Development Implementation and Assessment of A Gui

PHYSICAL REVIEW PHYSICS EDUCATION RESEARCH 14, 020115 (2018)
Development, implementation, and assessment of a guided-inquiry

teaching-learning sequence on vector calculus in electrodynamics
Laurens Bollen,1,* Paul van Kampen,2,† and Mieke De Cock1,‡
1
KU Leuven, Department of Physics and Astronomy & LESEC,
Celestijnenlaan 200c, 3001 Leuven, Belgium
2
Centre for the Advancement of STEM Teaching and Learning & School of Physical Sciences,
Dublin City University, Glasnevin, Dublin 9, Ireland
(Received 15 February 2018; published 9 November 2018)
We have developed, implemented, and assessed a teaching-learning sequence that aims to enhance
students’ understanding of the divergence and curl of electromagnetic fields. We designed guided-inquiry
worksheets based on student difficulties we identified during semiquantitative and qualitative studies and
discussions in the literature. A multiple representation approach was adopted to strengthen the link between
graphical representations of vector fields, calculations involving divergence and curl, and Maxwell’s
equations in differential form. We used the Design, Functions, Tasks framework to optimize learning with
the multiple external representations exhibited in our learning materials. These guided-inquiry worksheets
used in small-group tutorials comprise short open-ended questions that encourage discussions in
small groups of students. They were implemented in three consecutive years in a second-year
electrodynamics course at KU Leuven. The intervention was assessed using the same pretest and post-
test design adopted to evaluate learning in the original instruction. In addition, we gauged our students’
opinions on the intervention. We observed that our intervention positively affected our students’ structural
understanding of the vector operators, their ability to interpret divergence and curl in graphical
representations of vector fields, and their conceptual understanding of Maxwell’s equations in differential
form. In addition, our students indicated that they enjoyed the teaching approach, felt they learned
something from the worksheets, agreed with the difficulty level of the materials, and would like similar
tutorials on different subjects.
DOI: 10.1103/PhysRevPhysEducRes.14.020115
I. INTRODUCTION In introductory physics courses, Maxwell’s equations are

Five mathematical expressions form the foundations of typically formulated in terms of line and surface integrals.
the theory of electromagnetic phenomena: the Lorentz In intermediate and advanced courses the differential form
force law and Maxwell’s equations. The latter comprise is often more convenient. Writing the four laws in terms of
four fundamental laws, which in differential form read as divergence and curl not only increases the aesthetic beauty
follows: of the mathematical expressions, but also allows, e.g.,
straightforward derivation of the electromagnetic wave
∇ · E ¼ ϵρ0 Gauss’s law equations. In addition, Maxwell’s equations in differential
form are the starting point for gauge theory and Einstein’s
∇·B¼0 namelesslaw
theory of special relativity, both historically and in many
∇ × E ¼ − ∂B
∂t Faraday’s law physics curricula. However, a correct interpretation of these
∇ × B ¼ μ0 J þ μ0 ϵ0 ∂E
∂t Maxwell-Ampère’s law
equations is a challenging task, since it requires the use of
advanced mathematical knowledge in a physics context.
The use of mathematics in physics is an important topic
* in physics education research [1–8]. Various studies report
[email protected]
†
[email protected] that students struggle to incorporate their mathematical
‡ knowledge into physics, because they focus on equations
[email protected]
and calculations rather than on the underlying concepts
Published by the American Physical Society under the terms of [9–11]. As a result students generally develop a decent
the Creative Commons Attribution 4.0 International license.
Further distribution of this work must maintain attribution to operational understanding, since they are acquainted with
the author(s) and the published article’s title, journal citation, the mathematical processes, algorithms, and actions, but
and DOI. lack structural understanding of the meaning of the
2469-9896=18=14(2)=020115(18) 020115-1 Published by the American Physical Society

BOLLEN, VAN KAMPEN, and DE COCK PHYS. REV. PHYS. EDUC. RES. 14, 020115 (2018)
correct application of the vector operators and Maxwell’s

What are
What should
students
equations in differential form [17]. A follow-up study
students learn?
learning? probed students’ abilities to interpret, construct, and switch
Systematic between different representations of vector fields [18].
Establish Document student
course
learning goals
transformation difficulties: In this article, we intend to answer the question “How
develop and
How can we administer can we improve student learning?” The research aims of
improve student questionnaires this study may be summarized as follows:
learning? (1) develop a research-based teaching-learning sequence
on divergence and curl in mathematics and physics
Develop, implement, and
assess learning materials contexts that uses a multiple representation approach;
(2) implement the materials developed in small-group
FIG. 1. A model of research-based course transformation. tutorials that engage students and evoke student
Adapted from Ref. [14]. discussions in the electrodynamics course;
(3) assess the intervention in both a semiquantitative
and qualitative way:
mathematical entities used in physics [12,13]. For (a) evaluate the effectiveness of the intervention by
example, students who see divergence and curl merely comparing post-test results with pretest results
as the process of adding or subtracting partial derivatives of and findings after the original instruction;
a vector function have an operational understanding of the (b) determine student opinions on the tutorials.
concepts; students who see divergence as an object repre- In Sec. II, we discuss how we developed guided-inquiry
senting the source of a volume flux density, or curl as an worksheets based on the student difficulties with vector
infinitesimal rotation, that can be compared at various calculus in mathematics and physics identified in the
points and represented in various ways, have a structural literature and in our own studies. The implementation of
understanding. the tutorials at KU Leuven is described in Sec. III. In
To enhance students’ understanding of the use of vector Sec. IV the effectiveness of the intervention and results
calculus in electrodynamics, we have adopted the research- from informal evaluation by the participating students are
based course transformation procedure patterned after presented. In Sec. V we summarize the most important
Chasteen et al. [14] shown in Fig. 1. In answer to the conclusions, discuss the significance of these results, and
question “What should students learn?”, we established a discuss implications for instruction.
set of learning goals based on results from earlier studies,
teaching and learning experiences, course textbooks, and
aims of electrodynamics modules. We distinguish four II. DEVELOPMENT OF RESEARCH-BASED
skills or competencies that our students need to acquire LEARNING MATERIALS
regarding the use of vector calculus in electrodynamics: In this section we discuss the empirical and theoretical
(i) structural understanding [12,13] of the vector operators, foundations we used to develop learning materials that help
(ii) calculation of divergence and curl, (iii) graphical students improve their understanding of divergence and
interpretation of vector fields in terms of divergence and curl in electrodynamics. In Sec. II A, results regarding
curl, and (iv) conceptual understanding of Maxwell’s student difficulties with vector calculus in mathematics and
equations in differential form. physics from literature and our own studies are discussed in
In order to formulate an answer to the question “What are detail. In Sec. II B, we use the DeFT framework [19–21] to
students learning?”, we conducted multiple qualitative and explain aspects that optimize learning with multiple rep-
semiquantitative studies. Based on the list of learning goals, resentations. The key content selection is discussed in
we developed and administered written questionnaires to Sec. II C. The main principles on which the research-based
probe students’ use of the vector operators in mathematics worksheets were designed are exemplified in Sec. II D.
and electromagnetism contexts. The pretests and post-tests
were given before and after instruction, and were com-
pleted individually without recourse to textbooks or other A. Summary of results from investigations
resources. This pretest and post-test design allowed us to into student difficulties
evaluate students’ knowledge before and after the original As the learning materials are based on the student
instruction at KU Leuven and St Andrews, and document difficulties identified in our earlier studies [15–18] and
student difficulties with divergence and curl in mathemati- in the literature, we first describe the most important
cal and electromagnetic contexts at KU Leuven, St findings regarding students’ understanding of vector fields
Andrews, and Dublin City University (DCU) [15,16]. By and vector calculus in mathematical and electromagnetic
conducting eight semistructured student interviews, we contexts. We discuss student difficulties with vector field
gained a better understanding of our students’ reasoning representations (Sec. II A 1), their structural understanding
processes and were able to identify the elements that cue of the vector operators (Sec. II A 2), their ability to perform
020115-2
DEVELOPMENT, IMPLEMENTATION, AND … PHYS. REV. PHYS. EDUC. RES. 14, 020115 (2018)
calculations with divergence and curl (Sec. II A 3), their findings were confirmed: many students described diver-
skillfulness to interpret graphical representations of vector gence as “a measure for the spreading of the field” and curl
fields in terms of divergence and curl (Sec. II A 4), and their as “a measure of bending or rotation of the field.” Such
conceptual understanding of Maxwell’s equations in differ- descriptions resemble nonscientific dictionary definitions
ential form (Sec. II A 5). of both words rather than the mathematical concepts. In a
refined version of the pre- and post-test, correct and
1. Difficulties with vector field representations incorrect statements about divergence and curl were gath-
A vector field may be expressed symbolically, or ered and students were asked to indicate with which
visualized as a field line diagram or field vector plot. By statements they agreed. We found that difficulties with
conducting a research study at four European universities the concepts of divergence and curl were common in
[18], we established an overview of the difficulties students various educational contexts. Students mostly relied on
encounter when switching between algebraic and graphical operational understanding and showed little or no structural
representations of vector fields. Gire and Price [22,23] also understanding of the vector operators. To enhance students’
reported on a variety of errors students make when structural understanding, an intervention should clarify the
sketching graphical representations of vector fields, and meaning of divergence and curl.
argued that instructors should be aware that some repre-
sentational features may have two potential meanings (e.g., 3. Difficulties with performing calculations using
length meaning both distance between points and strength divergence and curl
of a field) or do not directly correspond to what is being About 60% of the KU Leuven students could calculate
represented (e.g., smaller spacing of field line diagrams
the divergence or curl of a given vector field on the pretest,
corresponds to a greater magnitude of the field). Likewise
which is a high percentage compared to the success rate on
Fredlund et al. [24] deemed it important that instructors
questions related to the other skills and competencies. This
take time to explain the intended meaning of all the features
shows students had been well trained in mathematical
of a representation. We share the view that students may
techniques in mathematics courses. Since their perfor-
benefit from instruction that focuses on teaching students
mance in the post-test was even better, we decided that
how to construct, interpret, and switch between symbolic
no further intervention was needed regarding the ability to
expressions of vector fields, field line diagrams, and field
perform calculations with divergence and curl.
vector plots.
2. Difficulties with the conceptual meaning 4. Difficulties with interpreting divergence and curl in
of the vector operators graphical representations
A limited amount of research deals with students’ In Sec. II A 1 we discussed students’ difficulties with
understanding of vector operators. Manogue and Dray vector field representations. Singh and Maries [27] dis-
[25] pointed out that in mathematics courses gradient, cussed how graduate students struggle with the interpre-
divergence, and curl are used in a general and abstract way, tation of divergence and curl in field vector plots. They
while in physics courses they are mostly used in certain argued that course materials often focus strongly on solving
symmetries (Cartesian, cylindrical, or spherical). They mathematically challenging problems rather than on clar-
called this the “vector calculus gap,” and argued that it ifying the underlying concepts. Similar findings were
is important to eliminate the gap as such differences in reported by Ambrose [28], when he asked students to
teaching approaches between mathematics and physics decide whether curl was zero or nonzero in several field
courses confuse students. vector plots. These findings were confirmed and extended
We have conducted studies [15,16] at KU Leuven, St in our studies, in which we asked our students to discuss
Andrews, and DCU in which we probed second-year where the divergence or curl vanished in certain field vector
physics and mathematics students’ concept images of plots. [15] Both before and after instruction, less than half
divergence and curl. The concept image is defined as a of them gave a correct response. Students were often
unique collection of all the mental processes activated when inconsistent in their problem-solving strategies and tended
a person encounters a certain concept [26]. An open-ended to rely on their intuitive ideas about divergence and curl that
pretest question asked students to “interpret (i.e., write often contradicted mathematical definitions. During the
down everything you think of when you see) the following interviews it became clear that many students tackle
operations: ∇ · A and ∇ × A.” When answering this ques- problems that involve vector fields by initially making a
tion, about half of our students wrote a symbolic expression field line diagram or field vector plot of the situation [17].
for divergence and curl, but fewer than 20% gave a Since sketching a situation is an important heuristic in
conceptual explanation. In addition, students’ descriptions physics problem solving, we decided that our learning
of what divergence and curl represent were mostly inaccu- materials would focus on how to switch between various
rate or incorrect. In individual interviews [17], these graphical and symbolical representations of vector fields
020115-3
and how to interpret divergence and curl in field vector which elements we focus in the tutorials and give some
plots and field line diagrams. excerpts from the worksheets.
5. Difficulties with interpreting Maxwell’s B. A multiple representation approach:

