Difficulties in learning science concepts

advertisement
Katherine Richardson
Cognitive Development and Learning
Why do many find science concepts so difficult to learn? Evaluate theory and
research that seeks to account for this, and also consider research that has sought to
address this problem at the level of classroom instruction.
Measuring science learning
What would successful learning of science concepts look like? How do we know that science
concepts are ‘difficult to learn’? Although the research literature often fails to articulate this
position, research measurements suggest that learners are judged by comparison with the
scientific community in three ways. Firstly, using accumulated science knowledge, comparing
the learner’s use of concepts, explanations and reasoning with correct science. Secondly, by
using example ‘scientists’ to compare novice and expert scientists. Thirdly, by using explicit
criteria of ‘scientific-ness’ derived from philosophy of science.
Learners' ideas, procedures and reasoning are gathered from responses to written, verbal or
practical assessments. Many variables may influence subject performance in these
assessments, and some of these will be confounding variables. Factors affecting learner
concepts are the extent and quality of prior instruction, changes or development of a learner's
concept, and the quality of mental representation and storage of the concepts. Additional
factors which may influence use of these concepts are understanding of the assessment;
identification, selection and application of relevant concepts; and conceptual change induced
by the assessment materials or researcher.
Clinical interviews are commonly used to uncover learner ideas. This is highly resourceintensive, but has advantages over written tests as researchers can perform tailored
conceptual probes. This allays Reif and Allen’s (1992) concern that ‘gross measures’ such as
scores do not reveal sufficient information about "the complexities of concept interpretation
and . . . performance deficiencies." (ibid, p11) Two-tier testing, in which students explain
their reasoning by means of a diagram or written prose, sits between standard testing and
interviews in terms of data quality and resources needed, and has been used in large scale
international surveys of science achievement. (Martin et al, 2004)
Posner and Gertzog (1982) reviewed uses of Piaget’s clinical interview and found several
flaws. In one study, children were rated once for a whole interview, ignoring the complexities
of their answers. Another study coded exclusively on learner comments without considering
the interviewer’s comments, which risks including comments ‘suggested’ by the interviewer.
More general cautions about clinical interviews date back to Piaget’s (1967) commentary
which warns researchers to distinguish a child's reasoning and spontaneous convictions from
researcher suggestion, idle speculation and 'romancing'. Treagust (1998) notes that clinical
1
Katherine Richardson
Cognitive Development and Learning
interviews isolate a learner from the normal contexts of their thinking, and so may not
adequately reflect situated problem-solving.
Adey (2005) stresses the importance of longitudinal studies for assessing the long-term
effects of cognitive change. In the original Cognitive Acceleration in Science Education
(CASE) experiment, a delayed effect on pupil achievement appeared one year after
completing the intervention. With short-term studies, these delayed improvements may be
missed. Conversely, long-term studies may find student ‘changed’ conceptions reverting to
prior convictions after instruction. Evidence of long-term conceptual change requires longterm assessment.
The Nature of Learning Science Concepts
Researchers take two broad approaches to account for difficulties in learning science,
reflecting different conceptions of learning. Those who view science learning as conceptual
change focus on learners’ domain-specific ideas, and how these change through interaction
with internal and external stimuli, such as instruction, cognitive conflict, and metacognition. In
contrast, the theory of learning as strategy use focuses on the learner’s use and application of
domain-specific ideas to solve a problem, and their use of scientific processes to acquire
scientific knowledge.
These approaches are strongly interconnected. For example, a learner’s improving use of
inference from experimental data may enable conceptual change in the domain illustrated by
the experiment (Kuhn, Schauble and Garcia-Mila, 1992). Conversely, a conceptual change
may lead directly to a change in strategy use. For example, understanding that an object is
accelerating when it changes direction allows a learner to use ‘acceleration’-based strategies
to solve problems of circular motion (Reif and Allen, 1992).
I will examine each approach, considering how they account for failure to learn science
concepts, how empirical studies support the approach, and the instructional strategies they
recommend. I agree with Duit and Treagust (1998) that different views of learning are not
necessarily competing explanations, but “complementary features that allow researchers to
address the complex process of learning more adequately” (p5). However, I will also highlight
where different viewpoints provide conflicting explanations.
