Sie sind auf Seite 1von 13

Original article

doi: 10.1111/j.1365-2729.2007.00259.x

Fostering elementary school students’


understanding of simple electricity by combining
simulation and laboratory activities
T. Jaakkola & S. Nurmi
Educational Technology Unit and Centre for Learning Research, University of Turku, Turku, Finland

Abstract Computer simulations and laboratory activities have been traditionally treated as substitute or
competing methods in science teaching. The aim of this experimental study was to investigate if
it would be more beneficial to combine simulation and laboratory activities than to use them
separately in teaching the concepts of simple electricity. Based on their pre-test performances,
66 elementary school students were placed into three different learning environments: com-
puter simulation, laboratory exercise and a simulation–laboratory combination. The results
showed that the simulation–laboratory combination environment led to statistically greater
learning gains than the use of either simulation or laboratory activities alone, and it also pro-
moted students’ conceptual understanding most efficiently. There were no statistical differ-
ences between simulation and laboratory environments. The results highlight the benefits of
using simulation along with hands-on laboratory activities to promote students’ understanding
of electricity. A simulation can help students to first understand the theoretical principles of
electricity; however, in order to promote conceptual change, it is necessary to challenge further
students’ intuitive conceptions by demonstrating through testing that the laws and principles
that are discovered through a simulation also apply in reality.

Keywords conceptual change, electricity, experimental study, inquiry learning, laboratory, computer
simulation .

before any formal instruction ever takes place. Although


Introduction
these intuitive conceptions are often poorly articulated,
Students build new ideas in the context of their existing internally inconsistent and highly context-dependent,
conceptual framework. Extensive research conducted they offer tremendous explanatory power for the stu-
during the last 30 years have shown that students come dents (Lee & Law 2001; Planinic et al. 2006). As a con-
to science classes with a wide range of informal ideas sequence, intuitive conceptions can be very resistant to
and conceptions that often differ considerably from change and easily interfere with students’ abilities to
accepted scientific views (e.g. Driver & Easley 1978; learn correct scientific principles; in the domain of
Shipstone 1984; McDermott & Shaffer 1992; Limon & electricity, there is a large body of research evidence that
Mason 2002). Students acquire these intuitive con- shows that students in all school levels have severe diffi-
ceptions from their everyday experiences and language culties and misconceptions in their understanding of
electric circuits even after formal instruction has taken
Accepted: 3 September 2007 place (e.g. Cohen et al. 1983; Osborne 1983; Shipstone
Correspondence: Tomi Jaakkola, Educational Technology Unit,
University of Turku, 20014-University of Turku, Finland. Email: tomi. 1984; McDermott & Shaffer 1992; Shepardson & Moje
jaakkola@utu.fi 1999). Therefore, learning of complex science issues

© 2007 The Authors. Journal compilation © 2007 Blackwell Publishing Ltd Journal of Computer Assisted Learning (2008), 24, 271–283 271
272 T. Jaakkola & S. Nurmi

such as electricity often requires not only acquisition situation, and this conflict is necessary for conceptual
of new knowledge, but changes in students’ deeply change to take place.
entrenched intuitive conceptions as well. This kind of Resnick (1998) has argued that it is the active
learning is referred to as conceptual change (see Limon manipulation – rather than the physicality as such – that
& Mason 2002, for multiple perspectives on conceptual may be the most important element in inquiry learning.
change). This would indicate that real equipment can be replaced
Since intuitive conceptions are grounded in personal by digital (or virtual) ‘manipulatives’ (Triona & Klahr
experience, educational environments should expose 2003). Indeed, recently there has been considerable
students to re-experience the phenomena, help them to interest within the science community in conducting
recognize the conflict between their prior knowledge research on the use of computer-based interactive
and new concepts, and support them during the simulations in order to help students to learn science by
demanding process of conceptual change. One promis- implementing inquiry learning (e.g. Frederiksen et al.
ing method of promoting conceptual change in science 1999; Tao & Gunstone 1999; Ronen & Eliahu 2000;
learning is inquiry-based (or discovery-based) learning Zacharia 2007; de Jong 2006). It has been argued that
(e.g. Hofstein & Lunetta 2004; de Jong 2006). Inquiry- teaching physics via computer-based models and simu-
based learning can be defined as a learning approach lations makes the content more easily graspable, espe-
that mimics authentic scientific inquiry; it involves a cially in areas that have been traditionally difficult to
process of actively exploring some realistic phenomena learn (Wiser & Amin 2002). Interactive electricity
or part of the natural world in a way that leads to asking simulation – i.e. a computer program that models the
questions, generating testable hypotheses, making dis- functioning of electric circuits – may offer several
coveries, and rigorously testing and evaluating the plau- advantages over traditional classroom teaching or
sibility of those discoveries in the search for new physical laboratory exercises (Hsu & Thomas 2002). In
understanding (de Jong 2006). general, simulations provide a safe and customizable
Inquiry learning has traditionally been implemented learning environments in which students can perform
in real laboratory environments via concrete, hands-on experiments virtually by setting up different circuits,
activities. The main aim of the laboratory activities is to changing circuit variables (such as resistance), and
provide students the means to investigate phenomena observing the outcomes of their actions (e.g. change in
through the active manipulation of physical materials, voltage). As a consequence, they may discover the prop-
as well as to open up the possibility of doing scientific erties of the model underlying the functioning of the
work as experts in authentic settings. In the area of elec- circuit (e.g. Ohm’s law). In contrast to a laboratory
tricity, students ought to infer the laws of electricity that working, a simulation can also reveal processes or
underlie circuit behaviour through physical experimen- abstract laws that are invisible in natural systems – for
tation with real equipment. However, although the use example, by visualizing the current flow within the
of concrete physical materials has a long history and a circuit (Hennessy et al. 2006). Furthermore, simulation
distinctive role in science teaching, many authors have can reduce the cognitive demands of physical labora-
questioned the effectiveness of laboratory work (e.g. tory experiments by providing students with a ‘cleaned-
Hofstein & Lunetta 1982, 2004; Hodson 1996; up’, idealized version of the complex and messy real
Kirschner & Huisman 1998). Hennessy et al. (2006) world, while still retaining a necessary level of theoreti-
have argued that the development of a theoretical under- cal authenticity (Hennessy et al. 2006). These capabili-
standing of complex phenomena (such as electricity) ties make simulation a promising tool for promoting
through practical manipulation can be problematic; in conceptual change (Tao & Gunstone 1999; Zacharia &
many cases students can only see what is happening on Anderson 2003; Blake & Scanlon 2007).
the surface level, while being unable to grasp the under- Even though previous research has documented the
lying processes and mechanisms that are invisible in potentially positive impact of simulations on students’
natural systems and important for theoretical under- learning performance (e.g. Ronen & Eliahu 2000;
standing (e.g. current flow). This means that students Hsu & Thomas 2002; Zacharia & Anderson 2003), the
are not necessarily experiencing the specific conflict positive effects are not self-evident (see de Jong & van
between their prior knowledge and the demands of the Joolingen 1998, for a review). The major criticism of the