equations in differential form The DeFT framework
When using divergence and curl in an electromagnetism The information presented in the previous section
context, it is appropriate to apply Maxwell’s equations in suggests that our students experienced difficulties with
differential form. However, it is also possible to visualize linking calculations, field vector plots, field line diagrams,
the field or set up an algebraic expression of the vector and Maxwell’s equations in differential form. In addition,
field and then determine the divergence or curl. A research they struggled to use multiple symbolic and graphical
project at the University of Colorado [29] included a representations in electrodynamics. However, being able
question where students are asked to determine where to interpret various representations and being able to switch
the divergence of an electric field vanishes for a thin, between formats of concepts and problems are important
nonconductive uniformly charged spherical shell. This type skills that correlate with a deep understanding of the topic
of question can be solved with the differential form of [32,33]. Others have reported that learners similarly did not
Gauss’s law in a straightforward way. However, the authors exploit the advantages related to the use of multiple
reported that only 26% of their students gave a correct representations in contexts different from ours [34–36],
answer. This was in stark contrast to their calculational or struggled to switch between different representations.
success, as students scored 90% on average when asked to [18,33,37] Since using multiple representations is generally
compute the divergence and curl of a vector field given in accepted to be an effective instructional approach when
Cartesian coordinates. Baily et al. [30,31] also showed that aiming to improve students’ understanding of a certain
students in an intermediate electromagnetism course strug- concept or operation, [38–46] the materials we developed
gle with the concept divergence, and added that learning strongly focus on students learning to link calculations,
about divergence in the context of Gauss’s law does not graphical representations, and Maxwell’s equations.
necessarily translate to students’ understanding in other In literature about multiple representations, one often
contexts like the continuity equation. distinguishes between internal and external representations.
These results are in line with our findings from analyzing Zhang and Norman [47,48] explained that internal repre-
written questionnaires [15] and conducting individual sentations are mental meanings that have to be retrieved
student interviews [17]. In both studies, we asked students from memory, like the numerical value of the symbol 7.
to interpret situations involving electromagnetic fields in External representations can be perceptually inspected
terms of divergence and curl. This type of question can be from the environment, like spatial relations between
solved using Maxwell’s equations in differential form. symbols. Internal and external representations can be
However, as stated before, many students initially made transformed into each other by externalization and inter-
a sketch of the situation. Both the semiquantitative and nalization, respectively. Zhang and Norman pointed out
qualitative study showed that students struggled to apply all that it is important to be aware of the differences and
four laws correctly and lacked conceptual understanding of dynamics between internal and external representations,
Maxwell’s equations. Their difficulties were often related even though most studies only consider one of them.
to the local character of the laws. For example, many The majority of research in physics education concerns
students stated that the divergence is nonzero everywhere in external representations. Ainsworth [34,36] discussed
space when there is a charge distribution somewhere in the importance of using multiple external representations
space. The final goal of our intervention is therefore to (MER), and described how they can support learning and
improve students’ understanding of Maxwell’s equations in problem solving. In other papers, Ainsworth and colleagues
differential form and enhance their ability to apply those [19–21] developed and discussed the DeFT framework,
laws in situations involving electromagnetic fields. which can be used to characterize multirepresentational
One of the most interesting findings of the qualitative learning environments. The Design, Functions, Tasks
part of our study was that many students gave different (DeFT) framework sets out a number of questions that
responses depending on the approach they used: a designers and teachers should ask in order to provide
calculation, the interpretation of a graphical representa- effective learning experiences for their students.
tion, or applying Maxwell’s equations. Moreover, the To benefit from MERs, students should understand how
interviewees were unsure which result was correct and every representation presents information, how different
struggled to argue how all three approaches could lead to representations relate to each other, and how the most
the same response. As a result, our learning materials appropriate representation would be chosen in a specific
strongly rely on using multiple representations and aim to context. Three functions of MERs may be distinguished:
clarify the link between the different problem-solving to complement, constrain, and construct [34–36]. MERs
strategies. In the next sections, we discuss in detail on can be used in complementary roles, where each of the
020115-4
representations may contribute to the information or often than not this approach led students to erroneous
process, like when velocity is represented by both an conclusions: they associated diverging field lines with
equation and a v, t graph. The second function of MERs nonzero divergence of the electric field, and bending field
is to improve students’ understanding of a domain by using lines with nonzero curl of a magnetic field. Most students
one representation that constrains their interpretation of a were able to revise their reasoning after they had been
second representation. For example, when students make prompted to use symbolic representations and had been
the common error of thinking that an object is at rest when given a vector field plot representation. We illustrate in
the v, t graph is a horizontal line, a teacher might employ Sec. II D how these findings influenced the design of two
the constraining, more familiar representation of enacting worksheets.
motion at constant speed while showing a real-time graph There are various teaching approaches that could be
to support the students’ interpretation of the less familiar used when intervening in a physics course (examples can
graphical representation. Finally, MERs can help students be found in Ref. [55]). We opted for an approach based on
to develop a structural understanding (“reification” in the guided-inquiry worksheets, each consisting of a set of
parlance of Ainsworth), which is the function that is most structured questions that guide students to reconstruct their
appropriate for our project. concept images, learn various solution strategies, learn to
The various possible functions of MERs leave many switch between different representations, and link math-
ways to design multiple representation approaches that vary ematical entities to physical phenomena. Many studies
in terms of content, target users, and teaching strategies. [56–66] have confirmed that such an actively engaging
While the choice of approach may influence learning approach is generally more effective than traditional
outcomes, Ainsworth and colleagues [19–21] argued that instruction.
there is a set of five design parameters that uniquely applies Using the results of the studies mentioned above, we
to MER approaches: (a) the number of representations, established that the contents of the worksheets should focus
(b) the form of each representation, (c) the way information on helping students to do the following:
is distributed, (d) the sequence of representations, and • obtain a structural understanding of the mathematical
(e) the support for switching between representations. entities “divergence” and “curl”;
Taking into account the various functions and design • interpret visualizations of vector fields in terms of
parameters, the DeFT framework then arrives at the develop- divergence and curl using conceptual approaches:
ment of cognitive tasks. To successfully learn with MERs, for example, a box mechanism to decide where the
the student should understand (i) how every representation divergence is nonzero [67] or a paddle wheel approach
presents information, (ii) the relation between the different to determine where the curl is nonzero [68];
representations, and (iii) how to choose the most appropriate • interpret, construct, and switch between field vector
representation in a specific context. plots, field line diagrams, and algebraic expressions of
vector fields;
C. Design of the materials: Content selection • determine the divergence and curl in vector fields with
The learning materials in the teaching-learning sequence 1=r2 (spherical) and 1=s (cylindrical) symmetry,
we designed are supplementary to Griffiths’s textbook which are exceptional from a mathematical point of
[49] and use multiple graphical and symbolic representa- view, but very common in electromagnetism;
tions. Many researchers argue that instruction on vector • develop a conceptual understanding of Maxwell’s
calculus should use methods that facilitate linking symbolic equations in differential form, with a focus on the
expressions and visualizations [30,31,50–54]; our studies local character of the equations;
[15,17,18] confirm these statements. By adopting a MERs • enhance their understanding of the link between
approach within the DeFT framework we aimed to improve different strategies to determine the divergence and
students’ abilities to interpret, construct, and switch curl, in both mathematics and physics contexts:
between different representations of vector fields, and help — interpreting graphical representations of vector
them make sense of the characteristics of divergence and fields in terms of divergence and curl;
curl in visualizations of vector fields. As our studies — calculating and interpreting mathematical expres-
revealed that vector fields with spherical and cylindrical sions involving vector operators;
symmetry are very challenging for students, and such fields — applying Maxwell’s equations in differential form.
are very common in physics, the worksheets address such We designed five sets of worksheets that can be
fields repeatedly. implemented in an electrodynamics course. Although they
In individual student interviews [17], we helped students can be used independently, we think learning will be
gain insight into the divergence and curl of electric and optimal if all parts are used. The first two worksheets
magnetic fields. When answering questions on these topics concern divergence and curl, respectively, from a mainly
students would often spontaneously draw field lines and mathematical point of view. Therefore, these tutorials could
use this representation in their reasoning. However, more also be useful for mathematics instruction. The remaining
020115-5
three tutorials link these mathematical concepts to a physics symmetry are global properties of the field easily gleaned
context, and concern the differential forms of Gauss’s law, from this representation. The field line diagram represen-
Faraday’s law, and ∇ · B ¼ 0 and Maxwell-Ampère’s law, tation, on the other extreme, provides qualitative informa-
respectively. All sets of worksheets comprise between two tion about the variation of the field strength and direction in
and five pages. a region of space, and does so in a way many students have
The guided-inquiry questions on the worksheets vary in come to find intuitive. While field line diagrams allow for
terms of format, including but not limited to questions that easy identification of the location of sources and sinks, they
should be answered in a few short sentences, computational indicate strength and direction of fields only qualitatively,
problems, questions in which students are asked to con- do not allow for precise manipulations, and are correlated
struct a sketch or diagram, questions in which students are with incorrect notions of divergence (see Sec. II A 2). In
asked to discuss student statements, hypothetical debate between these two extremes lies the field vector plot
problems [69], and questions which ask for an opinion that representation, which is the least familiar to our students.
requires a deep understanding of a concept or solution While not as precise or easily manipulable as the symbolic
strategy. Typically, the questions try to provoke common representation, relative field strength and direction are
student difficulties and then confront students with a displayed on grid points in a way that provides easy access
situation that is not in line with the incorrect ideas. to both point-by-point and global information.
Some examples are given in Sec. II D. By using the All information is distributed between these three
worksheets in a tutorial format, students are encouraged representations, of which two are graphical and one is
to engage in discussions, and many questions include symbolic. Thus our MERs are partly homogeneous and
statements that may evoke such conversations. partly heterogeneous. We chose not to include textual,
Since the tutorials were implemented during multiple pictorial, or audiovisual resources, since we believed that
years, we were able to make extensive field notes about the the three representations chosen afforded our students
use of the worksheets and the issues that students encoun- a sufficient number of productive vantage points. A
tered when discussing the questions and their responses. general description of sequencing and switching between
After each year, we improved or reformulated some of the representations—dimensions (d) and (e)—is probably not
questions by taking experiences from tutoring and the post- fruitful; instead, we will illustrate them in a discussion of
test results into account. Consequently, the worksheets three excerpts from the tasks—the tutorials. We will also
were developed using an iterative approach. The final use these excerpts to give specific instances of the functions
version of all five sets of worksheets can be found online of MERs: to complement, constrain, and construct.
[70]. An instructor’s guide is available upon request. An excerpt from the first set of worksheets, the diver-
In the next section we discuss a few typical excerpts from gence tutorial, is shown in Fig. 2. Here we opted to use just
the worksheets that show how the aforementioned focus the two representations that had proven fruitful for the
points are brought into practice. students we interviewed: a vector field plot and a symbolic
representation. Since students had spontaneously used a
D. Examples of tutorial worksheet design (different) graphical representation, we started with the
To give our students every opportunity to benefit from vector field plot representation. We asked students to make
MERs, we carefully took into account all five design a prediction based on this representation and only then
dimensions that are distinguished in the DeFT framework introduced the symbolic representation which allowed
when developing the tutorials. The choice of the number of them to calculate the divergence at any position. Thus
representations—dimension (a)—was informed both by the dimensions (a)–(d) are easily recognized.
aims of our instruction and our students’ prior attainment Dimension (e), switching, is where the function of the
and dispositions. We decided to include three representa- MERs comes to the fore and the MER approach becomes
tions in our tutorials: field line diagrams, field vector plots, more than the sum of its parts. Here we used a deeper
and symbolic expressions. The three representations each exploration of the symbolic expression to eventually
provide important information about the vector fields in constrain possible interpretations of the vector field plot
different forms—dimension (b). Each representation con- representation. The calculation of divergence is straightfor-
veys somewhat different information in different ways and ward away from the source of this field. Only at the location
to a different extent—dimension (c). Symbolic expressions of the source the divergence of the field is nonzero. We
are precise, analytical, mathematical expressions that chose to give the students a hint that they may need the
define a vector field or represent Maxwell’s equations delta distribution to describe the nonzero divergence
point by point for all space. This representation allows for mathematically. For many students the calculations chal-
precise manipulation and calculation, and contains all lenge their intuitive ideas about divergence being nonzero
information about the vector field in an abstract way with everywhere in a 1=sŝ cylindrical field. While the worksheet
explicit use of coordinate systems, some of which may be leaves it open to students to reconcile their ideas in any way
unfamiliar to students. Only in situations of extremely high they want, the tutors are primed to encourage them to
020115-6
FIG. 2. Excerpt of a worksheet used in the divergence tutorial.