Learning as conceptual change
The persistence of learners’ scientifically incorrect ideas after formal instruction has been
widely reported (eg Driver, 1990). These ideas have been variously conceptualised as
‘misconceptions’, ‘alternative frameworks’ and ‘naïve or intuitive science’, with each term
describing a different extent of coherence and reasoning.
2
Katherine Richardson
Cognitive Development and Learning
Kuiper (1994) argues that the nature of these ideas is important for planning instruction. If
learners have developed extended framework of ideas, then these must be challenged at the
level of this whole framework. Conversely, if learners are using a range of fragmented ideas,
then teaching should challenge at the level of an idea, and also aim to improve
systematisation. Kuiper presented learners (n=143) with sixteen mechanics problems
involving four different physical principles. Each principle was presented in four contexts,
such as identifying the forces acting on four different objects at rest. Factor analysis revealed
high factor loadings (>0.5) between types of answers for the same physical principle, but it is
possible that the close proximity of these test items affected learner responses.
Kuiper assumes that an ‘alternative framework’ is built at a higher level than a single physical
principle, and instead looks for an overarching alternative framework for force. However, in
looking for these frameworks, he assumes that students have a distinct concept of ‘force’
which has a single effect on motion, and thus concludes that most students’ conceptions are
incoherent:
“Students use the concept of force in all questions . . . but in a rather haphazardous
way: a (resultant) force is seen as either keeping an object at rest, setting it in motion,
keeping it in uniform motion or stopping its motion. . . . There is no logical coherency
between their ideas.” (Kuiper, 1994, p289)
In contrast, Kuhn’s (1962) idea of incommensurability suggests that we cannot always
translate between frameworks: there is no reason that a consistent concept ‘force’ should be
important to a learner’s framework, even if they use the word. Carey (1986) borrows from
Kuhnian interpretations of science history, arguing that conceptual change is equivalent to
paradigm shift, where successive theories are incommensurable, each theory is unfalsifiable
in the others’ terms. She cites McDermott's (as cited in Carey, 1986) work on mechanics,
which suggests that 'novice' use of the word 'force' is unstable and cannot be reduced to the
'expert' view of a force. It is therefore philosophically problematic to evaluate alternative
frameworks in the way suggested above. A consensus conclusion about the nature of
learners’ initial ideas therefore remains elusive, and this may affect the selection of
instructional strategies.
While Kuiper took an empirical approach to conceptual change, other researchers have
worked on more theoretical approaches. Chi et al (1994) define conceptual change as recategorisation of a concept, such as re-classifying ‘whale’ from fish to mammal. They predict
that the most difficult conceptual changes in science require a change in the ontological
category of the concept, often from ‘matter’ to ‘process’. For example, learners may initially
endow ‘force’ and ‘current’ with material properties, whereas the scientific concepts are
processes rather than entities. The researchers’ evidence comes from categorical analysis of
learners’ ideas. They found that naïve ideas which are in a different ontological category to
the desired scientific ideas (as in ‘heat’ and ‘energy’ were resistant to instruction. In contrast,
3
Katherine Richardson
Cognitive Development and Learning
when naïve and scientific conceptions shared an ontological category, initial ideas tended to
be fragmented and relatively conducive to instruction. In the experimental work for this study,
we see the difficulty in disentangling learner’s language from their concepts. Coding of pupil
responses used pupil language to determines whether concepts such as ‘heat’ were
represented material entities or processes. Here, student choice of language may not reflect
their mental representations/understanding, particularly if their vocabulary is limited or their
mental representations are primarily non-verbal. Analogy to material entities may also form a
valid part of reasoning, which further undermines the use of language as a determinant of
conceptual category.
Karmiloff-Smith (1992) presents a more general theory of conceptual change, modelling this
change as representational re-description (RR) which allows greater flexibility in strategy use.
She considers a child learning to balance a solid shape on a metal support. Once this
procedure is mastered implicitly (phase I), internal and unconscious conceptualisation of this
procedure can occur, leading to explicit representations (phase E1) which begin to integrate
the procedure found for different solid shapes. This concept can later become consciously
accessible (E2), and finally verbally communicable (E3). However, all previous
representations of the concept remain accessible to the learner. Karmiloff-Smith believes that
more explicit representations allow better application and use of concepts, and some
research supports this idea. Prain and Waldrip (2006) note that primary pupils who translated
effectively between visual and verbal representations (E3) were judged as ‘more able’ on
previous assessments. However, it is not clear if representational re-description contributes
to other conceptual changes which affect assessment, or effective translation is being
measured by assessments, as written interpretation of visual information such as diagrams
and graphs is expected in many science tests. In a second case study, Prain and Waldrip
found a correlation between inter-representational links and conceptually accurate accounts,
which suggests that E3 knowledge is useful for conceptual development as well as for
translation.