© 2007 The Authors. Journal compilation © 2007 Blackwell Publishing Ltd


Simulation–laboratory combination 273

use of simulations has been that when using simula- average urban Finnish elementary school. The students
tions, students are asked to learn in a fundamentally dif- had not had any formal education on electricity before
ferent way than scientists in an authentic environment the study took place.
do (Steinberg 2000). The other concern has been that a
simulation may oversimplify complex systems (Crook
Study topic
1994). Moreover, students do not always believe that the
laws and principles that a simulation demonstrates will In order to develop a proper understanding of electric
also apply in the real world (Couture 2004). In these circuits in students, many researchers have suggested
cases, simulations are not enough, and more authentic, that students should first adopt qualitative and a voltage-
stronger experiences may be needed to overcome the centred view of electric circuits (e.g. Cohen et al. 1983;
emotional barriers related to the processes of concep- White & Frederiksen 1990; Lee & Law 2001). A proper
tual change (Merenluoto & Lehtinen 2004; Sinatra & understanding of the concept of potential difference
Mason in press). As a solution, combining and linking (voltage) is very important in understanding electric
simulation activities with concrete hands-on activities circuits, because the qualitative rules that govern the
may increase the creditability of the simulations. The voltage redistribution process are related to well-known
results of a recent study by Zacharia1 (2007) indicate laws of quantitative circuit theory, namely Kirchhoff’s
that, in order to gain deeper understanding of complex Voltage Law and Ohm’s Law (White & Frederiksen
issues such as electricity, students may need to experi- 1990; Reiner et al. 2000). Therefore, as the study was
ence both virtual and physical experimentation (see also the first formal introduction to the subject of electricity
Hennessy et al. 2006). This finding suggests that we for these students, the aim was to familiarize students
should not see computer simulations and laboratory with simple electricity by teaching them the definition
activities as substitute or competing methods in science of a closed circuit and the principles for the potential
teaching (as is performed traditionally) but as being difference within series and parallel circuits. The focus
complementary to one another. Unfortunately, there was on acquiring a qualitative understanding of circuits.
exists a lack of studies investigating the impact that Teaching focused on the establishment of a relationship
these two environments can have on students’ learning between observable variables, namely the number of
when used together. batteries and bulbs, the circuit configuration and the
The aim of this experimental study is to investigate if variations in bulb brightness. An understanding of cir-
it would be more beneficial to combine simulation and cuits containing batteries, wires and bulbs (or resistors)
laboratory activities than to use them separately in requires students to reason about electrical potentials at
teaching simple electricity. It is hypothesized that the different points within a circuit, as well as to examine
simulation–laboratory combination environment will the flow of electrical charge that is caused by differences
lead to greater learning gains than the use of either simu- in these electrical potentials. Students should learn that
lation or laboratory activities alone. Furthermore, it is lightning depends not only on the number of batteries
expected that the use of simulation will also result in and bulbs in a circuit, but also on circuit configuration.
better learning outcomes than the use of laboratory
equipment. In this context, the main research questions
Procedure
investigated in this study are ‘Would it be better to
combine simulation and laboratory activities than to The empirical phase of the study took place during 2
use them separately?’ and ‘How do these three environ- weeks. In the first session, which took 1.5 h, the students
ments affect students’ conceptual development in were given pre-tests. Based on their subject knowledge
understanding electricity?’ pre-test scores, students were placed into laboratory,
simulation and combination learning environments. In
Methods order to ensure that each environment had the same
spread of achievement – i.e. that students in all three
Sample
learning environments were equal at the baseline – the
The participants were 66 fourth and fifth grade students placing was performed by first classifying students
(10–11 years old; 29 boys and 35 girls) from one according to their pre-test scores, then selecting them