In the original format, some blank space was left after every
subquestion for the students to formulate their responses.
qualitatively investigate the net flow per volume in the FIG. 3. Excerpt of a worksheet used in the Gauss’s law tutorial
vector field plot representation and then to apply the in differential form. In the original format, some blank space was
divergence theorem in symbolic representation. Finally, left after every subquestion for the students to formulate their
in preparation for the next tutorial students are asked to responses.
think of a physical field that behaves in this manner.
In Fig. 3, an excerpt of the Gauss’s law tutorial is
presented. It follows the divergence and curl tutorials,
and we feel that our students are now likely to be ready
to use all three representations. Our sequencing mimics that
adopted by our students during the interviews. In the
Gauss’s law tutorial, the switching dimension (e) takes
place at two levels. Students first draw their own field line
diagram, remember or derive an expression for the field in
cylindrical coordinates, and then use the knowledge
acquired in the divergence tutorial to comment on a
hypothetical student’s statement (obtained from the inter-
views) that diverging field lines indicate nonzero diver-
gence. Many students are sufficiently certain about where
divergence is zero and nonzero that they can now use these
MERs without help from the tutors to construct the under-
standing that the start and end of field lines show where
divergence is nonzero in this representation, and that field
lines spreading or coming together is irrelevant. The last two
parts of the excerpt encourage students to strengthen their
understanding of the relation between the vector field plot
representation and the symbolic representation of vector
fields.
Finally, we note that not all aspects were or must be FIG. 4. Excerpt of a worksheet used in the Maxwell-Ampère’s
tackled using MERs. For example, distinguishing the value law tutorial in differential form.
020115-7
of a field from its derivative proved a persistent difficulty supported by the teaching assistants [76], who provided
and recurrent theme in our tutorials. As shown in the guidance for students with the tutorial questions if needed
excerpt from the Maxwell-Ampère’s law tutorial (Fig. 4), and asked challenging follow-up questions. They were also
students must decide whether the magnetic field, its responsible for answering questions about the correctness
divergence, and its curl are zero or nonzero; thus the of students’ responses, since no written solutions were
notion that the derivative of a function at a point generally given to the students at any point. Student groups worked at
differs from the value of the function at that point is their own pace and were in charge of how much time was
reinforced. The students also get to consider the two spent on each part of a tutorial. Post-tests were adminis-
possible sources of magnetic fields in a single situation: tered at the end of the tutorial, so only those who attended a
a current density and a time-varying electric field. Finally, particular tutorial gave responses to the corresponding
students are asked to check their solutions for consistency post-test.
with Maxwell’s equations. While tutorial attendance was not mandatory, the instruc-
tor in charge of the course [77] strongly encouraged the
III. IMPLEMENTATION OF TUTORIALS IN AN students to take part. The tutorial content was officially
INTERMEDIATE ELECTRODYNAMICS COURSE recognized as examinable material in the course description.
Over the years, the context of both the electrodynamics
The students enrolled in the second-year undergraduate course and the intervention varied somewhat: the emphasis
electrodynamics course at KU Leuven are mainly physics during the lectures was modified, the exercises during
and mathematics majors. All have completed an introduc- the problem-solving session were replaced, the worksheets
tory electricity and magnetism course that uses the textbook in the tutorials were adapted, and the teaching assistants
by Giancoli [71] leading up to Maxwell’s equations in changed.
integral form, and at least one calculus course with a
chapter on vector calculus, following the textbook by
IV. ASSESSMENT OF THE INTERVENTION
Adams and Essex [72]. There are 13 weeks to a semester,
and attendance is not compulsory. The course content was In this section we discuss the effectiveness of the
based on Griffiths’s textbook [49]. Before the intervention, intervention and explain to what extent students perceived
the electrodynamics course comprised one 2 h lecture the tutorials as being useful and enjoyable. We are con-
session and one 1.5 h problem-solving session per week. vinced that evaluating learning materials is an important
The final exam was based mostly on exercises from the part of the research process that yields valuable informa-
problem-solving session. tion about optimizing learning in an intermediate physics
The intervention at KU Leuven was implemented in course, and that is why we include this section. However,
three consecutive academic years. The tutorials replaced the numbers must be interpreted with care. First, the
parts of lectures and problem-solving sessions on diver- number of participants was small, especially preinterven-
gence, curl, and Maxwell’s equations in differential form, tion. This is not uncommon for the postintroductory
but also required about two additional hours being spent on physics level. Second, we cannot determine with certainty
these subjects [73]. In total, about 5 h were spent on the whether self-selection effects play a role, but we can say
tutorials in a small-group setting [74]. that comparing results preintervention and postintervention
As explained in Sec. II, we developed five sets of has ecological validity. In both cases, the students were free
worksheets. At the start of the course, a pretest was given to attend lectures and tutorials, or not; and in both cases, the
to the students and the divergence and curl tutorials were examinable materials reflected the content of the tutorials.
implemented. The Gauss’s law, Faraday’s law, and It cannot be ruled out that some students who did not attend
Maxwell-Ampère’s law tutorials were implemented after some or all of our tutorials would have attended the more
the corresponding chapters in Griffiths’s textbook [49] regular tutorials, or vice versa, as we have only interviewed
were discussed during the lectures. Some time after the students who attended.
last tutorial, the post-test was given to the students. At KU In Sec. IVA, we discuss the post-test results at KU
Leuven, about 30–40 students are enrolled in the electro- Leuven in the years the intervention took place, and
dynamics course every year. However, since attendance compare these to data obtained after both the original
usually drops during the course of a semester, only about instruction and the pretest results. In all cases, the pretest
15–25 students per year completed the post-test. took place before tutorial instruction, and the post-test after
During the tutorials, the students were greatly encour- all instruction on the topics on hand. We assessed the
aged to work in small groups of two or three students and effectiveness of the tutorials at KU Leuven with regard to
discuss the contents of the worksheets, using teaching students’ structural understanding of divergence and curl
approaches adopted at the University of Washington [58] (Sec. IVA 1), their ability to interpret the vector operators
and the University of Colorado Boulder [60,65,75]. Not all in visualizations of vector fields (Sec. IVA 2), and their
students attended all tutorials; there were no alternative conceptual understanding of Maxwell’s equations in differ-
sessions for those who did not. These discussions were ential form (Sec. IVA 3). In Sec. IV B we explain how our
020115-8
students perceived the tutorials and discuss their informal