RR theory poses five major challenges for cognitive researchers and teachers. Firstly, a
learner may initially respond to a familiar task using their mastery of procedure rather than
more explicit representations because this allows for speed and automaticity. Researchers
therefore need to provide an opportunity for explicit representation. Secondly, the process of
explicit conceptualisation is marked by reduced performance on a series of tasks. In phase
E1 above, attention shifts from external data to the internal representation of the task.
Children have theorised that blocks can be balanced at their centre of volume, and rely on
feedback from this theory to position each block. They are therefore unable to balance
blocks which are weighted at one end. In contrast, children in phase I treat each block as an
independent task and rely on feedback from the task to position each block. This is a
competing explanation for learners’ inability to make logical inferences from observation: they
4
Katherine Richardson
Cognitive Development and Learning
are in a re-description phase and have adopted an internal theory-based focus. However,
these lower-performing learners (phase E1) have a better conceptual grasp of the task than
the higher-performing learners operating solely with procedural knowledge (phase I).
Measures of understanding and conceptual change must distinguish between these learners
appropriately.
Thirdly, the RR model draws a distinction between explicit representation, and conscious or
communicable access to that representation. Studies relying on learners’ explanations of
their concepts will be stymied by learners at E1 and E2 levels, who are unable to
communicate their representations. As yet, there has been little direct research on the RR of
science concepts, but many studies have reported on the seeming inconsistency of student
responses and reasoning, even in explanations of the same event. (diSessa et al, 2003).
Learners in phase E1 or E2 may be drawing on different representations to explain the event,
as these may seem inconsistent to the researcher, and yet retain coherence to the learner.
Fourthly, since learners who are undergoing redescription are prone to ignore observational
data in favour of their theory, instructional strategies which provide counterexamples are
unlikely to succeed. Karmiloff-Smith cites earlier work by Piaget in which children ‘observed’
a table become indented when an iron bar was placed on it, purportedly to maintain their
theory that forces have observable effects. If this sort of theory-laden ‘observation’ is
widespread, then teaching cannot rely on the universal perception of so-called ‘observables’
as evidence. Finally, it is not clear from the theory whether instructional strategies focusing
on metacognition (Mason, 2001) or developing argumentation (Simon, Erduran and Osborne,
2002) are of benefit unless learners are already consciously aware of their representations
(E2/E3). However, Kuhn (1993) argues that collaborative learning contexts may create
conscious awareness of conceptions, which suggests that higher levels of RR can be
triggered by external contexts, if not by external data. Further research is necessary to
discover the mechanisms which trigger re-description.
Karmiloff-Smith’s theory also emphasises the importance of positive feedback in developing
more flexible and explicit representations, unlike the theory of cognitive conflict which
emphasises negative feedback as a trigger. The RR model is therefore challenged by the
work of Stavy and Berkowitz (1980), who explicitly asked learners to compare their qualitative
and quantitative predictions of temperature changes when liquids were mixed. The cognitive
conflict triggered by the inconsistency of these representations led many learners to translate
correctly between qualitative and quantitative representations, effectively re-describing their
quantitative model (E2 or E3). However, the naive quantitative model held by many learners
was borrowed from mathematical addition, rather than from mastery of a procedure, so the
RR model may not apply in this case.
5
Katherine Richardson
Cognitive Development and Learning
Creating conceptual change
Research in classroom instruction often focus on predicted triggers of conceptual change,
such as the use of language and other representations, including analogies, explanations,
speculation and meta-cognition.
Recent studies on concept cartoons and meta-cognitive tools exemplify the difficulties of
studying the effects of classroom instruction. Classroom activities defy the concept of
controlled variables, and while the 'ecological validity' of these activities is undoubtedly higher
than that of a laboratory interview, it is very difficult to untangle a relationship between an
intervention and outcome from a host of confounding factors. These studies are also prey to
'confirmation bias', in which an improved outcome (often only compared with teacher
expectations) may be attributed to an intervention without consideration of other factors.