© 2007 The Authors. Journal compilation © 2007 Blackwell Publishing Ltd


274 T. Jaakkola & S. Nurmi

in groups of three (e.g. the first three students with the with an online electricity simulation, the ‘Electricity
lowest scores, then next three, and so on) and allocating Exploration Tool’ (see Fig 2). The representation level
these three randomly into one of the three learning of the simulation is semi-realistic, as it displays circuits
environments. on a diagrammatic level. With the simulation, students
The actual intervention lasted two consecutive hours are able to set up various circuits easily by dragging
and took place in the second week. In the beginning of wires, bulbs and resistors into desired points in the cir-
the intervention, students received a 15-min general cuits with simple mouse moves; the battery voltage can
introduction to the subject of electricity. This introduc- be changed this way as well. They can also conduct dif-
tion was identical in all three learning environments. ferent electric measurements with a multimeter simply
After the introduction, the students worked with various by dragging its probes onto the required testing points.
circuit assignments in randomly assigned pairs. Learn- The simulation also visualizes the current flow within
ing electricity is notoriously difficult, and previous the circuit. After making a certain configuration with the
studies have shown that working in pairs can be espe- circuit, students can observe the effects of their actions
cially effective when the work requires complex and get instant feedback.
problem-solving processes (e.g. Light & Glachan 1985; 3 Students in the Combination Environment (n = 22) also
Lou et al. 2001). worked in a computer suite. Here students used both the
In order to measure and compare the effectiveness of Electricity Exploration Tool and laboratory equipment
the different learning environments, a subject knowl- kits. Students were first asked to complete the assign-
edge post-test was administered to students 1 day after ment using the simulation and then, after succeeding
the intervention. Although students worked in pairs with the simulation, they were asked to repeat the
during the intervention phase, they completed all of the assignment with the laboratory equipment kits that were
tests individually. located right next to the computer. It was hypothesized
that it would be easier for the students to discover the
theoretical principles using the simulation first, and then
Learning environments to transfer them into the real circuits and to see that these
laws apply outside the simulation environment as well.
1 Students assigned to the Laboratory Environment
(n = 22) solved circuit assignments in a normal class- Instructional support
room, with laboratory equipment kits that included real
batteries, bulbs, wires, switches, and a multimeter (see The most effective learning results in inquiry learning
Fig 1). are found with instructional supports that structure the
2 The Simulation Environment (n = 222) was located in a learning process, help students to plan meaningful
computer suite where students solved the assignments experimentations, and guide them in reflecting their
working and interpreting obtained experimentation
findings (de Jong 2006; Fund 2007). In this study,
instructions were given in specially designed work-
sheets that requested and guided students to construct
various circuits and conduct various electrical
measurements. Each worksheet also contained several
instructional questions that encouraged students to pay
special attention to circuit configurations and investi-
gate how the changes and differences in these configu-
rations affected bulb brightness and voltages (see
Appendix I for an example). The students were required
to take notes regarding their observations and then write
down their answer on the worksheet. The students’ pro-
Fig 1 Example of parallel circuit constructed with the laboratory gression through the worksheets was tightly controlled
equipment kit. by the teacher (the teacher remained consistent for each

© 2007 The Authors. Journal compilation © 2007 Blackwell Publishing Ltd


Simulation–laboratory combination 275

Fig 2 The Electricity Exploration Tool simulation (© digitalbrain).

learning environment): once a student pair had com- tive ability, and a subject knowledge test was used to
pleted a worksheet, they had to ask the teacher to check measure domain-specific prior knowledge. Previous
their answer, and they could proceed into the next work- studies have shown that domain-specific prior knowl-
sheet only if they had completed the previous worksheet edge (Chambers & Andre 1997) and general intellectual
correctly. There were 12 worksheets in total, and each abilities (e.g. Cohen et al. 1983; von Rhöneck & Grob
worksheet consisted of one main topic. The worksheets 1990) can affect students’ learning about basic electric-
began with a very simple task wherein students were ity, and in order to compare the effectiveness of the three
asked to construct a circuit with one battery, wires and a learning environments on students’ understanding of
bulb, and later progressed towards more challenging simple electricity, it was important to control the effects
tasks in which students had to, for example, construct a of these two background variables on learning
circuit in which the brightness of four bulbs could be outcomes.
expressed as A > B > C = D. In order to treat the learn- The subject knowledge pre-test consisted of four
ing environments equally, the circuits and circuit ele- main questions measuring students’ understanding of
ments in the worksheets were always presented both on simple series and parallel circuits. The first two ques-
a realistic and a diagrammatic level. Each pair had 2 h to tions measured students’ understanding of the concept
complete as many assignments as they could. of a closed circuit. In the first question, students had to
evaluate different circuit configurations and decide
whether the bulb would light or not. In order to be suc-
Data collection
cessful, they needed to understand that there are two
The pre-test consisted of two separate tests. The polarities associated with the electrical force supplied
Raven’s (1958) Standard Progressive Matrices (Sets by a battery, and that for this electrical force to be
A–E) test was used to measure students’ general educa- applied to a resistor (bulb), both polarities must be