evaluation of the intervention.
A. Semiquantitative assessment of the

effectiveness of tutorials
To assess the effectiveness of the tutorials we used
the same pretest and post-test design we used to
evaluate learning after traditional instruction. The analy-
sis includes a comparison of the post-test results
(N ¼ 19) after all the original instruction discussed in
our earlier work [15] with post-test results after the three
years in which the tutorials were implemented (N ¼ 60),
where possible. In cases where we refined some of the
post-test questions we limit ourselves to comparing pre-
and post-test results.
For brevity and readability we have chosen to minimize
the numerical information on statistical significance.
However, it is important that we show that the pretest
and post-test data for all three intervention years are
equivalent and may therefore be added together. Since
we use categorical data, Pearson’s chi-squared test provides
a suitable means of discussing whether or not the overall
results were independent of each other. We found that the
post-test data after intervention were statistically the same
in the three years (all p values found by χ 2 testing were
larger than 0.05), which allowed us to aggregate them, even
though some conditions like the prior instruction, the
teaching approach during the lectures, the exercises dis-
cussed during the problem-solving sessions, and the ques-
tions asked on the final exam changed in small ways over
the years. We interpret our findings as confirmation that the FIG. 5. Questions designed to assess students’ correct and
tweaks to the lectures and tutorials were indeed “small,” incorrect conceptions of curl and divergence. The correct state-
and that we do not need to take these into account in this ments are C4, C6, C8, D1, D6, and D9.
discussion.
For the individual test items the error bars represent the
standard error p onffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
the proportion, calculated using the question was not on the preintervention post-test, we can
expression σ ¼ pð1 − pÞ=n. We have used the only test only compare responses before and after the intervention.
applicable to unpaired proportional data, a two-tailed two- Figure 6 shows the proportion of students who indicated
proportion z test, to put our conclusions on somewhat surer that each statement in the pretest and the post-test is correct.
footing. Only for products of np or nð1 − pÞ below five is Only students who took both tests were taken into account.
the test likely not valid; we have indicated where this may The error bars represent the standard error of the propor-
be problematic in the text below. tion. It is clear that students generally improved their ideas
of the concepts divergence and curl after the tutorials.
Significantly fewer students indicated that C2 [The curl is a
1. Structural understanding of divergence and curl measure for how much field lines bend (incorrect)] and C5
The first pretest and post-test questions were designed to [The curl is nonzero if and only if the direction of the field
gain insight into our students’ concept images of diver- changes (incorrect)] were correct, while more students
gence and curl. The concept image is defined as a unique checked the boxes of statements D1 [The divergence
collection of all the mental processes activated when a measures the source or sink of the field (correct)] and
person encounters a certain concept [26]. The questions D6 [The divergence indicates where field lines start or end
comprised prevalent student statements from open-ended (correct)]. However, no such improvements were seen in
questions and individual interviews. In three student inter- the most important distractor D2 [The divergence is a
views we verified that students interpreted the last version measure for how much field lines spread apart (incorrect)].
of the questions, shown in Fig. 5, as intended. They were This may in part be due to a potentially confusing statement
exactly the same on the pre- and post-test. Since this in the textbook [49]: “is a measure of how much the vector
020115-9
FIG. 6. The proportion of KU Leuven students (N ¼ 57) who