Keogh (2001) undertook a teacher survey, observations and interviews with teachers and
pupils to study classroom use of concept cartoons. Concept cartoons are a variation on an
interview-about-event, presenting a diagram of an event with three or four speech bubbles
suggesting explanations or theories about the situation. The study found little direct evidence
of conceptual change. Teachers reported an increase in productive pupil discussion, and
suggested that this contributed to an increase in understanding, but few ‘could offer evidence
about the nature of the change in ideas’ (p438). Pupils were also convinced that they were
learning something, though Keogh notes that the view of pupils after using the concept
cartoon ‘was not always more scientifically accurate than the original view expressed’,
reflecting the non-linear nature of conceptual change, and the persuasive and intuitive nature
of many non-scientific concepts. She also points out the difficulty of confirming a relationship
between the use of concept cartoons and positive conceptual change in a classroom setting.
A similarly ‘naturalistic enquiry' by Mason (2001) focused on 12 children as they learned
about micro-organisms, using reflective diaries, discussion and to promote meta-cognition.
Interviews before and after instruction confirmed that all the learners had altered their
conceptions of micro-organisms. Mason suggests that these changes were promoted by
collaborative learning and reflection which made students conceptions explicit and made their
thinking the object of reflection. Certainly, we can assume that learners were made more
aware of their conceptual change through tools designed to promote meta-cognition. This
emphasises a major difficulty in methodologies involving self-reporting: if students are more
aware of their learning, does this mean they are learning more? In the absence of a
comparison group or data for a similar classroom module, it is difficult to make any
conclusions. It also highlights a more general 'observation effect' in studying conceptual
change: if researchers ask for reasoning or metacognition from their subjects, it may alter
subsequent answers. For this reason, some researchers ask for reasoning on a 'second
pass' through presented problems, to avoid this interference (eg Reif and Allen, 1992).
6
Katherine Richardson
Cognitive Development and Learning
Other instructional studies have focused directly on addressing misconceptions. Levin et al
(1990) showed that students regularly attribute a single speed to an object, even when points
on these objects are moving at different speeds. Conceptual probes based on contradictory
data were ineffective at producing conceptual change in these students. Group discussion
was slightly more effective at producing conceptual change. However, kinaesthetic training
by physically modelling these situations proved most effective. Interestingly, most students
had physical experience of objects which contained points moving at different speeds
(roundabouts, cars turning corners). They suggest that the intuition survives because it is not
“the focus of attention when the potentially disconfirming physical experiences occur.” (ibid,
p277). Harris (1990) suggests an innate predisposition to regard a unified speed as part of
the visual concept for objects, citing Kellman and Spelke’s (1983, as cited in Harris, 1980,
p301) work with infants on rigid motion. They note that this intuition can be undermined by
kinaesthetic feedback which bypasses the visual system.
These classroom-based studies are most usefully considered as describing classrooms
where conceptual change occurs, rather than pinning down particular interventions which will
elicit conceptual change. The richness of detail provided about learner interactions and
instructional strategies is therefore the most important element of these studies.
In contrast, Novak (2005) suggests that children would simply benefit from instruction in the
abstract ideas of science at an earlier age. Echoes of the stage-theory belief that children in
pre-operational and concrete operational stages do not benefit from formal instruction in
abstract ideas can be seen in the National Curriculum for primary school science (DfES,
2004). In the Key Stage 2 curriculum (ages 7-11), pupils study changes such as evaporation,
dissolving and combustion, but do not study the particulate nature of matter. They study
electrical currents, photosynthesis, and motion, but do not study the concept of energy.
Novak included explicit instruction in particle theory and energy in an alternative science
curriculum for 6-8 year old children. Throughout twelve years of subsequent schooling, the
instructed learners remained significantly further ahead than uninstructed learners from the
same school.
In a similar vein, Ogborn et al (1996) also move away from the direct address of
‘misconceptions’ to focus on the nature of science explanations which are created in
classrooms. Their analysis of classroom observations produced a four-stage model of
explanation. The first stage creates a need for explanation, the second stage introduces
scientific entities (such as density) which can be used for explanation, the third stage
transforms ‘science’ into ‘school science’, and the final stage imbues physical phenomena
with the explanation. Of these, the first stage is the most revealing. diSessa (2000) notes
that phenomena which we explain by intuitive knowledge-fragments do not ‘need’ explaining,
7
Katherine Richardson
Cognitive Development and Learning
as they are considered self-explanatory. The pre-requisite for explanatory need in conceptual
change parallels the role of cognitive conflict.