© 2007 The Authors. Journal compilation © 2007 Blackwell Publishing Ltd


276 T. Jaakkola & S. Nurmi

connected, one to each of the ports of the device. In the


Data analysis
second question, students were asked to indicate how
they thought current flows within the circuit. Here they The Raven test answers were scored with the official
needed to know that once a circuit is closed, the current scoring key. Students’ answers to the subject knowl-
circulates around the circuit in a given direction, and edge questionnaires were analysed on two levels. First,
the current flow between the voltage source and they were scored quantitatively against the model
devices connected to the circuit is uniform and con- answer template (0 point for wrong answer; from 0.5 to
tinual. The third and fourth questions measured stu- 1 point for correct answer). Second, in order to obtain a
dents’ understanding of potential difference in series deeper insight into students’ conceptual understanding
and parallel circuits; i.e., how different circuit configu- of simple electric circuits in different learning environ-
rations affect bulb brightness. The third question (see ments, students’ answers to pre-test and post-test basic
Appendix II) requested students to rank five identical questions were also analysed on a qualitative level,
bulbs according to their relative brightness. Regardless and were then classified into conceptual categories
of whether the battery is perceived as ideal or real, the that characterized the qualitatively different models
relative brightness remains the same. To solve the ques- through which students understand simple electric cir-
tion, no calculations are required. It is sufficient to cuits. Data from the stimulated recall interview served
think in terms of a simple qualitative model in which as a validity confirmation for qualitative analysis.
the circuit configuration determines the magnitude of Extensive research conducted by Kärrqvist (1985),
the voltage across circuit components. This assignment McDermott and Shaffer (1992), Osborne (1983) and
was originally used in the studies of McDermott and Shipstone (1984) on students mental models of elec-
Shaffer (1992), and it has been found to be very effec- tricity served as a starting point for the qualitative cat-
tive in discriminating between students who under- egorization. The qualitative analysis was conducted by
stand the principles behind series and parallel circuits two independent raters with the inter-rater reliability
and those who do not. In the fourth question, students (Cohen’s Kappa) of 0.89. Disagreements were due to
were asked to alter bulb brightness by reconfiguring the fact that especially some of the weaker students
circuits. Even though this question was designed pri- constantly changed their reasoning pattern to suit the
marily to measure students’ understanding of series question in hand; thus, their reasoning did not consis-
and parallel circuits, it also gave us further evidence tently follow any single principle (see also, McDermott
on their conception of closed circuit (e.g. if they drew & Shaffer 1992; Shepardson & Moje 1999;
only one wire from either of the battery terminals Planinic et al. 2006). These discrepant answers were
to the bulb, thinking that was sufficient to light the negotiated and systematically coded into the lowest
bulb). model category, including characteristics of student’s
In the post-test, we repeated the pre-test subject reasoning.
knowledge questions (these will be called post-test
basic questions from now on) and added four more
challenging questions, called advanced questions.
Disregarding the fact that the circuits were more com- Results
plex (e.g. involved more bulbs, contained both series
Overall effects of learning environments
and parallel connections, and included simple voltage
calculations), the advanced questions also measured As students were matched into learning environments,
students’ conception of closed circuits and their under- all three groups were equal in general ability and prior
standing of series and parallel circuits. knowledge at baseline (see Table 1). Analysis of regres-
After the post-test, in order to increase the validity sion showed that only prior knowledge significantly
of the interpretations of students’ responses to the adjusted post-test scores (P < 0.05). Therefore, in order
subject knowledge questionnaire, 20 students were ran- to account for the variance associated with students’
domly selected for the stimulated recall interview, in prior experience with electric circuits, subject knowl-
which they were asked to explain and justify their edge pre-test variable was selected as a covariate for the
answers. post-test scores.

© 2007 The Authors. Journal compilation © 2007 Blackwell Publishing Ltd


Simulation–laboratory combination 277

Table 1. Descriptive statistics for measured variables in three learning environments and omnibus differences for these variables.

Laboratory Simulation Combination AN(C)OVA1


(n = 22) (n = 20) (n = 22)

Mean (SD) Mean (SD) Mean (SD)

Subject knowledge pre-test (Max. 8 points) 4.07 (1.80) 4.50 (1.56) 4.09 (1.45) F2,61 = 0.469, P = 0.628
Raven 42.86 (7.86) 44.35 (4.96) 43.23 (5.66) F2,61 = 0.311, P = 0.734
Subject knowledge post-test
Basic (Max. 8 points) 5.25 (1.78) 6.15 (1.67) 6.61 (1.88) F2,60 = 4.354, P = 0.017
Advanced (Max. 12 points) 5.95 (2.44) 7.30 (3.08) 8.55 (2.30) F2,60 = 8.685, P < 0.001
Total (Max. 20 points) 11.20 (3.91) 13.45 (4.30) 15.16 (3.33) F2,60 = 10.247, P < 0.001

Basic scores consist of post-test items that were identical to the subject knowledge pre-test. Advanced item scores consist of four more
challenging subject knowledge questions that were included only in the post-test. Total score is the sum score of all post-test subject
knowledge questions, thus the sum of basic and advanced scores.
1
Subject knowledge pre-test was used as a covariate for post-test scores.