indicated that a statement presented in the questions in Fig. 5 is
correct on the post-test, compared to the pretest. Only the students FIG. 7. Question on the pre- and post-test that were given to the
who took both the pre- and post-test were taken into account. students in order to determine to what extent they can decide
Correct statements are in bold and underlined. where the divergence or curl is nonzero in a field vector plot.
Fields (c) and (I) are similar and yield the same solution, as is the
case for fields (a) and (II).
v spreads out (diverges) from the point in question.” In
addition, responses to statements about the vector or scalar
character of both operators improved significantly (C4, terms of divergence and curl. The fields on the pretest are
C8=9 and D4, D8=9), and students rarely confused similar to those on the post-test, as can be seen in Fig. 7.
divergence and gradient anymore (D3). Figure 8 shows the results of the pretest and post-test, for
To put these results into perspective, on the pretest 21% students who took part in the original instruction and for
of the students answered all questions pertaining to curl those who participated in the tutorials.
correctly, compared to 56% on the post-test. For diver- Pretest field (c) and post-test field (I) both have zero
gence, 7% of the students answered all questions correctly divergence and nonzero curl everywhere. Figure 8 shows
on the pretest compared to 32% on the post-test. This shows that students benefited from doing the tutorials, as they
that there is a significant increase in the number of correct responded significantly better on the post-test question
answers, and students improved their understanding of the after the intervention than after the original instruction.
concepts divergence and curl. However, since we lack data However, this trend cannot be seen when comparing results
about the situation after the original instruction at KU for pretest field (a) and post-test field (II), which have a
Leuven, we cannot rule out that these results are partly due nonzero divergence and zero curl everywhere. This can be
to influences other than the tutorials. explained by looking at the most prevalent error, which is
stating that the divergence or curl vanishes at a location
2. Interpretation of graphical representations in where the magnitude of the vector field is zero. Such
terms of divergence and curl difficulties, related to confusing the derivative with its
As stated in Sec. II A 4, it is of great importance that value, were more often exhibited in field (c) and field (II)
students learn to interpret field vector plots, and in the than in field (a) and field (I). Even though the worksheets
tutorials we strongly focus on approaches that may help specifically addressed these issues, these difficulties were
them to decide whether the divergence or curl is nonzero at not resolved for some students.
a certain location in a visualized vector field. On both The tutorials also affected the students’ approach to
pretest and post-test we assessed our students’ ability to solving the questions in Fig. 7. In our earlier work, we
interpret field vector plots in a mathematical context in discussed how we distinguish between concept-based,
020115-10
FIG. 8. The proportion of KU Leuven students who determined

correctly where the divergence or curl is nonzero in a field vector
plot on the pretest [field (c) and (a) in Fig. 7] versus the post-test FIG. 9. Post-test question to determine whether students can
[field (I) and (II) in Fig. 7], after the original instruction (N ¼ 19) decide where divergence and curl are nonzero in field vector plots
and the intervention (N ¼ 57). Only the students who took both of electromagnetic fields.
the pre- and post-test are taken into account.
participated in the tutorials. It seems clear that the pro-

formula-based, and description-based strategies when inter- portion of correct responses is significantly higher for the
preting field vector plots in terms of divergence and curl latter cohort for field (III); the z test is not reliable enough to
[15]. A description-based strategy refers to explanations make definitive statements for field (IV), where in any case
similar to “the divergence vanishes where the field is zero” the gains would be smaller. We also found that the students
or “the curl is nonzero where the field changes direction.” who participated in our tutorials more frequently used
Students who employ a formula-based strategy use the approaches that are generally more successful, like the
mathematical expression for the vector operators, and may conceptual approaches described above or applying
even set up an equation for the vector field to calculate Maxwell’s equations in differential form, and gave descrip-
divergence and curl. In the worksheet, we strongly focused tive answers less often. This suggests that the tutorials
on concept-based strategies, like determining the net flux positively affected our students’ approach and ability to
through a box to find out where the divergence is nonzero, determine the divergence and curl in graphical representa-
or using the paddle wheel trick to decide where the curl is tions of electromagnetic fields.
nonzero. This involved both demonstrating how these To evaluate our students’ calculational skills and their
conceptual approaches work and giving the opportunity ability to construct field line diagrams, we developed the
to apply this strategy in several exercises. Students who
participated in the tutorials used such approaches in 40%
(divergence) to 70% (curl) of the cases, whereas only about
20% of the students did so after the original instruction.
Since such conceptual approaches correlate with a high
success rate, it can explain why results after the tutorials
were generally better than after the original instruction.
The post-test questions in Fig. 9 aimed to evaluate to
what extent students correctly determine where the diver-
gence or curl vanish in electric and magnetic fields. These
are rather tricky questions, since from a mathematical point
of view the divergence and curl behave exceptionally in
vector fields that fall off as 1=s (cylindrical coordinates) or
1=r2 (spherical coordinates). While such examples were
explicitly addressed in the divergence and curl tutorials, an
answer can also be found by applying Maxwell’s equations FIG. 10. The proportion of students who determined correctly
in differential form. where the divergence and curl are nonzero in a field vector plot of
Figure 10 compares the success rate of students who an electromagnetic field on the post-test (Fig. 9), after the original
took part in the original instruction with students who instruction (N ¼ 19) and after the intervention (N ¼ 57).
020115-11
field of a point charge [Fig. 12(c)], and a correct field vector

