relationship
, mous. The P-
between topics unfolds into hierarchies. Its details encompass conflict open 10 lhe i
detection and resolution, analogy, generalization, and models of inno- • tween person
vation and memory . our sense, P-
•
COllvcrs tltion Theory is a summarization of Ollr assumptions and I "P-Indiv
r<1tionale from thi s early period. Convcrsntions arc bchllviors, but spe- • it illlows Oile
, can be I
ci<1i kinds of beh<1viors with hLlrd-valucd observablcs in the form of 111<l1l)'
concept sh;'lrings, detect l~d LI S " understandings." Conversations Lire, we possibly conn
bL'iicvl', the first b<1sic d<1t<1 of psychological, socinl, or educal ion<ll theo- of 11l<lI1Y pefs
•• ry. We sec later thilt people can even have conversatiuns with them- The tech
•
I selves. Conversations which may lead to concept sharing need not be , tional opennl
,
•
verbal. Often they are gestural , pictorial, or mediated through a comput- of this paper
,•
u nderlie the t
! er intl'rf(lcl'.
, specific indic,
• Lllltfcr~ / l1IldillS , like many of the terms of COl1versntiull Theory, h<ls
•
• an everyday use that is evocative of its rigorous and almost technical review in Pas
, i
,,J meaning within Conversation Theory. An understanding involves not • netic, matheJ
, only the topics that are related and their relationship, but the ability to quite familiar
,
•

• tra nsfer and apply the relationship to new situations. An "agreement ! understandin
,· I
, ov cr an understanding" is a hard-valued event that can be detected in "what," "ho'
~ mutually soli
I ZlIl experiment. It invulves individuals, each of whom exteriorizes his or
.. . •• ,
,
•
her understandings (lnd confirms that the other's entZlilments reproduce ,, slightly rdim
•
I hi s or her own, previously internal, concepts. ing oct. Of Ct
I•
, The entailment mesh constitutes a network or n1<lp of topics which and often is
! have no hief<lrchy or direction. Paths on th is m<lp arc <l le<lrni ng str<ltcgy
, "knitting" 111,
,
or i.l teachin g striltl'gy. I\. learning strategy or il teaching str<llcgy consists ta ken to i nell
of p<llhs on till' ll1<lp ll1<1rkcd to indicate specific conceptual events like <Inti how we
examining, trying to learn about, learning about a topic (shared con- sharing, then
cept), or understonding a topic in the related context of others from cept of "tortl
which it ma l' be cons tructed , reconstructed, or recalled. Such paths may ,
1
alogue. Furtl

. "' .. '
 •
>

,
I
I,

,I
CDON I'ASK STRIITEGIES AND CONCErTlllIL STYLE 85

likourdis , consist of simple trees, a hierarchical branching and looping, and they
I and re- mayor may not employ analogies as structures. They often involve the
c concept creation 6f analogies, and they may employ or create genera lizations,
1975a, b; A style is a dispositiol1 to adopt one class of learning strategy or one
975), class of teaching strategy in the conversation of a tutorial. In an art
lS among school, this is usually by demonstration, In a high school, it is usually by
'om othe r , verbal communication or laboratory experience. The question remains,
one topic to whom does this stylistic disposition belong? Individual difference
OT ex pla - psychologists seem to take it for granted that the unitary, partly autono-
and hier- mous unit "owning" a disposition and opting for one or the other class
ion Thea- of strategy is a person, Of cou rse, this ma y be so, but a broader perspec-
II a model tive is needed if we (Ire to make sense of the facts. Conversation Th eory
ll~n((: lhe offers one such broader perspective, In this theory, we speCify an indi-
the pro- vidual (a psychological individual, or P-Individual) as pZlrtly autono-
ationship I mous, The P-Individual is only partly autonomous because he or she is
oS conflict open to the information transfer of a concept-sha ring conversation be-
-; of inno- ,, tween persons or between mental organizations , in one person (both, in
our sense, P-Individuals),
tlans and "P-Individual" may appear clumsy, but the idea is essential because
,
, but spe- I
it allows one of the symmetries of Conversation Theory to exist. There
, form of can be many P-Individuals within one person (when I take different and
I S are, we pOSSibly conflicting viewpoints) as well as a P-Individual that is made up
lIlal theo- of many persons, such as a school of thought or a religion,
• The technical criteria of "organizational closure" and "informa-
ith them- I

..:d not be tional openness," underlying this distinction, are hardly in the province •

I compu t- of this pilper. Likewi se, there are very thorough technicill criteria that
, underlie the theory's usa ge of the term IIl1rlcrsffillrlillg, ill ready noted as il
eary, has
,I specifi c indicator of concept sharing. The interested reilder will find a
technica l review in Prisk (1983), which <lIsa g ives reference to expositions in cyber-
alves not netic, mathematical and logical terms. The bilsic ideas are, however,
ability to quite familiar in ordinary language. A P-Individual is a personna; an
greernent understanding is the ability of two or more P-Individuals to exchange
, leeted in "what," "how," and "why by that method" questions and to provide
zes his or mutually satisfactory replies. Such a transaction, which embodies the
cproduce • slightly refined common meaning of understanding, is a concept-shar-
I ing act. Of course, the agreement over an understanding need not be
ics which !
and often is not complete; for example, your concept of "tortoise" or
~ strategy I "knitting" may be quite different from mine. Hence, agreement mllst be
y consists taken to include agreement to disagree, with some knowledge of why
vents like ,
,
and how we disagree. If there is some agreement and some concept
ared con- sharing, then you r personal concept of "tortoise" and my personal con-
lcrs from cept of "tortoise" are mutually enriched by whatever is shared in di-
lat hs may alogue. Further, a s hared o r publiC cO llcept of "tortoisl'," SOIlll' of it

,,
I,
I,
 I

,•
I
GORDON PASK STRATEGIES AND CONCEPTUAL STYLE 87

d may be in- like a pianist playing a chord. Also, thes~ tendencies are related in the
, domain of coding skills (Pask & Scott, 1971) to the s trategy preferred and
Jl'~crlrtion of t most effectively used in building up a complex response.
• •
lIZJng or con-
riving it from
I
Id meet with
re necessarily
I,
•
INTELLECTUAL LEARNING STRATEGIES