Table 2. Post-test pairwise mean differences between the three learning environments.

Simulation vs. laboratory Combination vs. laboratory Combination vs. simulation

Mean Diff P ES (CI.95) Mean Diff P ES (CI.95) Mean Diff P ES (CI.95)

Basic 0.64 0.418 0.41 (⫾0.61) 1.35 0.014 0.87 (⫾0.62) 0.71 0.435 0.45 (⫾0.61)
Advanced 0.92 0.467 0.43 (⫾0.61) 2.57 <0.001 1.22 (⫾0.64) 1.65 0.038 0.78 (⫾0.63)
Total 1.56 0.273 0.53 (⫾0.62) 3.92 <0.001 1.33 (⫾0.65) 2.36 0.033 0.80 (⫾0.63)

All probabilities are Bonferroni corrected.


Mean Diff, mean difference in original scale; ES, Hedges (1981) bias corrected standardized mean difference (i.e. mean difference in
standard deviation units); CI.95, 95% confidence interval.

There was significant pre-test–post-test (basic) deve-


Students’ conceptual understanding of circuits in
lopment within each learning environment (P < 0.05).
different learning environments
As can be seen in Table 1, there were statistical differ-
ences between the learning environments in post-test From students’ responses to the pre-test and post-test
basic, advanced and total scores. Pairwise comparisons basic questions (1 to 4), we were able to derive seven
(see Table 2) revealed that the students working in the conceptual models, as shown in Table 3. These models
combination environment outperformed the students illustrate the differences in students’ qualitative under-
working in the laboratory environment in all three post- standing of simple series and parallel circuits, and they
test scores with clear margin. They also outperformed are presented in ‘approximate’ developmental order
the students in the simulation environment in post-test (approximate because there is no univocal evidence on
advanced and total scores but not in the post-test basic the order the models 3a–3d emerge).
points. There were no statistical differences between the As can be seen from Table 4, in the pre-test phase and
simulation and laboratory environments in the post-test independent of the learning environment, there were
(P > 0.05). As an additional observation, the descriptive only total of five students who perceived circuits
statistics in Table 1 show that there seems to be overall through the scientifically accepted ‘Ohm model’. The
more diversity in the laboratory and simulation environ- rest of the students’ conceptions were classified into one
ments in the post-test scores as compared with combi- of the six intuitive model categories, the clashing model
nation environment, which implies the existence of a being the most popular. In the post-test, as many as
higher dispersion of developmental rates in these two 15 (68%) students in the combination environment
learning environments. possessed correct knowledge of the Ohm model. The

© 2007 The Authors. Journal compilation © 2007 Blackwell Publishing Ltd


278 T. Jaakkola & S. Nurmi

Table 3. Students’ conceptual models of simple electric circuits.

1. Sink model. In this most primitive model, students believe that only one wire between the bulb and the battery
suffices to light the bulb. They also think that the second wire is either unnecessary or that on some circuits it simply
provides more current to the bulb. Fredette and Lochhead (1980) call this model a ‘sink model’, and Kärrqvist (1985)
and Osborne (1983) a ‘unipolar model’.
2. Clashing current model. Students recognize that for a bulb to be part of a complete circuit, its two terminals must be
connected to different terminals of the battery. This thinking is characteristic in the remaining models as well. The
mistake students make in this category is that they constantly believe that the positive charge travels from the positive
end of the voltage source to the positive end of electrical device and that the negative charge travels from source’s
negative end to device’s negative end. When these two different charges clash, they produce electricity or energy and
cause the device to work. This model corresponds to Osborne’s (1983) ‘clashing currents model’, Shipstone’s (1984)
‘model I’ and Kärrqvist’s (1985) ‘two-component model’.
3a. Consumption model. This is the first model in which students understand that current circulates around the circuit in a
given direction. This thinking is characteristic in the remaining models as well. The distinctive feature in this model is
that circuits are perceived through a very common misconception that current is gradually consumed by the circuit
components. This model corresponds to Kärrqvist’s (1985) ‘current consumption model’, Shipstone’s (1984) ‘model II,
sequence model’ and Osborne’s (1983) ‘attenuation model’.
3b. Constant current model. These students think that the current supplied by the battery stays constant throughout the
circuit, as they perceive the brightness of a bulbs as being unaffected by the number of elements or the way that
circuits are configured. This model is equivalent to the constant current source models of Kärrqvist (1985) and Cohen
et al. (1983). It also shares most of the features of Shipstone’s (1984) ‘sharing model, model III’.
3c. Surface model. Another common misconception involves concentrating on the surface features of the circuit instead
of the way in which the circuits are configured. Students assume that the amount of bulbs is always negatively or
positively correlated with the brightness. McDermott and Shaffer (1992) refer to this misconception as a ‘tendency to
focus on number of elements or branches’.
3d. Preliminary Ohm model. Few responses indicated an emerging level of understanding that the ways the circuits are
configured affects the brightness. Although these students occasionally even used the concepts of series and parallel
circuits in their language, they were not able to identify these features correctly within a circuit. Similar findings have
been reported frequently, e.g. McDermott and Shaffer (1992).
4. Ohm model. Distinctively, the students holding the ‘Ohm model’, as named by Kärrqvist (1985), are able to distinguish
between series and parallel circuits, and they understand how circuit configuration is related to bulb brightness and
battery endurance. This model corresponds to the ‘scientific view’, as it was called by Osborne (1983) and Shipstone
(1984).