plot [Fig. 12(d)].
In the first year of implementation, 60% (N ¼ 15) of our
students sketched a field line diagram and only a single
student did so correctly. Two others made a sketch similar
to Fig. 12(b), making a total of 20% correct answers. Most
other students constructed a field line diagram similar to
Fig. 12(c) or sketched (correct or incorrect) field vector
FIG. 11. Post-test question that evaluates students’ computa- plots. Because of these poor success rates, we adapted the
tional skills and their ability to construct a field line diagram. tutorial worksheets, and this was reflected in the results. In
the final two years (N ¼ 45), about 25% of the students
constructed a correct field line diagram and another 15%
post-test question in Fig. 11. The majority of the students
sketched a correct diagram but did not indicate the direction
solved parts (a) and (b) by imposing the conditions
of the field. However, about 30% of the students still
∇ × E ¼ 0 and ∇ · B ¼ 0, respectively. Over the years,
constructed an incorrect field line diagram, and one out of
the percentage of correct answers increased from 70% to
four students sketched a field vector plot. Therefore, giving
90%. This confirms that students’ operational conceptions
students additional exercises regarding vector fields with a
were developed relatively well.
nonzero divergence does seem to positively influence their
Students struggled to sketch field line diagrams of vector
understanding of how to construct field line diagrams of
fields with a nonzero divergence at every location. In
such fields but leaves room for improvement.
Fig. 12, some typical responses are shown: a correct field
line diagram [Fig. 12(a)], a correct field line diagram
without an indication of the direction [Fig. 12(b)], an 3. Understanding of Maxwell’s equations in
incorrect field line diagram which resembles the electric differential form
To assess to what extent our students conceptually
understood Maxwell’s equations in differential form, we
asked them to determine whether the divergence or curl
vanishes in ten situations involving electromagnetic fields.
The post-test questions describing these situations are
shown in Fig. 13.
When Maxwell’s equations in differential form are
applied correctly in every situation of Fig. 13, one obtains
the answers: ∇ · E ¼ 0 always except for E1; ∇ × E ¼ 0 in
situation E1, E3, and E5; ∇ · B ¼ 0 always; and ∇ × B ¼ 0
in M4 and M5. We did not give the students a list of
Maxwell’s equations to avoid pointing students in a
particular direction.
The fraction of students who indicated that the diver-
gence or curl of the electromagnetic fields is zero, after the
original instruction and after the tutorials respectively, is
shown in Fig. 14. It is clear from Fig. 14(a) that students
score significantly better after the intervention when
determining the divergence in situations E2 and E3.
Figure 14(b) shows that students also more often evaluate
the curl in situation E2 correctly. However, even after the
tutorials, only 20% of the students showed a good under-
standing of Gauss’s law by evaluating the divergence
correctly in all five situations, and 50% seemed to con-
ceptually understand Faraday’s law as they correctly
determined the curl in every situation. While the tutorials
generally had a positive effect, there is room for further
improvement.
Figure 14(c) shows that there was no significant differ-
ence in how often students indicated that the divergence of
FIG. 12. Examples of student sketches in response to post-test a magnetic field vanishes between the original instruction
question 5(c) (Fig. 11). and the intervention, for all five situations individually.
020115-12
FIG. 13. Post-test question on Maxwell’s equations.
FIG. 14. Proportion of students who indicated the divergence or curl is zero in each situation on the post-test (Fig. 13) after the original
instruction (N ¼ 19) and the intervention (N ¼ 59). Correct answers are in bold and underlined. We omitted the data after the original
instruction for situation E4, since this situation was described ambiguously at that time.
However, while fewer than half of the students correctly B. Informal evaluation
checked the box of every situation on the post-test after the On multiple occasions we asked students to formulate
original instruction, more than three out of four did so after their opinion of the tutorials: in informal talk, after a one-
the tutorials. This suggests that considerably more students on-one discussion about the correct post-test answers, and
know that the divergence of the magnetic field is always zero by inviting them to anonymously write positive and
after the intervention. No significant differences were seen negative remarks about the intervention on a piece of
when students had to decide in which of these situations the paper. The most prevalent opinions from the first two years
curl is zero [Fig. 14(d)]. Both after the original instruction were used in eight 5-point Likert scale questions, and after
and after the intervention, only about 20% applied Maxwell- the tutorial sessions in 2017 students were asked to indicate
Ampère’s law correctly in all five situations. to what extent they agreed with those statements anony-
The results show that even after explicitly discussing all mously. The results are shown in Fig. 15.
four of Maxwell’s equations in differential form during the In general, the students were very positive about the
tutorials, students still exhibit serious difficulties with tutorials. They agreed that the tutorials were enjoyable and
applying the laws in situations where an electromagnetic that they learned something from them. While some
field is described. This seems to indicate that they still students told us that the tutorials were too long or that
struggle with the conceptual meaning and local character of there were too many sessions, the last cohort of students
the equations. In Sec. V, we will further discuss the generally seemed happy with the number of tutorials, and
significance of these results. the majority was open for additional tutorials on other
020115-13
FIG. 15. Student (N ¼ 29) responses to eight 5-point Likert scale questions, evaluating their experiences with the tutorials.
topics. However, for some students the tutorials were not with the responsible instructor, and together we tried to
challenging enough. While the post-test results seem to improve the link between the contents on the worksheets
suggest that we should not make considerable changes after and the subjects of the course. This is an ongoing process,
these comments, it is important that the teaching assistants and we acknowledge that further efforts are needed to
make sure to challenge all students during the tutorials. improve the integration of the tutorials.
When asked for (other) things that students liked about Understanding Maxwell’s equations in differential form
the tutorials, many acknowledged that the questions on the is one of the learning goals of the electrodynamics course
worksheets were intuitive and forced them to think about the tutorials focus on. It is important that questions on this
the concepts rather than mathematical expressions. Some topic are asked on the final exam. In the first year of
students explicitly stated that visualizing divergence and implementation, many students complained about the lack
curl had improved their understanding of both vector of relevance of these tutorials as it would not influence their
calculus and Maxwell’s equations in differential form. final grades. Since students work toward succeeding on the
Many students also stated the tutorials helped them to final exam, this is an understandable reaction. Thankfully
understand the fundamentals of the electrodynamics course the responsible instructor was convinced of the importance
better. of assessing conceptual understanding and communicated
The most prevalent negative comments were complaints this to the students. As he adjusted some of the exam
about not getting any solution manuals for the worksheets. questions, students explicitly stated in the two subsequent
While we tried to discuss some important or difficult years that they could use ideas from the tutorials during the
exercises on the blackboard, students felt that they should exam. This indicates how relevant evaluation may improve
have access to all the correct answers. We believe this is students’ engagement during the tutorials, and change their
unnecessary and even potentially detrimental to student impression of the contents of the worksheets.
engagement during the tutorials.
While many students indicated that the tutorials were
V. DISCUSSION AND CONCLUSIONS
useful, others stated they missed a close link to the lectures
and felt there was a great contrast between the post-test Based on the difficulties identified by evaluating stu-
questions and the type of questions that are typically asked dents’ understanding of divergence and curl in mathematics
on the final exam. We agree with the students that the and electromagnetism before and after the original instruc-
integration of the tutorials in the electrodynamics course is tion, conducting individual interviews, and assessing stu-
an important issue. After each year, we discussed the results dents’ use of vector field representations, we designed a
020115-14
teaching-learning sequence that aims to help students with final exam, but the post-test was typically given to the
the use of vector calculus in electrodynamics. The guided- students halfway in the semester. In addition, further
inquiry worksheets we developed aim to improve students’ alignment of tutorials and lecture may also improve the
reasoning processes by provoking common errors based on post-test results, but we do not have sufficient data to verify