The first studies of intellectual learning took place in 1969 and are
P-Individuals I
I reported in Pask and Scott (1972) and summa rized in Pask (1975a). In
y, 1955, and
·o re-co nstruct
I order to perform experiments of this kind, it is necessa ry to provide
unfamiliar material to be learned. For this purpose, Dr. Scott produced
taxonomies for two different types of "martian animal," the "elobbit"
ext o f educa- and the "gandelmuller." He provided graphic and written d escription s
cessari lyexist I of each species: 40 cards in all, of which 10 showed typical subspecies;
IS of rational
I,
15, contextual data about habits; 5, test types in the taxonomy; 7, phys-
ispositions to ical characteristics; and 7 to indi cate why names of parts or behaviors are
lat is, styles, ,
I used. Under free-learning conditions, the subjects could select cards as
:!.y are charac- I they wished and were required to give a tape recorded account of why
ed, they may anyo ne was selected.
,
llJUnd, rather Upon analysis of data from about 26 subjects, it was possible to
s of this vol- I discriminate subjects who tested large, global predicates. In all cases the
nso nant with subjects satisfied a criterion of 10 or more on a sta ndard 30-item tax-
onomy-based questionnaire administered when subjects thought them-
selves proficient. In addition, half of the selected subjects were requi red
to teachback the information to a participclnt experirncntcr prior to COIll-
TS pleting the qllcstionnnire. Teachback is an iterative process of explanation
(of each species and its proper discrimination), correction, and furth er
,'sa lion Theo- explanation. In comparative studies, teachback is replaced by s imilar-
obtoi ned by seeming but ineffective procedu res.
per concerns
I
The learning strategies are called "olis!, preferring global predicates
~ £ltory and in and relations of topi cs, and serinlist, preferring not to use such relations
experiments, and learning step-by-step. These are paralleled by teaching strategies of
>s ible shou ld 1 each type . One qualification is required. The holist learner may be either
of datil, that I irredundant, eschewing redundant data (often known as enrichment
I data), or redundant and apt to use such data. Generally, the redundant
I
holist learns less rapid ly, takes considerably longer, and has faster but
I less-accurate recall of any tutorial material. The irredundant holist is
I
I more selective about the material used in teaching. Schmeck (1983) re-
at in percep- fers to a redundant holist as an "elaborative processor."
,nSf learning
I
lsk & Lewis, I
tt-by-compo- MATCHED AND MISMATCHED LEARNING

:riminate the It is easy to design and use a holis t or serialist type of teaching
,sub jects oct prog ram to guide a participant experimenter. This was done, with clob-
. .

..
I -
 •
•

88 C(JI~{)ON /'IISK
I
• STl~A'lLCltS
•
bits first and gandelmullers second, balanced against "gnlldelmullers" jed altritio
first <llld "c1obbits" second . In (";"jeh easc , the instruction WilS iter<11i vp • pIc con tai n
until complete mastery was secured. After that, subjects responded, "'lu"lily of
•
several hours later, to test questions. cerned wit
•
The matched consistently performed better than the mismatched. I (i.e., the cI
For example, the mismatched holists needed 4 to 7 iterative repetitions I derstandin
in contrast to 1 to 3 for matched serialists. Also, learners performing ••• operon cye
teach back are superior to learners with the ineffective control treatment the influen
regardless of whether they are serialists or holists. These results are, if , Of the
iln y thing, strengthened by later, detailed replications wh ich also sub- to determi j
stantiate test-to-test reliability of the "clobbits" and the "gandelmuller" subjects en
,
s ituations when their order is alternated (interscore correlations of .89). • in a match·
In common with all such methods of exteriorizing concepts, any were taugl
teachback process is likely to occupy 3 to 4 hours of effort. not counting and finally
•
the subsequent ana lysis of detailed data. Attempts to red uce the labor given inefl
involved, by automat ic control of administration and automatic record- teach back <
ing, have been used with varying degrees of sophistication since the I
tive teachb
early ]9705. However, in order to be effective, the techniquc must be lers and ef!
based upon a firm sct of principles and, in the case of our group, those ,• were unre ~
of conversation theory. • sions, but I
To summarize our findings at this early stage, the condition of
, grammed t
, understanding of a shared concept was obtainable by iterated teachback, ,• In both gro
i and the resilience of an understood concept was affirmed by the superi- had been s
~., presented
or retention of teachback learners. It was possible to specify learning
j
J s trategies and teaching strategies as path markings (or classes of them)
, The fu
•
:i, on a data representation, although, at the stage concerned, it was diffi- style indict
I cult to compare or refine them. Even so, the existence of d istinct classes •
I
most eflect
of strategy was suffiCiently evidenced by the data in question. Strategic • measures:
1 match or mismatch showed an influence upon learning-mismatch as to their
when leaf!
,~
leading to difficulty in understanding and sometimes to complete mis-
....•
.
,,
••
understanding of relevant topics. The converse result, that matching
,
I subject ma
certainty, s
improves learning, was suggested, but it is less convincingly demon-
. •
.1 I
.. ,1• testing a s
.' '

strated because of an inherent ignorance of what the learner knows to

• begin with. These statements could, at this stage, only be demonstrated
for the persons who participa ted, learners or teachers, but studies uf the
I
•
tional hyp<
free recall <
act of design suggested that the statements had greater generali ty (e.g., certain top
that designing can involve distinct P-Individuals in a single person). • mirrors thl
,
,.'~
• further SOl
style were
!., i were inflUt
,• A TYPICAL LARGER-SCALE SERIES OF EXPERIMENTS
I or ineffect i
J••
• tests (emb
,.•
•
A series of experiments, using partly automated methods, were
i logical pro
carried out using 62 participants (who completed the series) from poly-
technics of Kingston and Chiswick, United Kingdom. To indicate sub- although I
•
• I
I•
•
I
,I I
I
 .. ....
.
. '