Reynolds and Brosnan (2000)investigated conceptions of scientific entities by presenting
sentences and asking learners to judge whether they were sensible. They observed fruitful
discussion from asking students to speculate about entities, and suggested that speculation
may help students to learn science. Interestingly, the most scientifically competent learner
cited in the study used very little speculation, displaying a “precise but rigid reasoning style”
(ibid, p65). Speculation can therefore be seen as providing an alternative approach to
learning science. The case study of the most scientific learner also illustrates a problem with
novice-expert comparisons: existing experts are ‘selected’ from those who learned science
concepts successfully, probably during formal instruction. Their characteristics may not
provide useful advice for those who are struggling with formal instruction.
Posner et al (1982) also developed a conceptual change theory which draws on the ideas of
Kuhn. They considered the conditions under which a central concept in a problem domain
(such as mechanics) would be replaced. Unlike standard empiricism, which removes
concepts when they are falsified, Kuhn’s paradigm theory relies on the resources of a concept
to solve current problems in that domain. Posner et al selected four intrinsic characteristics of
a new theory, and several environmental factors which ‘govern the selection’ of a new theory.
A new theory must arise from dissatisfaction with the current theory. It must be intelligible,
plausible, and potentially fruitful for problem-solving. Environmental factors help to provide
dissatisfaction, intelligibility and plausibility. For example, dissatisfaction with the old theory
could be caused by anomalous data which is explained by the new theory. As discussed
elsewhere, analogies may help to ‘bridge’ learning and so help plausibility. Posner et al also
note the importance of more fundamental beliefs regarding metaphysics, knowledge from
other fields, and the epistemological character of a successful theory. They cite learning
relativity as a case where learners will ‘explain away’ observables in order to maintain their
underlying belief in absolute time, ascribing relativistic phenomena to a problem in
measurement or observation. Although this is seen as ‘unscientific’ by the researchers, it
may represent a sensible use of all the evidence that has been presented to the learner
(since the law of absolute time is widely applicable in everyday life). It is also worth noting
that an intelligible belief in the physical sciences is often negotiated by mathematical
description rather than linguistic description. This may begin to account for the difficulties of
learners who are only dealing with qualitative examples without access to the quantitative
underpinnings.
Learning as strategy use: challenging the purpose of conceptual change
8
Katherine Richardson
Cognitive Development and Learning
Driver (1990) emphasises formal science as sense-making on a grand scale, with many
formal criteria which ensure an integrated set of explanations. Formal science would
therefore seem to be an excellent ‘strategy’ for problem-solving, because of the wide
problem-set which formal science attempts to address, and its coherence as a body of
knowledge. However, a significant body of research purports to explain the persistence of
other strategy use by considering learners' application and use of scientific and alternative
concepts.
For example, diSessa, Elby and Hammer’s (2003) case study of an undergraduate student's
explanations for free fall motion suggests that some learners perceive science as incoherent
and non-pervasive, at least when applied to 'real-life' situations.
“you learn these formulas in school . . . half the time they only apply to certain
perfect models . . . you can't apply them to absolutely everything . . . when you're in
real life, I mean, so many things have so many different things going on that you can't
always say.” (p267)
Here, scientific explanations are rejected as a sense-making tool by the undergraduate J
because ‘real-life’ situations do not match the abstract requirements of scientific theories.
For example, the ideal gas law requires particles with zero volume, no attraction between
particles, and random motion. Many adjustments are needed to predict the 'real life'
behaviour of gas particles. However, a learner's experience of gas laws is likely to
commence with the ideal gas law, with ‘real life’ scenarios covered later in the curriculum.