intuitive reasoning and showing them how these incorrect such a statement. Nevertheless, our students’ comments
responses contradict mathematical and physical principles. indicate that a solid link between lectures and tutorials
In addition, they demonstrate and rehearse strategies that should be one of the main points of attention in the
are based on conceptual understanding and clarify the link implementation process.
between the divergence and curl in visualizations of vector A third comment is related to the type of questions on
fields, calculations, and Maxwell’s equations in differential the pretest and post-test. Some questions can be answered
form. The worksheets were designed using a multiple by simply checking a box, others demand a relatively
representation approach, and we have discussed how the short response. This allows us to easily compare pre- and
aspects that influence the effectiveness of learning with post-test results and analyze how responses change over
multiple representations, which are identified by the DeFT the years. However, it also gives students the opportunity
framework, are exhibited in our learning materials. to guess, and it is impossible to track such actions.
There are several ways to implement the worksheets, but Therefore, it is likely that in some cases results tended to
for practical reasons we opted for a tutorial approach where converge to 50%, making differences appear less signifi-
students discuss the questions in small groups. The cant than they actually are. This effect may be quite
intervention took place in an intermediate electrodynamics strong for the questions in Sec. IVA 3, since those were
course in three consecutive years (2015–2017). While no generally perceived as being difficult. Additional quali-
solutions were given to the students, teaching assistants tative research might give a better idea of the size of this
were present to help students and check their responses, effect. It would be interesting, for example, to video tape
including for correctness. students while solving questions on the worksheets, and
To assess the effectiveness of the tutorials, we used the to analyze how their reasoning processes change over
same pretest post-test design that was adopted to evaluate time. This may also help to further improve the learning
learning after the original instruction. The results show that materials.
our students gained insight into the mathematical concepts With all these provisos our results do seem to indicate
divergence and curl, were more successful when interpret- that many students benefit from the tutorials we developed
ing field vector plots in terms of divergence and curl, and with a stronger focus on conceptual understanding and the
generally improved their understanding of Maxwell’s ability to interpret graphical representations. This is in
equations in differential form. However, the differences agreement with the literature about the use of multiple
in post-test results between the original instruction and the representations [38–46]. Students generally enjoyed work-
tutorials were not always significant. Therefore, it is ing through the worksheets. In addition, the majority also
important to make some side notes when interpreting these felt they were challenged, indicated they learned some-
results, which may be helpful for future instructors that plan thing, and seemed interested to take additional tutorials.
to use (part of) our developed learning materials. This shows that our approach engaged and motivated
The most obvious reason for the disappointing signifi- students. Therefore, we encourage instructors to either
cance of the results is the relatively small sample size. use the worksheets discussed in this paper or design their
Ideally, the type of research that we discussed is conducted own learning materials based on the information we
in large groups of students. This may be achievable for presented.
interventions in introductory courses [60,78,79], but in
intermediate and advanced courses, student numbers are
ACKNOWLEDGMENTS
almost always lower, which makes it difficult to justify
statements on a strong statistical basis [61,62,65,80]. Such We gratefully acknowledge the advice of Charles Baily
issues can be resolved in part by repeating the experiment with the design of the worksheets. We thank Wojciech De
over a couple of years, as we did in this study. Nevertheless, Roeck and the teaching assistants for making the inter-
since other conditions, like the textbook used, the teaching vention possible, and for their flexibility during the
staff, prior courses, and other factors may also change over implementation of the tutorials. We would also like to
the years, results should always be interpreted with care. thank all participating students whose cooperation made
A second issue that may have affected the significance of this research project possible. This publication was made
the results is the timing of the tutorials and the post-test. possible through funding support of the KU Leuven Fund
Many students tend only to study in the weeks before the for Fair Open Access.
020115-15
[1] T. J. Bing and E. F. Redish, The cognitive blending of [19] N. Van Labeke and S. E. Ainsworth, Applying the DeFT
mathematics and physics knowledge, AIP Conf. Proc. 883, Framework to the Design of Multi-Representational In-
26 (2007). structional Simulations, in AIED’01—10th International
[2] T. J. Bing and E. F. Redish, Analyzing problem solving Conference on Artificial Intelligence in Education, 2001,
using math in physics: Epistemological framing via war- San Antonio, Texas (IOS Press, Amsterdam, Netherlands,
rants, Phys. Rev. ST Phys. Educ. Res. 5, 020108 (2009). 2001), pp. 314–321.
[3] R. Karam, Framing the structural role of mathematics in [20] S. E. Ainsworth and N. Van Labeke, Using a multi-
physics lectures: A case study on electromagnetism, Phys. representational design framework to develop and evaluate
Rev. ST Phys. Educ. Res. 10, 010119 (2014). a dynamic simulation environment, Int. Work. Dyn. Vis.
[4] C. A. Manogue and T. Dray, Bridging the gap between Learn. (2002).
mathematics and the physical sciences, APS Forum Educ. [21] S. E. Ainsworth, DeFT: A conceptual framework for
13 (2004). considering learning with multiple representations, Learn.
[5] E. F. Redish, Problem solving and the use of math in Instr. 16, 183 (2006).
physics courses, in World View in Physics Education 2005 [22] E. Gire and E. Price, Graphical representations of vector
(2005). functions in upper-division E&M, AIP Conf. Proc. 1413,
[6] J. Tuminaro, Understanding students’ poor performance on 27 (2012).
mathematical problem solving in physics, AIP Conf. Proc. [23] E. Gire and E. Price, Arrows as anchors: An analysis of the
720, 113 (2004). material features of electric field vector arrows, Phys. Rev.
[7] J. Tuminaro and E. F. Redish, Elements of a cognitive ST Phys. Educ. Res. 10, 020112 (2014).
model of physics problem solving: Epistemic games, Phys. [24] T. Fredlund, C. Linder, J. Airey, and A. Linder, Unpacking
Rev. ST Phys. Educ. Res. 3, 020101 (2007). physics representations: Towards an appreciation of dis-
[8] B. R. Wilcox, M. D. Caballero, D. A. Rehn, and S. J. ciplinary affordance, Phys. Rev. ST Phys. Educ. Res. 10,
Pollock, Analytic framework for students’ use of math- 020129 (2014).
ematics in upper-division physics, Phys. Rev. ST Phys. [25] T. Dray and C. A. Manogue, The vector calculus gap:
Educ. Res. 9, 020119 (2013). Mathematics (Does Not Equal) physics, PRIMUS 9, 21
[9] E. Kuo, M. M. Hull, A. Gupta, and A. Elby, How students (1999).
blend conceptual and formal mathematical reasoning in [26] D. Tall and S. Vinner, Concept image and concept
solving physics problems, Sci. Educ. 97, 32 (2013). definition in mathematics with particular reference to limits
[10] J. Larkin, J. McDermott, D. P. Simon, and H. A. Simon, and continuity, Educ. Stud. Math. 12, 151 (1981).
Expert and Novice performance in solving physics prob- [27] C. Singh and A. Maries, Core graduate courses: A missed
lems, Science 208, 1335 (1980). learning opportunity?, AIP Conf. Proc. 1513, 382 (2013.
[11] B. L. Sherin, How students understand physics equations, [28] B. S. Ambrose, Investigating student understanding in
Cognit. Instr. 19, 479 (2001). intermediate mechanics: Identifying the need for a tutorial
[12] A. Sfard, On the dual nature of mathematical conceptions: approach to instruction, Am. J. Phys. 72, 453 (2004).
Reflections on processes and objects as different sides of [29] R. E. Pepper, S. V. Chasteen, S. J. Pollock, and K. K.
the same coin, Educ. Stud. Math. 22, 1 (1991). Perkins, Observations on student difficulties with math-
[13] J. Tuminaro, A cognitive framework for analyzing and ematics in upper-division electricity and magnetism, Phys.
describing introductory students’ use and understanding of Rev. ST Phys. Educ. Res. 8, 010111 (2012).
mathematics in physics, Ph. D. thesis, University of [30] C. Baily, M. Dubson, and S. J. Pollock, Research-
Maryland, 2004. based course materials and assessments for upper-
[14] S. V. Chasteen, K. K. Perkins, P. D. Beale, S. J. Pollock, division electrodynamics (E&M II), AIP Conf. Proc. 54,
and C. E. Wieman, A thoughtful approach to instruction: 54 (2013).
Course transformation for the rest of us, J. Coll. Sci. Teach. [31] C. Baily and C. Astolfi, Student reasoning about the
40, 70 (2011). divergence of a vector field, Proceedings of the Physics
[15] L. Bollen, P. van Kampen, and M. De Cock, Students’ Education Research Conference 2014, Minneapolis, MN
difficulties with vector calculus in electrodynamics, Phys. (AIP, New York, 2015), pp. 31–34.
Rev. ST Phys. Educ. Res. 11, 020129 (2015). [32] D. Rosengrant, E. Etkina, and A. Van Heuvelen, An