!
j
•
I
DON PASK I
•
STRATEG IES AND CONCEPTUAL STYLE 89
•
Imullers"
iterative
I
t
ject attrition in such detailed and lengthy experiments, the original sa m-
ple contained 115 people, all selected by pretest in order to obtain rough
;ponded, II equality of preferred strategy difference. These experiments were con-
,, cerned with severa l issues: the means of determing learning s trategy
·llatched. (i.e. , the clobbit and the gandelmuller tests), the effect of securing un-
-pe litions derstanding by teachback, s tability over different subj ect matters (the
rformi ng ~ operon cycle, the menstrual cycle, and an inductive inference task ), a n d
I
rcalment , the influence of matched versus mismatched learning strategies.
Its are , if ,• Of the participants, 32 were exposed to the clobbit fre e-learning task
, to d e te rmine preferred learning strategy (holist or serialist). The n, the
also sub-
, Imuller" , subjects engaged in a further session to p erform the gandelmuller task
in a ma tched or mismatched teaching program . Two-weeks later, they
,s of .89).
' pls , any I were taught the operon cycle in matched and misma tch ed conditions,
cou nting
th e labor
I and fin ally, they were given retention tests and teachback. Partici pants
give n ineffective control gro up teach back on clobbits received effective
c record- tea chback on ga ndelmullers and on the operon cycle. Those given effec-
sin ce the ti ve teachback on clobbits received ineffective teach back on gandelmul-
t,
must be lers and effective teachback on the operon cycle. All learning conditions
IP, those
I were unrestricted . The rema ining 30 students experienced similar ses-
I
I sions, but the free-learning task was the mens trual cycle, and the pro-
cli tion of ,i
grammed ta sks we re the operon cycle and the inductive inference task.
'ac hback, ,, In both groups of participants, hal f of those receiving effective teachback
c superi- I
had been selected and assigned as holists and half as serial ist; half were
IC<lrning I, presented wilh match ed programs and half, mismatched .
or them) The fuJI experimental design nnd data for the development of the
NOS d iffi- I style indi ces appear in a series of re ports (Pask, 1977, 1978, 1979) . The
ct classes
,,• most effecti ve s ty le assignment index consisted of the following set of
St rategic I mea s ures: the freq uency of naming, the intentions expressed by learners
lliSIllJtch to their gOil) for lea rnin g, and the details of the free-learning d ata
(1S

.Jete mis-
I, when leilrners were directing their own path s (e.g" exploration of the
natc hing I s ubject matter), deliberate information search based on cu riosity or un-
certainty, search for one specific datum, search for several speci fic data,
demon-
~ nows to
,
t
tes ting a single-predicate hypothesis, testing a large predicate or rela-
'n strated tional hy pothesis . Some additional indices are obtained in sessions of
les of the free reca ll administered by the apparatus, notably when lea rners re peat
ily (e.g ., certt'lin top ics and the extent to wh ich the order of top ics in free reca ll
,"son). I mirrors the order in which topics were covered during learning . As a
I further source of discriminative data, retrospective determination s of
I style were carried out after teachba ck a nd reca ll. Most determinations
rs I,
· were influenceC: by whe ther or n ot the previou s teach back was effective
,, or ineffective (in the latter case, n o explanations were obtained) . Other
Js, we re ,
•

•
tests (embedded figures, the "circles" test, the "analogies" test, a nd
.un poly- I log ica l problem solving) we re cilrried Oll t, but yielded corrcliltions that,
: ale sub- although marginally significant, are modest.
I
 "
90 GORDON PASK STRATeGIES AN
•
•

Seri<lllenrners showed intention to search for specific data . Holist may be derivl
learners showed intention to test a large predicate or relational hypoth- and it shows '
esis. Holist lea rners selected a large number of data cards, while se- na tion or acti
rialists selected a s miJII number of data cards. Matched performance was pIe from Can
better than mismatched performance. Matched task scores were better using the in:
than mismatched task scores. Teachback performance was better than when used a
ineffective, control group teachback performance. understandil
In addition, confidence was measured on the following dimensions: However, (a
uncertainty about hypo theses, correct belief, uncertainty about the next equally valie
topic to select, and belief in correctness of what to select next. These made upon
~
components are a multidimensional index of doubt. For holist learners compass.").
"
uncertainty about the next topic to select and belief in correctness of Thus, the el'
what to select next were significantly higher than the others. For opera- arrowS and
tion learners, correct belief was significantly higher than the other "chemistry"
dimensions. ic" is one fl
Perhaps the most puzzling, but also the most consistent, finding is not only sa )
the constancy of strategy type over all these tes ts as well as different-in- also how tl~
content Icnrning s ituations. The enigma lies in the invnriancc uf I'CfSUllol merely to p'
style, rothcr than P-Individual style. We should predict change in per- that shows \
spective with learning experiences fur different learning contexts. In ments ilnd 1
fact, these experiments started with a half-and-half split in each group, CASTE v
and there are in the 32-member group, 18 serialist and 14 holist learners . portable, Ie
In the 30-member group, there are 17 serialist and 13 huli,t lcorners. could be en
Further, this constancy is manifest at all of the sampled interrnediary sentl1lion ul
phases . This was perplexing in that we approached the whole project menls. Alsl
with the assumption that context, or contextual demands, would far m eiosis a11 1
outweigh stylistic consistency as a determiner of choice of strategy, that gines seen
is, we expected less consistency in an individual's choice of strategy. For the Tse mb,
more complete descriptions of these projects, the reader is referred to eases of th l
- .. "~
Pask (1977, 1978, 1979, 1983), and de Gelder (1981). These
uf m<ltcrial:
completed
THE USE OF "CASTE" AND OF "INTUITION" AND THE NOTION experimen l
OF STYLE sessions be
original fin
Just before 1973 a computer-regulated interface for exteriorizing do adopt '
conceptual events became operational as a laboratory exteriorizing tool. strategIes
It is known as CASTE, or course assembly system and tutorial environ- effective II
••
ment. Further, the "conceptual maps" of topics Pask (1975a,b) had been silu<:ltion.
replaced by sophisticated, less-arbitrary forms of cartography. The rep- As en
resentation employed was a form of entailment mesh. criminate,
As before, a cOllceptl/a/11lap is an entailment structure, showing hier- and their
• archical connections (directed arrows) which indicate how the topics termined