Moreover, the problems which are presented to them in class are likely to be solvable using
the ideal law, since we rarely produce problem sheets which are irrelevant to the topic of
instruction. Thus, while the body of scientific knowledge itself is not fragmented, students’
access to that body of knowledge is often very fragmented, consisting first of special cases
before extrapolating to a more general cases. This is particularly the case where overarching
theories (such as relativity, or particle-wave duality) are complex or mathematically notated.
diSessa’s (2000) model of fragmented intuitive knowledge comprising visually and spatially
encoded ‘phenomenological primitives’ (p-prims) offers a potential explanation for J’s
confusion. According to diSessa, practical judgements about problems involve summoning all
possible relevant fragments of intuitive knowledge (p-prims) and applying those which seem
to fit the situation. This has obvious parallels with the information-processing emphasis on
recognising a schema for a problem, and may also extend to judgements made on the basis
of fragmented 'school science' knowledge. In Reif and Allen’s (1992) study of novice and
expert approaches to qualitative acceleration problems, novice science learners presented
similarly fragmented reasoning for different problems, whereas expert scientists were more
likely to use definitional knowledge of acceleration as the derivative of velocity (which was
also accessible to the novice learners). Reif and Allen argue that novices have not
understood the applicability rules for specific cases of acceleration (such as acceleration of
9
Katherine Richardson
Cognitive Development and Learning
free-falling objects or circular motion at a constant speed). Instead, they use these specific
cases in superficially similar but fundamentally different problems, such as all those involving
circular motion. Similarly, Chi, Feltovich and Glaser’s (1981) study of novice and expert
groupings of physical problems found that novices grouped problems according to superficial
features, whereas experts grouped the same problems according to the equations needed to
solve them. Information-processing theorists would argue that both groups are ‘recognising’
familiar problem features, but the experts recognise features in a way which will help them to
solve the problems more efficiently. In the absence of a systematic approach to physics
problem-solving, novices may fall back on the more familiar features of problem contexts, and
approach the problems using either intuitive p-prims or similarly fragmented 'special cases'
from school science.
Does the use of different strategies during this microgenetic study equate to conceptual
change? J self-reports a lack of conceptual change as she shifts between one-force and twoforce explanations, claiming that “I don't think that now I understand what's going on any
better . . . But I can explain it to you in the right way” (p249). The researchers note that she
used “terms like force, momentum and velocity as if they were interchangeable (p254).” The
interviewer challenges J on incorrect use of the word 'force', suggesting 'momentum'
instead. J accepts the change, but does not see a meaningful distinction. This undermines
Ogborn et al’s (1996) description of science explanation which focuses on “the right way to
talk about things” (p127). Is learning the right language equivalent to conceptual change?
Conversely, is there anything more to understanding science concepts than using the right
language? This is a cautionary tale for other researchers: the linguistic categories which
researchers use to distinguish between student conceptions may be considered equivalent by
their research subjects. J's lack of commitment to any particular explanation suggests that
she has not undergone conceptual change, and undermines the idea that all conceptual
change is as 'radical' and 'incommensurate' as Kuhnian interpretations imply. In this case
study, focus on conceptual change is problematic, and focus on strategy use seems more
appropriate as a descriptive tool.
Both Caravita (1994) and Posner et al (1982) warn against students who learn to "play the
game of physics” in a limited set of circumstances. However, I would argue that these
learners are engaged in effective sense-making of the world. When confronted with 'real
world' situations, they use 'real world' solution strategies which differ from those of formal
science. Driver (1990) notes that where formal science privileges general and penetrating
models, everyday science privileges correct production of outcomes. If a student is uncertain
in her application of science principles, particularly if they have experienced only limited
instruction, theories from accumulated experience are likely to prove more pragmatic.
10
Katherine Richardson
Cognitive Development and Learning
This suggested that the use of familiar 'real world' situations in clinical interviews may actually
hinder learner use of science concepts, as students choose to interpret the problems using
their everyday experience and ‘school science’. This is particularly likely where contextual
aspects of the situation reduce the similarity between a formal principle and the situation,
leaving learners unsure of which strategies represent the best fit.
"When in instruction we introduce examples from the students' everyday experiences
. . . we want the students to observe only those particular aspects of the situation that
are of relevance within a given theoretical framework." (Caravita, 1994, p107)
Dias and Harris (1990) cite a number of studies showing that learners use contextual
information to approach problems. For example, testing children with syllogisms which
contained untrue propositions such as birds living in waters led to judgements based on the
empirical 'trueness' of the propositions rather than the logical properties of the syllogism.
However, Dias and Harris also offer a cure for the empirical bias. They found significant
increases in logical reasoning by using a make-believe intonation, a remote context, and
visual imagery when presenting these empirically nonsensical problems.