[16] C. Baily, L. Bollen, A. Pattie, P. van Kampen, and M. overview of recent research on multiple representations,
De Cock, Student thinking about the divergence and curl in AIP Conf. Proc. 883, 149 (2007).
mathematics and physics contexts, Proceedings of the [33] J. F. Wagner, C. A. Manogue, and J. R. Thompson, Rep-
Physics Education Research Conference 2015, College resentation issues: Using mathematics in upper-division
Park, MD (AIP, New York, 2015), pp. 51–54. physics, AIP Conf. Proc. 1413, 89 (2012).
[17] L. Bollen, P. van Kampen, C. Baily, and M. De Cock, [34] S. Ainsworth, The functions of multiple representations,
Qualitative investigation into students’ use of divergence Comput. Educ. 33, 131 (1999).
and curl in electromagnetism, Phys. Rev. Phys. Educ. Res. [35] S. Ainsworth, P. Bibby, and D. Wood, Examining the
12, 020134 (2016). effects of different multiple representational systems in
[18] L. Bollen, P. van Kampen, C. Baily, M. Kelly, and M. learning primary mathematics, J. Learn. Sci. 11, 25 (2002).
De Cock, Student difficulties regarding symbolic and [36] S. Ainsworth, The educational value of multiple represen-
graphical representations of vector fields, Phys. Rev. Phys. tations when learning complex scientific concepts, in
Educ. Res. 13, 020109 (2017). Visualization: Theory and Practice in Science Education,
020115-16
edited by J. K. Gilbert, M. Reimer, and M. Nakhleh [58] L. C. McDermott and P. S. Shaffer (PEG at the University
(Springer, Netherlands, 2008), pp. 191–208. of Washington), Tutorials in Introductory Physics (Pren-
[37] R. Even, Factors involved in linking representations of tice Hall, Upper Saddle River NJ, 2002).
functions, J. Math. Behav. 17, 105 (1998). [59] C. A. Manogue and K. S. Krane, Paradigms in physics:
[38] J. I. Heller and F. Reif, Prescribing effective human Restructuring the upper level, Phys. Today 56, 53 (2003).
problem-solving processes: problem description in phys- [60] N. D. Finkelstein and S. J. Pollock, Replicating and under-
ics, Cognit. Instr. 1, 177 (1984). standing successful innovations: Implementing tutorials in
[39] A. Van Heuvelen, Learning to think like a physicist: A introductory physics, Phys. Rev. ST Phys. Educ. Res. 1,
review of research-based instructional strategies, Am. J. 010101 (2005).
Phys. 59, 891 (1991). [61] C. Singh, Interactive learning tutorials on quantum me-
[40] R. J. Dufresne, W. J. Gerace, and W. J. Leonard, Solving chanics, Am. J. Phys. 76, 400 (2008).
physics problems with multiple representations, Phys. [62] S. V. Chasteen, S. J. Pollock, R. E. Pepper, and K. K.
Teach. 35, 270 (1997). Perkins, Transforming the junior level: Outcomes from
[41] P. C.-H. Cheng, Unlocking conceptual learning in math- instruction and research in E&M, Phys. Rev. ST Phys.
ematics and science with effective representational sys- Educ. Res. 8, 020107 (2012).
tems, Comput. Educ. 33, 109 (1999). [63] D. E. Meltzer and R. K. Thornton, Resource letter ALIP-1:
[42] A. Van Heuvelen and X. Zou, Multiple representations of Active-learning instruction in physics, Am. J. Phys. 80,
work energy processes, Am. J. Phys. 69, 184 (2001). 478 (2012).
[43] D. Rosengrant, Case study: Students’ use of multiple [64] L. Doughty, Designing, implementing and assessing
representations in problem solving, AIP Conf. Proc. guided-inquiry based tutorials in introductory physics,
818, 49 (2006). Ph.D. thesis, Dublin City University, 2013.
[44] C. J. De Leone and E. Gire, Is instructional emphasis on the [65] C. Baily, M. Dubson, and S. J. Pollock, Developing tutorials
use of non-mathematical representations worth the effort?, for advanced physics students: Processes and lessons
AIP Conf. Proc. 818, 45 (2006). learned, in 2013 Phys. Educ. Res. Conf. Proc. (American
[45] P. B. Kohl and N. D. Finkelstein, Effect of instructional Association of Physics Teachers, 2014) pp. 61–64.
environment on physics students’ representational skills, [66] C. Singh and E. Marshman, Review of student difficulties
Phys. Rev. ST Phys. Educ. Res. 2, 010102 (2006). in upper-level quantum mechanics, Phys. Rev. ST Phys.
[46] J. F. Wagner, Transfer in pieces, Cognit. Instr. 24, 1 (2006). Educ. Res. 11, 020117 (2015).
[47] J. Zhang and D. A. Norman, Representations in distributed [67] An approach based on the divergence theorem: When
cognitive tasks, Cogn. Sci. 18, 87 (1994). drawing a three-dimensional box in a field vector plot, one
[48] J. Zhang, The nature of external representations in problem can check whether the divergence inside the box vanishes
solving, Cogn. Sci. 21, 179 (1997). or not by determining the net flux through the box. If the
[49] D. J. Griffiths, Introduction to Electrodynamics, 4th ed. net flux is nonzero, the divergence will be nonzero in at
(Addison-Wesley, Reading, MA, 2012), p. 624. least one location inside the box.
[50] K. F. Warnick, R. H. Selfridge, and D. V. Arnold, Teaching [68] If a paddle wheel is placed at a certain location in a field
electromagnetic field theory using differential forms, IEEE vector plot, the curl in that location will be nonzero if and
Trans. Ed. 40, 53 (1997). only if the paddle wheel rotates about its own axis.
[51] T. Dray and C. A. Manogue, Using differentials to bridge [69] D. Hu and N. S. Rebello, Shifting college students’
the vector calculus gap, Coll. Math. J. 34, 283 (2003). epistemological framing using hypothetical debate prob-
[52] B. M. Notaros, Geometrical approach to vector analysis in lems, Phys. Rev. ST Phys. Educ. Res. 10, 010117 (2014).
electromagnetics education, IEEE Trans. Ed. 56, 336 (2013). [70] See Supplemental Material at http://link.aps.org/
[53] H. Huang, J. Wang, C. Chen, and X. Zhang, Teaching supplemental/10.1103/PhysRevPhysEducRes.14.020115
divergence and curl in an electromagnetic field course, Int. for all five sets of worksheets; They will also be available
J. Electr. Eng. Educ. 50, 351 (2013). on http://fys.kuleuven.be/slonwf.
[54] D. F. Styer, The geometrical significance of the Laplacian, [71] D. C. Giancoli, Physics for Scientists and Engineers with
Am. J. Phys. 83, 992 (2015). Modern Physics, 4th ed. (Addison-Wesley, Reading, MA,
[55] B. J. Guzzetti, T. E. Snyder, G. V. Glass, and W. S. Gamas, 2008), p. 1328.
Promoting conceptual change in science: A comparative [72] R. A. Adams and C. Essex, Calculus: A Complete Course,
meta-analysis of instructional interventions from reading 8th ed. (Pearson Education, Upper Saddle River, NJ,
education and science education, Read. Res. Q. 28, 116 2013), p. 1136.
(1993). [73] As the theory and end-of-chapter problems on divergence,
[56] P. S. Shaffer and L. C. McDermott, Research as a guide for curl, and Maxwell’s equations were discussed separately
curriculum development: An example from introductory during various sessions in the original instructional ap-
electricity. Part II: Design of instructional strategies, Am. J. proach, this is a rough estimation.
Phys. 60, 1003 (1992). [74] This is an average value, as students worked at their own
[57] L. C. McDermott, P. S. Shaffer, and M. L. Rosenquist pace. We documented that approximately 1 h was spent on
(University of Washington, Physics Education Group), both the divergence and curl tutorials, 1.5 h on the Gauss’s
Physics by Inquiry: An Introduction to Physics and the law tutorial and 1.5 h in total on the Faraday’s law and
Physical Sciences (John Wiley & Sons, New York, 1996), Maxwell-Ampère’s law tutorials, excluding the pretests and
p. 468. post-tests.
020115-17
[75] S. V. Chasteen, R. E. Pepper, S. J. Pollock, K. K. Perkins, [78] P. S. Shaffer and L. C. McDermott, A research-based
N. S. Rebello, P. V. Engelhardt, and C. Singh, But does it approach to improving student understanding of the vector
last? Sustaining a research-based curriculum in upper- nature of kinematical concepts, Am. J. Phys. 73, 921 (2005).
division electricity & magnetism, AIP Conf. Proc. 1413, [79] M. Kryjevskaia, A. Boudreaux, and D. Heins, Assessing
139 (2012). the flexibility of research-based instructional strategies:
[76] One of the teaching assistants was directly involved in the Implementing tutorials in introductory physics in the
research project. There were one or two teaching assistants lecture environment, Am. J. Phys. 82, 238 (2014).
for every session. [80] S. Goldhaber, S. Pollock, M. Dubson, P. Beale, and K.
[77] The instructor in charge was not directly involved in the Perkins, Transforming upper-division quantum mechanics:
research nor the tutorials, but supported the project and Learning goals and assessment, AIP Conf. Proc. 1179, 145
made sure the necessary time slots were available. (2009).
020115-18

Development Implementation and Assessment of A Gui

Uploaded by

Copyright:

Available Formats

Development Implementation and Assessment of A Gui

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Development Implementation and Assessment of A Gui

Uploaded by

Copyright:

Available Formats

PHYSICAL REVIEW PHYSICS EDUCATION RESEARCH 14, 020115 (2018)

Development, implementation, and assessment of a guided-inquiry

(Received 15 February 2018; published 9 November 2018)

I. INTRODUCTION In introductory physics courses, Maxwell’s equations are

2469-9896=18=14(2)=020115(18) 020115-1 Published by the American Physical Society

correct application of the vector operators and Maxwell’s

5. Difficulties with interpreting Maxwell’s B. A multiple representation approach:

FIG. 2. Excerpt of a worksheet used in the divergence tutorial.

students perceived the tutorials and discuss their informal

A. Semiquantitative assessment of the

FIG. 6. The proportion of KU Leuven students (N ¼ 57) who

FIG. 8. The proportion of KU Leuven students who determined

participated in the tutorials. It seems clear that the pro-

field of a point charge [Fig. 12(c)], and a correct field vector

FIG. 13. Post-test question on Maxwell’s equations.

You might also like