-e
 • •

" WON PASK STRATEGIES AND CONCEPTUAL STYLE 91

aLa. Holist may be derived. The entailment mesh has relations rather than arrows,
al hypoth- and it shows how particular entailments may arise for purposes of expla-
while se- nation or action. For example, J can explain (to use the very tired exam-
llance was ple from Conversation Theory) a circle using a compass and a plane ("By
'cre better using the instrument of a compass, which works like such and such
Idler than when lIsed on npinne, 41 figure called a circle is made"). This shows how
understanding .n circle might entail the topics of compass and plane.
mensions: However, (and this was the next step from entailment structure) it is
It the next equally valid to explain a compass from circle and plane (" A circle is
~xt. These made upon a plane by the manipulation of this instrument called a
.::;t learners compass."). And again to explain a plane from circle and campass.
ectness of Thus, the entailment mesh was developed, which does not contain any
For opera- arrows and allows more of the everyday complexity, flexibility, and
the other "chemistry" of understanding to be represented. An " understood top-
ic" is one for which teachback has been obtained. Teachback involves
finding is not only saying what other topics that topic may be derived from, but
fferent-in- also how those topics relate to reproduce the topic. It is not enough
a f persollal merely to parrot back a definition; the topic must be used in a novel way
ge in per- that shows the learner's ability to manipulate the topics and their entail-
ntexts. In ments and produce a novel interaction.
ch group, CASTE was a laboratory installation that occupied two rooms. But a
t lcarners. portable, less-powerful facility called INTUITION was developed that
learners. could be employed in schools and colleges, thus allowing for the repre-
-' rm ediary sentation of different types of learners in different institutional environ-
Ie project ments. Also, programs were designed for many subject matters, such as
Nould far meiosis and mitosis, probability theory, physical chemistry, heat en-
tegy, that gines seen in the context of thermodynamics, the reign of Henry VIlJ,
ll0gy. For the Tsembaga ritual regulation cycle (an anthropological subject), dis-
,ferred to eases of the thyroid system, and many others.
These studies lasted for 6 years or more and covered an admixture
of materials and types of learner, about 60 in the laboratory and some 95
completed (out of over 170 initiated) in colleges or schools. Typically,
: NOTION experiments with detailed, meticulous recordings lasted over spaced
sessions between 7 and 12 hours each. These studies confirmed all of the
. '.
original findings. For example, "understood" topics are resilient, people
~no nzlng do adopt strategies of learning, a mismatch of learning and teaching
":ing tool. strategies leads to no relevant learning, and matching leads to more
enVlron- effective learning than that achieved in an arbitrarily programmed
had been situation.
The rep- As entailment meshes became recognized, it was possible to dis-
criminate, with greater refinement, between types of learning strategy
'ing hicr- and their relation to inherent structures of subject matter, whether de-
l e topics termined by an author's predilections, institutional constraints, or the

•
I
1••

!
I 92 GORDON P/I SK STR/ITEGIES /I

I'-
subject mot ter itself. These dis tinctions, refined by using CASTE and matter. The :
j INTUITION, are more clearly related to the content of this paper later rule s and ti-
! when dealing with those strategies influencing dispositions called
"s tylcs. "
!
I
previou sly r
specified (i.e
Some disti nctio ns, however, merit attention at this point. For exam- with pure (i .
ple, among alI holists, it is possible to recognize (a) those who depend s till s uggest:
upon the use of valid analogies or generalizations given within the mate- misleading.
rial , that is, those having both contrast and similarity; and (b) those does involve
holists who depend upon the creation of valid analogies and generaliza- Ii and global n
tions between topics. This la tter method incidentalIy seems to be one of • understandi l
the mos t productive ways of learning-by invention or discovery . These
• valid analogies are distinct from mere similitudes and overgeneraliza-
j tions that do not contain contrasts and which are commonly found to be TE~

Ii unproductive . The analogies that learners set down may be either valid
or positively misleading. In the context of " hard to learn" subject mat- I,
,
Soon afl
ters, counterproductive holistic learning gives risc to casual g/oIJc:lroflillg, the need for
I a lea rning pathology. to predict "
Teachers, authors, or experts do much of the sa me (i.e., analogize certain indiv
productively or not). As a matter of fact, in the CASTE environment, the covaiiditY,1I1
nUl hor- tc;:lchcr-lcarning dis tinction becomes only olle of dominant ing topics (5
, ro le . D uring the creation of the subject matter representation, clearly the They both n
autho r is dominant. Then, CASTE adopts the role of teacher wh e never it motic materi
'presen ts material based o n this underlying re presentation . However, Bruner's (19'
the learner must teach the teacher what is not known , as well as certain quired of n
aspects of hi s or her learning style . Also, the learner adopts the teacher's
role in order to teach back the concepts to CASTE to confirm understand-
I learned.
One tesl
ing, and fin ally, in its most full-blown form,
become author.
CASTE allows the learner to I opment of il l
(local rules),
Among sorial ist ic learners, it is possible to distinguish at leas t the rather than i
follDwing two types. One is the o peration, or local rule, learner who Questions a
prog resses logica lly, s tep-by-step, moving to a different context only possible to i
when he Dr s he has assimilated one portion thoroughly and often asking inform<ltion
for tutori<11 gui d<l ncc in which topics to tackle next. The other is the rote The oth
learner, w ho fDllows the prescribed and narrow path. Unless the subject ufaclurc and
ll1<1tter is of trivial extent or the learner virtually hLls eidetic imagery, rote in test is Clll
lea rners hil ve little chance of success in making use of the knowledge wa s mth er I
the y have acqui red , although they may perform quite competently in administerel
multiple choice or ot her types of examinations where special memoriza- provided tha
lion skill pki Cl 'S thclll at an ildvllntage . in varinlll , "
f.n.JIll these description s it may appear that serial learners learn s uited to a gi
rules that holist ic learners do not. If so, that misconception must be "Spies" and
remedied. Successfu l se rial and holistic learners both come to grips with prod Ll ced in
rules, perhaps of th ei r own making or perhaps inherent in the subject Since in
 IROON PASK STRATEG IES AND CONCEPTUAL STYLE 93

CASTE and matter. The seria l learner is inclined to focus upon local, sma il-predicate
Japer later rules and the holstic learner upon global, large-predicate rules. The
o ns ca lled previously noted redundant holist tends to use dispersed and over-
specified (i.e., enriched) rules, whereas the irredundant holis t is happier
For exam- with pure (i.e., minimally specified) rules. Finally, if this terse account
ho depend still suggests that one type or another is at an advantage, then it is
1 th e mate- misleading. The evidence strongly su ggests that successful learning
I (b) those docs involve innovation or discovery but also the illtegratioll of both local
generaliza- and global rules. Bo th serial and holistic s trategies are need ed to achieve
) be one of unders tanding.
·ery. These
;eneraliza -
o und to be TESTS OF CONCEPTUAL AND LEARNING STYLE
·ither valid
,bject mat- Soon after CASTE and INTUITION w ere in use, we became "ware of
oVetrolling, the need for test learning situations yielding scores which could be used
to predict which learning strategies were likely to be successful with
analogize certain individu als. A pair of tests were constructed. They have excellent
lment, th e covalidity and arc based upon the sa me form of entailment mesh, relat-
d o mina nt in g topi cs (s h<lrcd concepts) that have mod els as well <IS descriptions .
clearly th e They both require participants to learn a quantity of unfamiliar but th e-
'henever it matic material. Th ere are global as well as local rules, but prediction, in
However, Bruner's (1975) se nse of "going beyond the informati on given," is re-
I ilS ce rlai n quired of respon den ts, as well (IS the recall of whatever ha s been
e teacher'S learne d .
nde rsta nd- One test, the "Spy Ring History Test," is concerned with the devel-
. learn e r to opment of an espionage network in which the agents h ave specific roles
(ioca l rules), and thei r activity, found ed upon a logic of permission
It leas t the rather than implication, influences the form of network by global rules.
arner who Questions of both a predictive and a recall type are detailed, a nd it is
ntext o nly possible to infe r a great deal about h ow a s tudent goes beyond the
len <1sking inforl11iltion predictively as well as how recall tilke s pla. ce.
is th e ro te The other test, matched and covalid, is concerned with the man-
_he subject , ufac ture and d is tribution of heroin . The syntheSiS proposed in the hero-
" ' ..
II 'lT) ' nlt- ·
. Ilowlctigc
in test is dlL'Jll il'.tlly possible but nowad.1Ys is uneconomic and always
wns mthcr h;1z.1rd nus. Both tests occupy about 3J h ours and may be
lC lcntly in ildministered either l11ilnu<1l1y or through a computer program. Further,
•
neJrH)rlZi.l- provided lhi.lllhe form of the entailment rnes h, wit h model s in cluded, is
invariant, a use r is <1ble to manufacture versions which may be more
ners learn suited to a given population .. The originals are, hencefo rwa rd, called the
,,
1 must be "Spies" and the "Smuggle rs" tests. Several alternative tests have been
grips with prod uced in othe r labo ratorie s.
he subject Since in strument s of this kind arc crude, relillive to th e scrutiny of
• •