An alternative approach to improving strategy use is to focus on developing general
reasoning skills, such as logical inference. Adey (2003) argues that working on instruction at
the level of conceptual change is fruitless. He claims that learners lack the working memory
capacity to analyse and synthesise evidence, or to achieve knowledge transfer. He theorises
that improving working memory will foster conceptual change, and argues that cognitive
acceleration programmes achieve this improvement. However, evaluation of science
instruction programmes have not included measures of working memory.
The largest UK study to target science concept learning is the Cognitive Acceleration through
Science Education (CASE) programme, developed by King’s College. Shayer and Adey
(2002) claim that the CASE programme provides significant gain in GCSE grades for English
and maths, as well as science, compared with controls matched by cognitive attainment on
entry.) They attribute this ‘far transfer’ effect to an increase in information processing
capacity. However, these results come from a self-selecting sample of schools, who were
committed to teaching the CASE programme, and further committed to an optional
Professional Development programme. It is therefore difficult to exclude other variations in
teaching, learning and school effectiveness from this study.
However, there is evidence that a high working memory capacity is correlated with attainment
in summative tests. Jarvis and Gathercole (2003) investigated measures of working memory
as predictors of National Curriculum test results. Structural equation modelling of the data
suggested that working memory affects the latent variable 'attainment', which adequately
predicts National Curriculum levels in science, English and maths at Key Stage 3. That is,
there is no relationship in the improvement of working memory which is specific to science. In
11
Katherine Richardson
Cognitive Development and Learning
contrast, other science education researchers focus on the specific procedures of science
inference and their interaction with concepts.
For example, Kuhn, Schauble and Garcia-Mila (1992) investigated whether developments in
reasoning affected knowledge across science content domains. They undertook microgenetic studies in which feedback is provided only by engagement with the task. This is very
different from most formal science learning, which emphasise the use of external feedback
such as teacher or peer evaluation to improve learning. (Black and Harrison, 2004). They
examined learners’ uses of correct inferential reasoning to account for motion of a vehicle and
a ball using multiple variables. Over the course of the study, learners improved in both
explicit and intuitive reasoning about both situations. However, it is not clear whether these
problems represent different science content domains.
Overall, strategy use research may offer more explanatory power than conceptual change
theory where learner concepts and representations seem to be unstable and contextdependent.
Conclusion
"There is . . . plenty of theoretical speculation about how misconceptions might
possibly be converted to scientific conceptions, but there is precious little empirical
evidence of anyone actually achieving this goal.” Adey, P (2003, p21)
The dual lenses of conceptual change and strategy use help us to understand barriers to
science learning: fragmented intuitive knowledge, curriculum design, rivals to logical inference
and working memory limits. Comparing these lenses indicates the insufficiency of any one
model for explaining science learning. Theoretical and empirical research points to fruitful
directions for instructional research, including the development of explanations and the use of
meta-cognitive tools for conceptual change. However, the complex nature of the learning
being investigated makes it unlikely that this research will provide universal recommendations
regarding instructional strategies. Many instructional studies have therefore focused on
descriptions of conceptual changes as it is situated in classrooms. These provide a rich
phenomenology for further investigation.
Word count (excluding bibliography): 5250Bibliography
Adey, P (2003) Changing Minds. Educational and Child Psychology, 20(2), p19-30
Adey, P (2005) Issues Arising from the Long-Term Evaluation of Cognitive
Acceleration Programs. Research in Science Education 35, p3–22
12
Katherine Richardson
Cognitive Development and Learning
Black, P. & Harrison, C. (2004) Science inside the black box: Assessment for learning in the
science classroom London: nferNelson
Caravita, S. & Hallden, O. (1994). Re-framing the problem of conceptual change. Learning
and Instruction, 4, 89-111.
Carey, S. (1986). Cognitive science and science education. American Psychologist, 41, 11231130.
Chi M T H, Feltovich P J & Glaser R (1981) Categorization and representation of physics
problems by experts and novices. Cognitive Science 5:121-52.
Chi, M. T. H., Slotta, J. D. and de Leeuw, N. (1994). From things to processes: A theory of
conceptual change for learning science concepts. Learning and Instruction, 4, 27-43.
Department for Education and Skills (2004) National Curriculum for Science. London: HMSO.
