You are on page 1of 8

Student understanding of electric and magnetic fields in materials

Savannah L. Mitchem, Dina Zohrabi Alaee and , and Eleanor C. Sayre

Citation: American Journal of Physics 85, 705 (2017); doi: 10.1119/1.4991376


View online: http://dx.doi.org/10.1119/1.4991376
View Table of Contents: http://aapt.scitation.org/toc/ajp/85/9
Published by the American Association of Physics Teachers
PHYSICS EDUCATION RESEARCH SECTION
The Physics Education Research Section (PERS) publishes articles describing important results from the
field of physics education research. Manuscripts should be submitted using the web-based system that can
be accessed via the American Journal of Physics home page, http://ajp.dickinson.edu, and will be forwarded
to the PERS editor for consideration.

Student understanding of electric and magnetic fields in materials


Savannah L. Mitchema)
Department of Physics, Florida State University, Tallahassee, Florida 32306
Dina Zohrabi Alaeea) and Eleanor C. Sayreb)
Department of Physics, Kansas State University, Manhattan, Kansas 66506
(Received 24 July 2016; accepted 16 June 2017)
We discuss the clusters of resources that emerge when upper-division students write about
electromagnetic fields in linear materials. The data analyzed for this paper come from students
written tests in an upper-division electricity and magnetism course. We examine how these clusters
change with time and context. The evidence shows that students benefit from activating resources
related to the internal structure of the atom when thinking about electric fields and their effect on
materials. We argue that facilitating activation of certain resources by the instructor in the classroom
can affect the plasticity of those resources in the student, making them more solid and easily
activated. We find that the wording of the questions posed to students affects which resources are
activated, and that students often fill in resources to link known phenomena to phenomena described
by the question when lacking detailed mental models. VC 2017 American Association of Physics Teachers.
[http://dx.doi.org/10.1119/1.4991376]

I. INTRODUCTION structure of student reasoning. We use Resource Theory17 to


describe the ways students link pieces of information, or
Learning to think scientifically is extremely important for resources, to form more complex ideas, improve understand-
all students. All science, technology, engineering, and math ing, and solve problems. As researchers gain more insight
(STEM) majors must take physics; thus, research into how into the diversity of student thought processes, patterns in
students learn physics is important for all STEM students. A how students use and combine resources help us better
significant body of research in physics education has been understand student learning. In this paper, we discuss student
devoted to improving the teaching and evaluation1 of elec- responses to conceptual questions about linear materials in
tricity and magnetism (E&M) at the introductory level,24 electric and magnetic fields. We categorize the responses by
but less attention has been geared towards upper division the students resource usage, and we look for changes across
E&M.510 Learning physics at the upper division level is a
contexts and over time.
complex endeavor, and each student is unique in the way
they process information to understand upper level physical
concepts. Prior research into student learning of E&M at the II. THEORY
upper level has focused on student understanding of specific
topics like Maxwells equations.812 Other research uses Previous research on Resource Theory and mental models
E&M as a context in which to study mathematical tools that in Electricity and Magnetism guides the research described
support physical reasoning,13,14 student identity develop- in this paper. A resource, broadly speaking, is a discrete
ment,15 or epistemological framing.16 Of the studies that piece of an idea that a student activates when considering or
focus on student understanding of specific physics topics, solving a problem.17 It is important to note that resources
none focus on how students understand the behavior of linear themselves are not inherently right or wrong (e.g., closer
materials in electric and magnetic fields. This is an important means stronger),18 and that student difficulties arise from
topic to research in science and technology. For instance, misapplying resources (e.g., its hotter in the summer
dielectrics have numerous practical applications in the home because Earth is closer to the sun). Students build resources
and industry, including nanotechnology, electronics, photon- throughout their lives and education,19 and they use these
ics, chemical and mechanical systems, and the emerging resources by activating and linking them to form mental
fields of biology and biochemistry. Thus, we need to help models of complex physical and mathematical situations.
students develop mechanistic reasoning while covering these Unlike phenomenological primitives, or p-prims for short,
topics that are fundamental to many STEM fields. which are the simplest, most atomic pieces of knowledge,18
The development of mechanistic reasoning plays an resources can have internal structure that is accessible to the
important role in student understanding, and evaluating user.19,20 Because resources can be more complex and struc-
understanding in STEM areas often involves unpacking the tured, students can develop resources for sophisticated topics