94 CORDON PASK ST/M 1'[CIES AN

CASTE records or to the personal attention of one participant experiment- There are
er, it is only possible to detect general dispositions or styles of concep- istic tests and
tualization and learning. However, these results have predictive va lue. rna trices, AH:
The scoring scheme of the latest version admits to considerable large correlatil
latitude. No particular scoring method is sacrosanct and there are nu- and learning '
•
merous possibilities. Braten (1982), Laurillard (1980, 1981; Laurillard & teacher trainil
Mirante, 1981), and Lindstrom (1983) have produced equally reliable Academic
and, in certain res pects, more-informative scoring systems. All of the strongly to a'
sys tems are reliable and yield useful correlation with aptitudes as well as there is also .•
specific skills and their performance. academically'
Our own group uses the following indices: ing sensible, l
1. A versatility index (V) based upon forecasting rather than given correlatio ns '"
data, interpolative or predictive, and requiring the invention of puter prograr
more or less perspicuous analogy relations. life situations
2. An index (X) chiefly of global pattern recall.
3. An index (Y) of global rule recall. The sum of Y and X is the score STYLISTIC TES'
called cOlll prehension learning, or C score.
4. An index (U) of local rule and operation recall. Rote learning of It is uscll
li sts of mess<lges, in con trast to th e local rules that agents use, is only in vLlriu l
discounted, so far as possible. Rule learning is known as an the context a
operation learning skill. have maintai
5. An index (R) of rote recall (chiefly of message lis ts) . On th e which is ope
whole, with performance in mind, this index should be counted material is ell
against the learner. index. It is P'
6. An index (S) of semantic recall. appreciate no
7. A neutral (N) score, indicating that the material said to be stud- way. Th ey ar,
ied has, in fact, been s tudied . This index has no discriminative as individual
stylistic value but, if it is below a reasonable level, the data may organize and
as well be discarded as guesswork . from ou tside
skills) may e\
In addition, learners are required to provide a written or pictorial some exten t ,
account of what events they believe to have occurred during the period desirabl e; or .
covered. Thi s story given an adequate length and sufficien t content, Amo ng a
plays a quite crucial role in the analYSis of results. decision skill;
Cross correlation between the scores is passably high and reached group and th'
the level for the V, U, and C scores of between .40 and .58 or .50 and .81 group has en
(depend ing upon the calculation of indices) for 74 learners. We used of "Smuggler
paired participants in teacher training colleges, youngsters at Queens- explaining ro
gate school with mandatory "Iearning-to-Iearn" classes, and old er per- distinct strat,
sons at Henley grammar school/6th form and college/first-year univer- free d iscussio
si ty. Apart from the Queensgate group and one Henley group, there is a a cued recall I
stro ng opera ti on lea rning bias . The reverse difference appears for the situations ha '
groups first mentioned. The V score has the property of increasing if texts (such a:
both C and U are high. But it was highes t at H enley. have em ploy,
 ,.
'.

OROON PASK
STRATEGIES AND CONCEPTUAL STYLE 95

-,xperimcnt-
There are significant correlations between components of these styl-
:) uf concep-
is tic tests and other psychometric instruments (analogies, progressive
ctive value.
matrices, AH5, and others), but the corre lations are smaller. There are
onsid e rable
large correlations between the V, C, and U indices (between .53 and .70)
lere are nu-
and learning su ccess on CASTE for the Henley group and one gro up of
.a willard &
teacher training stude nts.
lily reliable
Academic performance is harder to assess. The U score relates most
. All of the
strongly to academic success in biology and physical che m is try, but
'" as well as there is also predominance of this operation learning style amo n g all
academically successful students, some being also versatile (V) and hav-
ing sensible, but not often high, C scores. Others ha ve found significant
than given corre la tions with task performance, esp ecially in va rious aspects of com-
lVcntion of puter programming, a nd problem solving, in both laboratory and real-
life situations.

is the Score
STYLISTIC TESTS AND LEARNING TO LEARN
learning of It is useful to add that these stylistic tests ha ve been employed not
~ n ts
use, is o nly in various psychometric, sty le-d iscriminatory s tudies, but also in
own as an the contex t of specific "learning-to-lea rn" studies. Various research ers
have maintained the existence of a general skill of learning to learn
;). On the which is open to acquisition by most people. Learning to learn novel
.Je counted material is closely related to innovation and probably to the versatility
ind ex. It is possessed by some people who are able to approach and to
app reciate n ove l con tingencies or subject matters and to make their own
to be stud - way. They are able, on the one hand, to adopt roles and make decisions
-riminative as indi vidual s o r as leaders in a gro up and, on the other hand, to
. data may orga nize and assimilate a body of subject matter w ith minimal direction
from outside. The skills required (gen erally called learnin g-to-learn
skills) may even be manifest in a desig n or invention context, where to
)r pictorial
some extent such skills are mandatory; in scholarship, where they are
the p eriod
desirable; or in an action-oriented situation.
t can tent,
Among oth e r methods of promoting th ese generallearning-to-Iearn
decision skills, we have the p ossibility of adm inistering s tylistic tests to a
d reached
g roup and th en giving them feedba ck regarding their scores. Afte r this
00 a n d .81
group has e ngaged in the sa me learning experience (say, of "Spies" or
We used
of "Smuggle rs"), the resulting sco res are offered as indices of difference,
t Qu een s-
explaining roughl y h ow they are d erived. The differences pinpoint the
older per-
• dis tinct strategies e mpl oyed by the pa rticipants, and they encourage
J r ll niVe r -
free discussion of the relative advantages of each approach. At the e nd ,
there is a
a cued recall rev iew of th e discussion is con ducted by the leader. Similar
rs for the
situa tio ns have bee n engineered usi ng CASTE in m ore inte res ting con-
rea s ing if
texts (such as the Team Decision System, discussed late r) and others
ha ve em ployed mo re socially oriented techniques.