Dias, M. & Harris, P.L. (1990). The influence of the imagination on reasoning by young.
children. British Journal of Developmental Psychology, 8, 305-318
diSessa (2000) Changing Minds. Cambridge: MIT Press
diSessa, A, Elby, A and Hammer, D (2003) ‘J’s Epistemological Stance and Strategies.’ In
Sinatra, G and Pintrich, P (eds) Intentional Conceptual Change. New Jersey: Lawrence
Erlbaum
Driver, R (1990) Everyday science: Is it right or does it work? British Journal of
Developmental Psychology, 8, 295-7
Duit, R and Treagust, D (1998) Learning in science - from behaviourism towards social
constructivism and beyond. In B. Fraser and K. Tobin (eds), International Handbook of
Science Education (Dordrecht, Netherlands, Kluwer), 3-25
Harris, P (1990) The nature of everyday science. British Journal of Developmental
Psychology, 8, 299-303
Jarvis, H.L., & Gathercole, S.E. (2003). Verbal and nonverbal working memory and
achievements on national curriculum tests at 11 and 14 years of age. Educational and Child
Psychology, 20(3), 123-140
Karmiloff-Smith, A. (1992) Beyond modularity: a developmental perspective on cognitive
science. London: A Bradford Book. The MIT Press.
Keogh, B. (2001) Concept cartoons, teaching and learning in science: an evaluation,
International Journal of Science Education, 21(4), 431 – 446
Kuhn, D, Schauble, L and Garcia-Mila M (1992) Cross-Domain Development of Scientific
Reasoning. Cognition and Instruction, 9(4), 285-327
Kuhn, D. (1993). Science as argument: Implications for teaching and learning scientific
thinking. Science Education, 77, 319–337.
Kuiper, J (1994) Student ideas of science concepts: alternative frameworks?' International
Journal of Science Education, 16(3), 279 - 292
Levin, Siegler, Druyan and Gardosh (1990) Everyday and curriculum-based physics
concepts. British Journal of Developmental Psychology, 8, 269-279
13
Katherine Richardson
Cognitive Development and Learning
Martin, M.O., Mullis, I.V.S., & Chrostowski, S.J. (Eds.)(2004) TIMSS 2003 Technical Report.
Chestnut Hill, MA: TIMSS & PIRLS International Study Center, Boston College.
Mason, L. (2001). Introducing talk and writing for conceptual change: a classroom study.
Learning and Instruction, 11(4-5), 305-329.
Novak, J (2005). Results and Implications of a 12-Year Longitudinal Study of Science
Concept Learning Research in Science Education, 35, p23–40
Ogborn, J., Kress, G., Martins, I. and McGillicuddy, K. (1996) Explaining science in the
classroom. Buckingham: Open University Press
Piaget, J. (1967). The child's conception of the world. (J. &. A. Tomlinson, Trans.).London :
Routledge
Posner, G and Gertzog, W (1982) The Clinical Interview and the Measurement of Conceptual
Change. Science Education, 66, 195-209
Posner, G, Strike, K, Hewson, P and Gertzog, W. (1982) Accommodation of a Scientific
Conception: Toward a Theory of Conceptual Change. Science Education, 66, p212-227
Prain, Vaughan and Waldrip, Bruce (2006), 'An Exploratory Study of Teachers' and Students'
Use of Multi-modal Representations of Concepts in Primary Science', International Journal of
Science Education, 28(15), 1843 – 1866
Reif, F and Allen, S (1992) Cognition for Interpreting Scientific Concepts: A Study of
Acceleration. Cognition and Instruction 9, 1-44
Reynolds, Y. & Brosnan, T. (2000). Understanding physical and chemical change: the role of
speculation. School Science Review, 81, 61-66.
Simon, S., Erduran, S. & Osborne, J.F. (2002) 'Enhancing the Quality of Argumentation in
School Science'. Paper presented at the Annual Conference of the National Association of
Research in Science Teaching, New Orleans
Shayer. M. and Adey, P. (2002) Learning Intelligence: Cognitive Acceleration Across the
Curriculum Buckingham: Open University Press
Stavy, R. & Berkowitz, L. (1980). Cognitive conflict as a basis for teaching quantitative
aspects of the concept of temperature. Science Education, 64, 679-692.
Treagust, D (1998) Conceptual Change as a Viable Approach to Understanding Student
Learning in Science. In B. Fraser and K. Tobin (eds), International Handbook of Science
Education Dordrecht, Netherlands: Kluwer, p25-32
14
Download