705 Am. J. Phys. 85 (9), September 2017 http://aapt.org/ajp C 2017 American Association of Physics Teachers
V 705
such as diode construction,21 separating differential equa- electricity, current, and energy. The subjects knew that
tions,22,23 or quantum mechanics.24 a causal agent must come between perceived related events,
Several kinds of resources have been identified. such as a wire being connected and a light bulb lighting up,
Conceptual resources are pieces of knowledge, such as coor- and they filled in resources about something flowing, with-
dinate systems19 or activating agent.18 Procedural resour- out knowing what that thing was.
ces23 are actions such as bringing constants out of a Borges and Gilbert investigated mental models of magne-
derivative25 or summing forces. Epistemological resources26 tism as well. Although their discoveries were of mental mod-
are elements of beliefs about the nature of knowledge, such els of magnets creating magnetic fields, those models may be
as whether results can be figured out or if they need to be closely related to the mental models students have of materi-
looked up. To solve problems, students coordinate all three als subjected to a magnetic field. Again, the most complete
kinds of resources, activating connections between concep- mental models (meaning the ones with the most power in
tual resources and procedural ones to build arguments, as making predictions and inferences about a situation)28 were
mediated by epistemological resources about the problem also the most microscopic models (very detailed). They also
solving they are engaged in. We are specifically interested in found that the participants interviewed with the most prior
conceptual resources, and how instructors can recognize the experience and formal instruction about magnetism were the
resources students bring to the course and how they can only ones who thought of magnetism as a result of micro-
encourage the use of those resources as applied to electricity currents circling around the atoms in a material that line up
and magnetism.4 We explore in this paper which conceptual to produce a field. They concluded that deliberate instruction
resources are activated in clusters when students explain can affect students mental models, noting that when com-
polarization and magnetization. paring students models, the effects of their models are evi-
Another characteristic of resources is their plasticity ver- dent, as shown by the vocabulary and the type of constructs
sus their solidity.19 Solid resources are durable and well they use.28
established, and their internal structure need not be accessed It is important to note that it is possible to have a complete
for use. In other words, they require less justification and can mental model that is applied incorrectly. For example, some
be more easily and readily used and linked with other resour- participants in our study showed evidence of having a com-
ces. Plastic resources are unstable and require more elabora- plete mental model of polarization of light that they applied
tion and justification to be used. They are usually new incorrectly to our question about polarization of materials. It
resources for the student, and their connections to other is also possible to have an incomplete mental model that is
resources have yet to be solidified. More broadly, stable applied correctly. For example, a student could make correct
networks are networks of resources that are cognitively predictions about the results of a physical situation, but may
nearby,21 and tend to activate together.19 In this paper, we be only able to do so for very specific cases.
examine how the plasticity of conceptual resources and the
networks and clusters of resources students use change with
increased conceptual instruction. III. METHODS
Solid networks of resources form students mental models. A. Data collection
Mental models enable individuals to make predictions and
inferences, to understand phenomena and events, to make The data analyzed for this paper come from students writ-
decisions, and to control their execution. They are incomplete, ten test responses spanning two years at Kansas State
despite being structural analogues of the processes taking University. The upper-division E&M course is available
place in the world.27 After interviewing students across a every fall semester, and typically covers the first half of
wide range of ages and backgrounds, Borges and Gilbert28 Griffithss text,29 Maxwells equations in matter and vac-
gained insight into mental models of electricity, and found uum. The class meets four days a week for fifty minutes at a
that the more successful and complete the model was, the time. About one third of the days are tutorial days in which
more microscopic (very detailed) it tended to be. They also students complete guided practice problems in small groups
discuss students use of causal agents in their reasoning. They instead of listening to a lecture.
discovered that many of the more incomplete mental models There were two cohorts of students who took the class in
the subjects had about electric current moving through a wire 2014 and 2015. In 2014, on the final exam (Final 2014),
were characterized by the interchangeability of words such as 19 students were asked both parts (1) and (2) from Fig. 1. In

Fig. 1. Final exam question for 2014 and 2015 cohorts. At mid-semester in 2015, the students were quizzed only on part 2.