!
•
I

-
 96 GORDON I'A SK STRATEGIES /

The immediate results, although difficult to quantify, have yielded /\ series

positive but far frolll dramatic results, the most salient beneficial effects over four se ~

being a marginal improvement. A far more useful and convincing result for their "Stl
is that greater than 20% of participants, commonly mature people who age. Each Sf
have often to act in real-world group or individual projects, return after was balance
intervals of 6 months to 1 year and often at considerable inconvenience the four ses
to ask advice upon their specific projects_ They say that they have "got test. The ne
the idea" of the training sessions and, further, "h<lve used their ac- Advanced r
quired skill." In th e studies conducted by my own group, consultancy of The next w;
this type is always promised. This is not a typical laboratory result, but a ment, and t
practical and heartening finding which deserves record since it suggests and the Inv
that researchers and professionals must perhaps expect an interval of 6 So far a
months to 1 year's duration before their labors bear fruit. Spies in the
sian. Carre
prehension ,
dex). The c(
THE ENIGMA OF STRATEGY AND STYLE Betwee
comprehen~

In ilGtdcmic learning or innov(ltive tmining, it looks <IS though peo- dispositions

COJl1prchcn~
pic do adopt one or other type of strategy as a personal style. The
prediction is that P-Individuals (persons, separate mental organizations correia tions
in one person, or organizations identifiable as a group of people) should ven's Progr,
do so, but at the actual person level, our results seemed to be unduly and .60_ Co
positive. This finding, noted previously, remains enigmatic. factory valu
OUf conjecture is that the person-oriented correlation can be largely son's stratel
ascribed to institutional constraints. These constraints, especially for high, as I n
preuniversity science students, are imposed by examination techniques Due to
and the necessary serializa tion of tuition, a point confirmed by the domi- this experir
nantly serial bias in style for this category of students. A less-prominent ended ques
but holistic bias pertains to architecture and art students evaluated by believed to
jury methods or cumulative assessment. The institution , viewed as a P- prediction s
Individu<ll, determines that academic success shall depend upon a cer- Only s(
tain, and fo r the most part, serial orientation, and it is hard to recom- ble to rcana
mend practical means of alleviating this state of affairs, especially since dent respor
institutions provide a significant social as well as academic utility. The of impartia l
pity is that effective learning is not optimized by an enforced mismatch (outside the
of s tyle, and that point is borne out by noting that other-than-academic or smuggiel
pursu its (be they astrology or pop music) have often been learned in a material ). E
mntchcd but sc hob stic<1l1y atypical manner. and ci<lssifil
In order to dClllonstrate that fixity to persons does not always occur, ry of "anal)
rather fixity to P-Individuals (i ncluding whole institutions), it proved tive from 0
necessary to eX<1ll1inc occupations in which changes of perspective are Among the
essl'nti.1l constitul'nts and to reanalyze much of the s tandard data. different P-

--
JJWON PASK STRATEGIES AND CONCEPTUAL STYLE 97

ve yielded I\. series of experiments was carried out under classroom conditions
cia I effects over four sessions in a high school setting. The students were matched
cing result for their "streaming" in class, their assessed academic ability, and their
oople who age. Each session lasted a double period of about 2J hours. The design
..' turn after was balanced with respect to the stylistic tests Spies and Smugglers. Of
•
nvenlence the four sessions, one was devoted to the administration of a stylistic
have "got test. The next session was devoted to a test battery including Raven's
.their ac- Advanced Progressive Matrices and the Lancaster Learning Inventory.
ultancyof The next was devoted to the stylistic test not used in the first experi-
;u lt, but a ment, and the last to a test battery of the AH5 test, the analogies test,
s uggests and the Inventory of Learning Processes.
orval of 6 So far as this study is concerned, there were 53 students completing
Spies in the first session and 74 completing Smugglers in the first ses-
sion. Correlations exist between stylistic tests and versatility, com-
prehension, and operation learning indices (as well as the neutral in-
dex). The correlations ranged from .47 to .70.
Between versatility and the normalized product of operation and
comprehension learning there is a correlation of .84, indica ting that both
Ig h peo- dispositions are needed in order to be versatile. The intercorrelations of
vie. The comprehension learning and operation learning are about .55, and inter-
izations correlations between these and problem-solving test scores (notably Ra-
I s hould ven's Progressive Matrices and the AH5) are fairly high, between .40
unduly and .60. Correlations between the Spies and Smugglers test gave satis-
factory values between .54 and .77. In fact, this consistency of the per-
. largely son's strlllegic preference across tests (and subject matters) is curiously
ally for high, as I noted above.
•
lnlques Due to the experiences of other researchers, the data obtained from
~ domi- thi s experiment were reanalyzed. The crucial matter W<lS one opcn-
minent ended question asking for an illustrated story of what the respondent
Ited by believed to have caused the thematic changes that occur and how their
as a p- predictions are made.
I a cer- Only some of the students furnished intelligible accounts, suscepti-
'ecam- ble to reanalysis. Some provided pictorial augmentation. Forty-four stu-
I s ince dent respondents did so, and these results were examined using a pair
y. The of impartial judges. The judges opted for classification as "analytic"
match (outside the system) and " involved" (inside the system, i.e., as an agent
demic or smuggler or inhabitant of the countries described in the thematic test
d in a material). Each of these 44 intelligible story responses were analyzed •
and classified with virtually no disagreement and assigned to the catego-
)cell r i ry of "a nalytic" or "involved." Of these, 25 people did change perspec-
roved tive from one test (context) to the next, and 19 people did not do so.
'e are Among the 25 people who changed perspective between tests (acted as
I. different P-Individuals in the different tests), the intercorrelations be-