Table 4. Students’ conceptual models of simple electric circuits in different learning environments.

Sink Clashing Consumption Constant Surface Preliminary Ohm Ohm

Laboratory (n = 22) Pre-test 4 8 3 0 6 0 1


Post-test 3 5 5 1 4 0 4
Simulation (n = 20) Pre-test 1 9 3 1 3 1 2
Post-test 0 2 3 2 4 2 7
Combination (n = 22) Pre-test 4 9 1 2 4 0 2
Post-test 0 1 0 2 3 1 15

conceptual shift was also apparent in the simulation three leftmost models in Table 4), and only 18% of the
environment (from the most intuitive categories on the students were able to acquire the correct Ohm model.
left towards the more coherent models on the right in In order to allow statistical comparisons, conceptual
Table 4), even though the total amount of students categories in the pre-test and post-test were transformed
(35%) holding the Ohm model stayed relatively low. In into binomial distribution; Ohm models were coded as
the laboratory environment, the conceptual shift was ‘correct’ models (1) and the other six intuitive models
modest, as the majority of the students still perceived were coded as ‘incorrect’ (0). The McNemar test for
circuits through the three most elementary models (the paired proportions showed that the increase in correct

© 2007 The Authors. Journal compilation © 2007 Blackwell Publishing Ltd


Simulation–laboratory combination 279

Ohm models between the pre-test and the post-test was test was statistically significant only in the combination
statistically significant only in the combination environ- environment. Even though the number of correct
ment (P < 0.05). The chi-squared test with Yates conti- models in the simulation environment remained rela-
nuity correction revealed that the total proportion of tively low in the post-test, the conceptual shift towards
correct Ohm models in the post-test was significantly more coherent understanding was observable in this
higher in the combination environment (68%) as com- environment. In the laboratory environment, the con-
pared with the laboratory environment (18%), c2 (1, ceptual understanding remained on an initial model
n = 44) = 9.26, P = 0.002. The difference in the propor- level even after the intervention.
tions of correct models between the combination and Our findings suggest that we should not consider
simulation (35%) environments fell just outside statis- computer simulations and laboratory activities as sub-
tical boundaries, c2 (1, n = 42) = 3.390, P = 0.065. stitute or competing methods in teaching electricity as
Between the simulation and laboratory environment, they both have unique characteristics that are needed to
there were no statistical differences, c2 (1, n = 42) = promote deeper conceptual understanding and concep-
0.786, P = 0.38. tual change. The simulation offered two distinctive fea-
tures that appeared to have critical impact on students’
conceptual development: it provided students with an
Discussion
idealized model of a circuit, and visualized circuit
The aim of this study was to investigate if it would be functioning; the students possessing the most intuitive
more beneficial to combine simulation and laboratory conceptions (e.g. sink, clashing and consumption
activities than to use them separately in teaching the models) in the pre-test were predominantly able to over-
concepts of simple electricity. The results revealed that come these misconceptions during the intervention in
there was significant overall pre-test–post-test develop- the simulation and combination environments, but more
ment within all three learning environments. Post-test than half of the students working in the laboratory envi-
score comparisons between the three learning environ- ronment still possessed such primitive conceptions even
ments revealed that it was especially the students who after the instruction. Even though the simulation pro-
worked with both the simulation and laboratory equip- vided students with a clear and informative learning
ment (the combination environment) whose learning environment, it was also important for students to obtain
outcomes stood out from the rest of the field. The mean experience with real circuits. One explanation is that as
scores were significantly higher and the dispersion the simulation was only semi-realistic in representing
around the mean was notably smaller in the combina- circuits on a diagrammatic level, the students, particu-
tion environment when compared with the simulation larly owing to their young age at the time of our study,
and laboratory learning environments. The difference in needed assurance that the laws and principles of simula-
the dispersion indicates that the impact of the learning tion also apply in reality (cf. Couture 2004; Srinivasan
environment was relatively homogenous among the stu- et al. 2006). In other words, students may have kept the
dents in the combination environment, whereas in the laws and principles they had learned in the simulation
other two environments there were more individual dif- environment apart from their ‘real’ beliefs until these
ferences (i.e. some benefited from the environment beliefs were further conflicted in an authentic setting
whereas others did not). There were no statistical differ- with the laboratory equipment; this may have finally
ences between the simulation and laboratory learning ‘forced’ students to abandon their intuitive conceptions
environments. (Merenluoto & Lehtinen 2004).
The superiority of the combination environment was To conclude, the results of this study highlight the
even more evident on a conceptual level; the amount of benefits of using simulation along with hands-on labo-
correct models in the post-test was significantly higher ratory activities to promote students’ understanding of
as compared with the laboratory environment, and the electricity (see also Ronen & Eliahu 2000; Zacharia
difference to the simulation environment fell just 2007). A simulation can help students to first understand
outside the alpha level 0.05. The development from the theoretical principles of electricity; however, in
incorrect models of electric circuits in the pre-test to the order to promote conceptual change, it is necessary to
correct, scientifically accepted Ohm model in the post- challenge further students’ intuitive conceptions by