706 Am. J. Phys., Vol. 85, No. 9, September 2017 Mitchem, Zohrabi Alaee, and Sayre 706
2015, 18 students were asked only part (2) of this problem When a linear paramagnetic material is placed in a mag-
(electric fields) near the beginning of instruction on this topic netic field, whatever magnetic dipoles that already exist
in the 12th week of the semester (Mid-Semester 2015, we become more aligned with the field, tending to point in the
treated this as a pre-test), and all parts of the problem on same direction as each other. As opposed to polarization,
their final exam (Final 2015). Between the 2015 pretest magnetization of a material does not separate any charges,
and final, students were given targeted instruction on this but merely changes the random direction of spins within the
problem. Thirty-seven students in total completed the class atoms to be more uniform. The magnetization M of a mate-
in 2014 and 2015, and 55 assessments (written test rial is given by M vm H, where H is the magnetic field and
responses) were collected, both during instruction and at the vm is a constant dependent on the material. The linear rela-
final exam. We analyzed all usable answers to find patterns tionship between M and H breaks down after H is strong
and trends in the students use of resources, and we also enough to align all of the magnetic moments of the atoms
searched for patterns among the 2015 students responses parallel to it. After that threshold, H can increase but M can-
during the semester and after they completed the class. not. No breakdown of the material occurs.

B. Data analysis V. RESULTS


We reviewed all 55 student exams, seeking patterns in stu- A. E-field
dent understanding. Three students did not answer the ques-
tion and two students wrote about the unrelated phenomenon As part of our investigation, we looked at how students
of polarization of light. We discarded these students exams considered what happens to a dielectric material in an
from further analysis. In addition, one student did not take extremely strong electric field. We examined their responses
Final 2015 and one student was absent for Mid-Semester for evidence of resources being activated. When observing
2015, leaving 50 usable responses in all, 33 for 2015 and 17 the students answers to the polarization question, we found
for 2014. that the resources students activated generally fell into two
To find patterns, we considered each responses resources, groups. The first group, which we call sub-atomic, consists
drawing resource graphs20 to generalize them across of students who used resources related to the internal struc-
responses. We use resource graphs to link resources acti- ture of the atoms within the material. This includes mention
vated in particular contexts. As themes emerged in the data, of electron clouds along with nuclei, stretching of the nega-
we sought a categorization scheme that would both capture tive and positive charges within the atom, drawings of
the rich variety of student responses and make meaningful dipoles with positive and negative sides, and internal forces
categorical differences among them. We iteratively devel- of the atom versus the external force of the electric field;
oped an emergent coding scheme linking specific student 23 of 50 responses are in this category.
language to specific resource use. The second group, super-atomic, consists of students who,
We classified each response as sub-atomic or super-atomic instead of thinking about the internal structure of the atom,
(explained in Sec. IV). Then we noted which resources each talked only about the material in general. These students
student used by the criteria in our codebook. If the students often considered the problem in terms of the model of posi-
resources did not fall into any of the categories we defined, tive and negative charges moving throughout the material
we marked them as other. and collecting on one side, as opposed to thinking about the
When our coding scheme was stable, we invited an addi- atoms themselves becoming dipoles aligned with, and pulled
tional researcher to separately code student responses. He apart by, the field. Included in this category are drawings of
categorized 20 randomly selected answers, using our catego- material with pluses and minuses lined up within it (that do
rization scheme for evaluation during inter-rater reliability not form polarized atoms), mention of electrons or charges
testing (IRR). Before the discussion, 82% of his identified without reference to an atom, and mention of the word
resources and ours overlapped. After discussion, which dipole without description or depiction of its structure;
included clarifying and improving parts of the codebook and 27 of 50 responses are in this category.
discussing each response categorized differently, we When comparing the answers of the sub-atomic and
repeated this procedure with 20 different responses and super-atomic groups represented in Table I, we found that if
found 100% agreement. a student considers the internal structure of the atom during
polarization, they are very likely to arrive at the correct
IV. THE PROBLEM AND ITS SOLUTION answer, dielectric breakdown. This is true for all three testing
situations we evaluated. Furthermore, on the two tests where
The following is a complete and correct response to the students had not been given the question before (Mid-
exam question students were asked. When a dielectric mate- Semester 2015 and Final 2014), the students that took the
rial is placed in an electric field, its atoms become polarized. super-atomic route were not likely to get to breakdown over
This means that the negative electron cloud around the posi- other answers. This is especially true for Mid-Semester
tive nucleus is slightly displaced from the center. This dis- 2015, when the students had not yet been taught the material.
placement of charges within the atoms forms electric dipoles There were other variations of resources used within the
with separation linearly proportional to the strength of the large groups of sub-atomic and super-atomic resources.
electric field, as given by the equation P 0 ve E, where P is Students who used sub-atomic resources and reached break-
net polarization, E is the electric field, ve is a property of the down tended to justify their answers one of two ways. In the
material (susceptibility), and 0 is constant. When the electric first approach, students described the internal atomic forces
fields strength exceeds a certain threshold value, electrons holding the electrons and nuclei together, and how an elec-
become separated from their nuclei, converting the dielectric tric field above a certain threshold would provide a large
into a conductor. This is called dielectric breakdown. enough force to overcome the internal force and break