i
I
 •

98 GOR~ON PASK STRATEGIES AND C<

tween tests were between .04 and .50, and for those 19 who did nut
change, intercorrelations were between .65 and .90. Thus, sume ie(lrn-
ers were more sensitive to context and more versatile in approach than It seems evid ,
were other learners. McCarthy and Schmeck (Chapter 6, this volume) be derived from I
and Schmeck (Chapter 12, this volume) maintain that this suggests a sentations (maps ,
difference in personality development. rationally believec
Design appears to be an occupation in which success depends upon of both subject n
frequent changes of perspective, but our first experiments, in the design detailed com paris
of electronic circuits, gave disappointing results . The fact was that op- learn about a sha
timal design is, under these circumstances, substantially predetermined There are alse
by the relevant stylistic stereotypes. of strategy. The 0
However, architectural design, examined in a fairly informed way would be charact,
at the Architectural Associatiort School of Architecture in London, is a mental organizati
different matter altogether. In order to compromise between infor- people and appli'
mality, open-endedness, and controlled rigor, experiments were carried distinct entaiImen
out upon the design of an intruder alarm system (several designs being More recent a
submitted due to variations in the brief). In this context, there is no origina l hypothes,
signifi cant correlation between the interpolated Spies and Smugglers seve ral coexisting
stylistic tests, so that it seems people do not adopt the same style or act mind) or to use [
as the same P-Individuals throughout this design process. argue (along witl·
In addition, further evidence comes from studies using a Team tional constraints
Decision System (TDS) in which stylistic test indices are admirable pre- responsible for II
dictors of an ability to plan (a prerequiSite of complex decision making) [earning style, wI
but not of !Ising plans under conditions of overload. The TDS merits ample, must chal
attention since it is able, used as a computer-regulated interface, to Howeverr I am st
exteriorize strategies of action, tactics, and planning strategies where persist in using t
responsibility is delegated. pear, of sheer nee
The method involves one or more "commanders," each in charge of and ineffective fo
at least two "ships," or any commander of necessity delegating tactic their style does I
responsibility for at least two ships. The tactics ordained are subse- institutions. Also.
quently executed (the "ship" is an independent microprocessor openit- text-sensitive tha i
ing in a further independent, environmental processor, and a tactic is a
context-dependent program). The method used is a measure of complex
decision making.
In these studies, planning ability is predictable from stylistic tests .
In fact, for a sample of 19 subjects who undertook two or three missions Bratcn , 5. (1982). Sim I
each lasting between 3 and 4 hours, the versatility correlation with plan- survc.Y, 2, 18Y- 21 1
Bruner, J. S. (1975). Bl
ning was .95, the operation score .91, and the comprehension score .56. !-low stmfcl1ts [co r
These figures, obtained by quantifying tactic construction, are highly Post.Cu mpulsory
significant. But there is no significant correlation between style and the de Gelder, 13. N. F. ( I'
effective use of tactics. Readers interested in planning should also con- cybcnlctics. New '
sult Das (Chapter 5, thi s volume). Kelly, G. A. (1955). T,

, •
 ON PASK STRA TEGIES AND CONCEPTUAL STYLE 99

did not GENERAL CONCLUSIONS

e Jearn-
ch than It seems evident that distinctive learning strategies exist. These may
'olume) be derived from unfoldments (pathways) of canonical, evolving repre-
,gests a sentations (maps of topics and their relations) of what may be known or
, rationally believed, that is, an entailment mesh. Given this tangible form
.5 upon of both subject matter and individual strategy, it is possible to make
design detailed comparisons of strategies, whether they are of learning (how to
',at op- learn about a shared concept) or action (how to reach a goal).
-mined There are also certain distinct styles, or dispositions to adopt classes
of strategy. The original conversation theory prediction was that styles
'd way would be characteristic of specific P-Individuals or partly autonomous
•
n, IS a mental organizations. Later it seemed that styles belonged to specific
infor- people and applied to learning situations involving a great variety of
carried distinct entailment mesh.
being More recent and detailed analyses of the initial data also support the
•
15 no original hypotheses, especially that one person may be the residence of
gglers several coexisting P-Individuals (points of view, perspectives, frames of
or act mind) or to use Laurillard's phrase "people in context." I would also
argue (along with Ramsden, Chapter 7, this volume) that the institu-
Team tionai constraints of an institution-sized "P-Individual" are to be held
e pre- responsible for the initially apparent person-fixity of conceptual and
lking) learning style, which seemed both counterintuitive (a deSigner, for ex-
nerit-s ample, must change point of view) and inherently counterproductive.
,e, to However, I am still perplexed by the extent to which some individuals
vhere persist in using the same strategy across contexts. Some students ap-
pear, of sheer necessity, to be institutionally conSigned to a mismatched
"ge of and ineffective form of intuition, examination, and the like, insofar as
tactic their style does not match the serialist style of academic or training
Jbse- institutions. A1so, some persons seem more flexible, versatile, and con-
"

ler" t- text-sensitive than do other persons .

•
:: IS a
lplex
REFERENCES
ests .
•
Ions l3rnten, S. (1982). Simulation and self-organization of mind. COlllcmporary philosophy: A lIew
lian- survey, 2, 189-218 .
.56. Bruner, J. S. (1975). Beyond the information given. In N. J. Entwistle & D. Hounsell (Eds.),
,hly How SllidcHts lcam. Lancaster, England: Institute for Research and Development in
Post-Compulsory Educntion. '
the
de Gelder, B. N. F. (1981). Cognitive transactions. In G. E. Laster (Ed.), Applied systems (!nd
.:on-
eybert/dies. New York: Pergamon Press.
Kl'lly, G. A. (1955). Till' psychology of persollal constructs (Vols. I nnd II). New York: Nortun.
 ,
I