© 2007 The Authors. Journal compilation © 2007 Blackwell Publishing Ltd


280 T. Jaakkola & S. Nurmi

demonstrating through testing that the laws and prin- identical to real equipment. In this case, the physical
ciples discovered through a simulation also apply in experience may be less important for promoting con-
reality. Thus, the combination of simulation and labora- ceptual change.
tory exercise can bridge the gap between theory and
reality. It is important to emphasize that the results of
this study were obtained in a normal school environ- Notes
ment, not in laboratory setting, which is likely to add to 1
Zacharia’s study consisted of parts A, B and C, all of which covered different
the validity and applicability of the results. Neverthe- aspects of electricity. The control group used laboratory equipment in all parts,
less, this study has limitations that warrant further whereas the experimental group switched from laboratory equipment into elec-
tricity simulation in the beginning of part C. The overall development (from A to
research. For instance, we do not know what would have C) was significantly better in the experimental group, but no differences were
happened if the students in the combination environ- reported between the conditions for parts A and B. Instead, the post-test scores
ment had used real equipment prior to simulation. If the for part C were statistically higher in the experimental group. Therefore, one
may question as to whether it was really the combination of laboratory and
positive effect in Zacharia’s (2007) study was created
simulation activities that caused the difference, as Zacharia interprets his
truly by a laboratory–simulation combination and not results, or whether it was just the simulation that caused the positive learning
by a simulation alone1, then one might expect that stu- effect. In order to make this claim (in either direction), the study would have
needed more complex design; e.g. additional experimental group(s) using simu-
dents in a combination environment would do well even
lation in parts A and B.
if the order was reversed. Furthermore, it may be that 2
Only 20 students in the simulation environment completed the study as two had
students in the simulation environment would have per- to pull out owing to illness.
formed better had the appearance of the simulation been

© 2007 The Authors. Journal compilation © 2007 Blackwell Publishing Ltd


Simulation–laboratory combination 281

Appendix I Compacted example of a worksheet (translated)

© 2007 The Authors. Journal compilation © 2007 Blackwell Publishing Ltd


282 T. Jaakkola & S. Nurmi

Appendix II Subject knowledge pre-test questionnaire item example (translated)

3. Look at the circuits below and compare the brightness of the bulbs (a–e). The battery in each circuit is identical
(Max. 2.5 points).

Fund Z. (2007) The effects of scaffolded computerized


References
science problem-solving on achievement outcomes: a com-
Blake C. & Scanlon E. (2007) Reconsidering simulations in parative study of support programs. Journal of Computer
science education at a distance: features of effective use. Assisted Learning 23, 410–424.
Journal of Computer Assisted Learning doi:10.1111/j. Hedges L.V. (1981) Distribution theory for Glass’s estimator
1365-2729.2007.00239.x. of effect sizes and related estimators. Journal of Educa-
Chambers S. & Andre T. (1997) Gender, prior knowledge, tional Statistics 6, 107–128.
interest, and experience in electricity and conceptual Hennessy S., Deaney R. & Ruthven K. (2006) Situated exper-
change text manipulations in learning about direct current. tise in integrating use of multimedia simulation into sec-
Journal of Research in Science Teaching 34, 107–123. ondary science teaching. International Journal of Science
Cohen R., Eylon B. & Ganiel U. (1983) Potential difference Education 28, 701–732.
and current in simple electric circuits: a study of students’ Hodson D. (1996) Laboratory work as scientific method: three
concepts. American Journal of Physics 51, 407–412. decades of confusion and distortion. Journal of Curriculum
Couture M. (2004) Realism in the design process and credibil- Studies 28, 115–135.
ity of a simulation-based virtual laboratory. Journal of Hofstein A. & Lunetta V. (1982) The role of laboratory in
Computer Assisted Learning 20, 40–49. science teaching: neglected aspects of research. Review of
Crook C. (1994) Computers and the Collaborative Experi- Educational Research 52, 201–217.
ence of Learning. Routledge, London. Hofstein A. & Lunetta V. (2004) The laboratory in science
Driver R. & Easley J. (1978) Pupil and paradigms: a review of education: foundations for the twenty-first century. Science
the literature related to concept development in adole- Education 88, 24–54.
scent science students. Studies in Science Education 5, Hsu Y. & Thomas R.A. (2002) The impacts of a web-aided
61–84. instructional simulation on science learning. International
Frederiksen J.R., White B.Y. & Gutwill J. (1999) Dynamic Journal of Science Education 24, 955–979.
mental models in learning science: the importance of con- de Jong T. (2006) Computer simulations: technological
structing derivational linkages among models. Journal of advances in inquiry learning. Science 312, 532–533.
Research in Science Teaching 36, 806–836. de Jong T. & van Joolingen W.R. (1998) Scientific discovery
Fredette N. & Lochhead J. (1980) Student conceptions of learning with computer simulations of conceptual domains.
simple circuits. The Physics Teacher 18, 194–198. Review of Educational Research 68, 179–201.