100 GORDON PASK I

i
••
Lnurillard, O. M. (l9RO). III/('mc/ilw colII/llller ~i/lll/ll1liol/ ." ;11 IIml,'rsmdlllll,' fmd,iI ,s. !"'Ill'r
prl·sl·nl l·d at till' 2nd Inlen1lltiun .. 1 Micrucomputers in Educntion C ungress, London.
Lilurillilrd, O. M. ( 1981 ). Intera ctive computer simulations in undergraduate leaching.
Proceedillgs of Ille Secolld [llterl/a liol/al Microcomputers ill Educatioll COl/gress . London:
Harves ter. •
LOlurill<l rti, D. M " & Mnrank, G. J. (19tH). A vicwof WIII/ Jllh'r -as:;;s/t'd kart/ illS illll/{' /iSM of
COl/verslllio ll theory . England: Milto n Keynes .
Lewis, B. N., & Pa sk. G. (1965). The theory and practice of adaptive teaching systems. In R.
•
Glasser (Ed.), Teachillg machines and programmed learning (Vol, 11). Was hington, DC:
National Education Association.
Lewis, 13. N. , & Pas k, G. (1965) . The theory and practice of adaptive teaching sys tems. In
R. Glasser (Ed .), Teaching machines and programmed learning (Vol. 11). Washington, DC: I,
National Education Association.
Linds trom, B. (1983). Learnillg styles and learning strategies: Conversation theory- the work of
Gordoll Pasko Sweden: University of Goteborg Press.
Pask, G. (1961 ). All appr{)(/ch to cybernet ics. London : Hutchinson.
Pask, G. (1970). The mean ing of cybernetics in the behavioral sciences. In J. Rose (Ed .), ,•
Progress of Cybernetics, Vol. 1 (pp. 15- 45). London: Gordon and Breach.
Pask, G. (1972). A fresh look at cognition and the individual. Ill/ema/iollal IOl/ mal of Mal/ -
MIIChjlll' Siudies, 4, 211 - 216.
Pask, G. (1975a). COl/versa /io ll, cognition and leamillg: A cybemetic theory alld methodology.
Amsterdam: Elsevier.
Pa sk, G. (1975b). The cybernet ics of IlIIlIIall leamillg alld perfommllce. London : Hutchinson.
P... sk, G. (1976... ). ConvcrS<ltion techniques in the study (lnd pril cticc of cduc(ltion. lir;ti:;',
JOlln1ll1 of [d llm /jo llal Psyc/w/oSy, 46, 12- 25 1
POl sk, G. (l976b) . Styles and strategies of learning. British JOllmal of EdllcatiOllal Psycl/Ology,
46, 128-148
Pask, G. (1976c). COllversatioll theory: Al'plicntiolls ill edllcatioll alld ellis tl'lI/oloS:Y. Amsterdam: This t
El sl..'viL' r. two I
Pas k, G . (1977). uamillg styles, edllcatiollal sl rategies alld represelliat ;0115 of knowledge: Methods and I
ami (fl'p/icatiolls : Progress r~port 3. London: Social Science Research Council. feren
Pa sk, G. (1978). Leamillg styles, educational strategies alld representations of knowledge: Met//Ods
(llId applicatiolls: Progress report 5. London : Social Science Research Council.
betw,
Pa sk, G. (1979). Leamillg styles , educational strategies and represclltnt iOlls of knowledge: Me/hods era I,
GllrI,I/lplicatiolls: filial reporl. London: Social Science Research Council. prom
Pask, G. (1983). Kllowledse and illllOVtltioll of decisioll makers: Filial tecJmical reJ1Orl. London: 'I
U.S. Arm y Research Ins titute. book
Pask, G., K.1l!iko·urdi s, D., & Scott, B. C. E. (1975). The representation of know(lbles. •

lllleri/otiolla/Jolimal of Mall-Machillt Studies, 17, 15-134.

cesslI
r .lsk, G ., & Lewis, B. N. (1968). The use of a null pOint method to study the acq uisition of arbih
simple and com plex trans formational skills. British Joumaf of Mathematics nud Statistics, styles
21 61-84. Engli'
Pask, G., & Scott, 13. C. E. (1971). Learning and tea ching strategies in a transformation
defint
ski ll. Brit ish JOllmal of Mathematics alld Statistics, 24, 205-229.
Pask, G., & Sco tt , U. C. E. (1972) . Learning strategies and individual competence. ill/ema-
Learlll
tiOllal Jtll/wal tlf Man-Machine Studies, 4, 217-253. Astra
Pa sk, G., & Scott , B. C. E. (1973). CASTE: A system for exhibiting learning strategies and to pre
rl.' ~U I 'ltin~ unce rta inty. lll/ematiollal lOllmal of Mall-M(lcJ,ille Studies, 5, 17-52.
Sch ml'(k, R. R. (1983). Learning styles of college st udents. In R. Dillon & R. Schmeck
(Eds.), 11I/III'ilflial DiffaellCt's ill Cognitioll. New York: Academic Press.
Sh.1\\-" , M. L. G. (1980). Epistemological hierarchies ill kllO"wledgc alld expert systellls. Un- f. P. D
pul>ti:>h\.·d rt'porl. Yo rk University, Department of Computer Science, Toronto . Alberta
,
 •
, Perspectives on Individual Differences
Series Editors: Cecil R. Reynolds and Robert T. Brown
New!
Using individual
edited by Ronald R. Schmeck
Southern Illinois, U niversity, Carbondale
"[n education, if we understand the styles of ind ividual
students, we ca n often anticipate their perceptions a nd
subsequent behaviors, anticipate their misunderstandings.
take advantage of their strengths, and avoid (or corren)
their weaknesses," says Ronald Schmeck, editor of this new
book. Here, he presents papers that describe research
methods used by international experts to assess learning
styles and strategies . .With data deri ved from widely differ~
ing cultural populations, they arrive at remarkably similar
conceptions of the educational process. Their chapters in~
elude theoretical perspectives on the development of
cognitive style and practical suggestions o n how to use the
style of an individual st uden t to improve the o utcome o f
his or her educational experience. The final chapter in-
tegrates all of the information in the book and points to
directions for fu ture research.
----------------------------------------------------()FlIJ]EIt IF()Fl~----------------------------------------------------- ·
Please send US/outside US & Canada
_ _ LEARNING STRATEGIES S4i.,O/$,7.00
AND LEARNING STYLES
ISBN 0·306·42860·1
Plenum Text Policy
For information about special discounts available o n text adoptlon o rders of
six or more copies, wrire to Adopt ion Sales Dept. Specify the si:e of the adop-
tion.
Book orders by individuals must be paid fully in advance.
AI! major credit cards accepted.
type of card expo date
account no. bank no.
sIgnature
• • •
CONTENTS, The Concepts of Strategy and Style, An in-
troduction to strategies and styles of learning , R. R. Schmeck.
Motivatio nal factors in students' approaches to learning, N.
Entwistle. Describing and improving learning, F. Marron. Learning
strategies, teaching strategies , and conceptual or learning style, G.
Pask. Simultaneous-successive processing and planning, J. P. Das.
Students' self-concepts and the quality of learning in public
schools and universities, P. McCarrhy and R. R. Schmeck. Applica-
tions of the Concepts of Strategy and Style: Context and
strategy: situ ational influences on learning, P. Ramsden. Ap-
proaches to learning and to essay writing. ]. Biggs. Style, strategy I
a nd skill in reading,]. Kirby. Styles of thinki ng and creativity, P.
Torrance and Z. Rockensrein . Assessment and training of student
learning st rategies, C. Weinstein . Strategies and styles of learning:
an integration of var ied perspectives, R. R. Schmeck. Index.
0-306-42860-1 /347 pp. + index/ i11. / 1988/$47.50 ($57.00 outside
US & Canada)
name
affiliation
address
cIty statelzip
Mail )'Our order today.
PLENUM PUBLISHING CORPORATION
233 Spring Street
New York, NY loolJ-1578
Telephone orders, 1·21 2·620-8000/ 1·800-221·9369