© 2007 The Authors. Journal compilation © 2007 Blackwell Publishing Ltd


Simulation–laboratory combination 283

Kirschner P. & Huisman W. (1998) Dry laboratories in science Ronen M. & Eliahu M. (2000) Simulation – a bridge between
education: computer-based practical work. International theory and reality: the case of electric circuits. Journal of
Journal of Science Education 20, 665–682. Computer Assisted Learning 16, 14–26.
Kärrqvist C. (1985) The development of concepts by means Shepardson D.P. & Moje E.B. (1999) The role of anomalous
of dialogues centred on experiments. In Aspects of Under- data in restructuring fourth graders’ frameworks for under-
standing Electricity (eds R. Duit, W. Jung & C. von standing electric circuits. International Journal of Science
Rhöneck), pp. 215–226. IPN, Kiel, Germany. Education 21, 77–94.
Lee Y. & Law N. (2001) Explorations in promoting conceptual Shipstone D.M. (1984) A study of children’s understanding of
change in electrical concepts via ontological category shift. electricity in simple DC circuits. European Journal of
International Journal of Science Education 23, 111–149. Science Education 6, 185–198.
Light P. & Glachan M. (1985) Facilitation of problem solving Sinatra G.M. & Mason L. (in press) Beyond knowledge:
through peer interaction. Educational Psychology 5, 217– learner characteristics influencing conceptual change. In
225. Handbook on Conceptual Change (ed. S. Vosniadou).
Limon M. & Mason L. (2002) Reconsidering Conceptual Lawrence Erlbaum Associates, Mahwah, NJ.
Change: Issues in Theory and Practice. Kluwer, Dordrecht, Srinivasan S., Perez L.C., Palmer R.D., Brooks D.W., Wilson
The Netherlands. K. & Fowler D. (2006) Reality versus simulation. Journal
Lou Y., Abrami P.C. & Apollonia S. (2001) Small group and of Science Education and Technology 15, 137–141.
individual learning with technology: a meta-analysis. Steinberg R.N. (2000) Computers in teaching science: to
Review of Educational Research 71, 449–521. simulate or not to simulate? American Journal of Physics
McDermott L.C. & Shaffer P.S. (1992) Research as a guide for 68, 37–41.
curriculum development: an example from introductory Tao P. & Gunstone R. (1999) The process of conceptual
electricity. Part I: investigation of student understanding. change in force and motion during computer-supported
American Journal of Physics 60, 994–1013. physics instruction. Journal of Research in Science Teach-
Merenluoto K. & Lehtinen E. (2004) Number concept and ing 36, 859–882.
conceptual change: towards a systemic model of the pro- Triona L.M. & Klahr D. (2003) Point and click or grab and
cesses of change. Learning and Instruction 14, 519–534. heft: comparing the influence of physical and virtual
Osborne R. (1983) Towards modifying children’s ideas about instructional materials on elementary school students’
electric current. Research in Science and Technology Edu- ability to design experiments. Cognition and Instruction
cation 1, 73–82. 21, 149–173.
Planinic M., Boone W.J., Krsnik R. & Beilfuss M. (2006) White B.Y. & Frederiksen J.R. (1990) Causal model progres-
Exploring alternative conceptions from Newtonian dynam- sion as a foundation for intelligent learning environments.
ics and simple DC circuits: links between item difficulty Artificial Intelligence 42, 99–157.
and item confidence. Journal of Research in Science Teach- Wiser M. & Amin T.G. (2002) Computer-based interac-
ing 43, 150–171. tions for conceptual change in science. In Reconsidering
Raven J.C. (1958) Standard Progressive Matrices. Sets A, B, Conceptual Change: Issues in Theory and Practice
C, D and E. H.K. Lewis & Co. Ltd., Cambridge, UK. (eds M. Limon & L. Mason), pp. 357–387. Kluwer,
Reiner M., Slotta J.D., Chi M.T.H. & Resnick L.B. (2000) Dordrecht, The Netherlands.
Naïve physics reasoning: a commitment to substance-based Zacharia Z.C. (2007) Comparing and combining real and
conceptions. Cognition and Instruction 18, 1–34. virtual experimentation: an effort to enhance students’ con-
Resnick M. (1998) Technologies for lifelong kindergarten. ceptual understanding of electric circuits. Journal of Com-
Educational Technology Research and Development 46, puter Assisted Learning 23, 120–132.
43–55. Zacharia Z.C. & Anderson O.R. (2003) The effects of an inter-
von Rhöneck C. & Grob K. (1990) Psychological aspects of active computer-based simulation prior to performing a
learning about basic electricity. In Learning and laboratory inquiry-based experiment on students’ concep-
Instruction. European Research in an International tual understanding of physics. American Journal of Physics
Context, Vol. 2.2 (eds H. Mandl, E. De Corte, N. Bennett & 71, 618–629.
H.F. Friedrich), pp. 589–603. Pergamon, Oxford, UK.

© 2007 The Authors. Journal compilation © 2007 Blackwell Publishing Ltd

Das könnte Ihnen auch gefallen