707 Am. J. Phys., Vol. 85, No. 9, September 2017 Mitchem, Zohrabi Alaee, and Sayre 707
Table I. Resource groups activated by students versus their answers while atoms, so we classified her as using sub-atomic resources,
writing about material in fields. seeing as she conceptualized the situation on a sub-atomic
level. Kate continued, writing As the field increases, the
Sub-atomic Super-atomic electron clouds get pulled farther and farther away from the
Final 2014 nucleus, until they can go no farther without leaving the
Break down 5 3 nucleus entirely. We classified this resource as maximum
Saturation 1 6 displacement because she mentioned the distance between
Others 0 2 the electron clouds and the nuclei reaching a breaking point.
For part two of the question, she wrote, The electrons are
Mid-Semester 2015
ripped away from the nuclei, causing a brief flow of charge
Break down 7 1
(dielectric breakdown).
Saturation 0 7
Students who did not think about polarization at the sub-
Others 0 1
atomic level tended to justify their answers using a descrip-
Final 2015 tion of the dipoles or charges aligning with the field until
Break down 9 4 none of the dipoles or charges could be more aligned. This
Saturation 1 2 resource, alignment, was usually paired with the final answer
Others 0 1 of saturation. These students wrote that nothing significant
happened to the material besides alignment, failing to men-
tion breakdown. We called this cluster of resources Cluster
electrons away from their atoms. We called this resource C which included super-atomic resources, alignment, and
balancing forces, and Cluster A is the name we gave the saturation. Several students responses (fifteen responses in
cluster of resources including sub-atomic resources, balanc- all) that used this cluster made no mention of separation of
ing forces, and breakdown. In the second method, students charges during polarization and only mentioned alignment of
described how maximum polarization pulls the electron the dipoles, treating polarization and magnetization as the
cloud so far away from the nucleus that electrons break from same phenomenon in their answers. Many students used a
their nuclei. We called this resource maximum displacement, model of charges aligning in some way across the material
and Cluster B is the name we gave the cluster of resources while separating as well.
including sub-atomic resources, maximum displacement, and An example of Cluster C is due to Alex, in Final 2014.
breakdown. They are essentially the same phenomenon, but He wrote, Polarization occurs when an E-field pushes all
balancing forces is a more rigorous description. Figure 2 charges one way and the  charges move the opposite direc-
depicts the common clusters of resources used by students. tion. This causes an E-field from the difference between sep-
An example of a student response that includes sub-atomic arated and  charges. There is no mention of atoms here,
resources and the balancing forces resource (Cluster A) is and the term charges, especially when he discusses posi-
that of Kate,30 in Mid-Semester 2015. Polarizing atoms tive charges, indicates his use of the model of charges mov-
means slightly offsetting the negative electron cloud from ing throughout an insulator and causing net effects rather
the positive nucleus. It is clear that Kate uses sub-atomic than electrons moving within an atom. He went on to write,
resources to frame the question. She goes on to say, If the Eventually all the will be as far to one side as possible
field doing this is strong enough, it will separate the two and all of  will be as far as possible in the opposite direc-
entirely, since it will overcome the Coulombic force attract- tion. The charges are as far away as possible with all the
ing the electrons to the nucleus. The electrons will gather on charges segregated. At this point, no more charges can move
one side of the material. The use of the words overcome so E will increase, but P cant. Here, he activates the resour-
and force attracting the electrons to the nucleus shows that ces alignment and saturation. In summary, he uses super-
balancing forces activated along with breakdown. atomic resources, alignment, and saturation.
In her answer for Final 2015, Kate used Cluster B instead. The rest of the students answers indicated the activation
She wrote, The external electric field pulls on the electron of other resource clusters that we grouped into the miscella-
clouds in the individual atoms/molecules, creating a polar- neous Cluster D. This includes any answers that were neither
ization. She mentioned electron clouds in relation to their breakdown nor saturation. It also includes students who used

Fig. 2. Resource graph for materials in electric fields. The diagram on the left represents Clusters A and B, and the diagram on the right represents Cluster C.

708 Am. J. Phys., Vol. 85, No. 9, September 2017 Mitchem, Zohrabi Alaee, and Sayre 708
sub-atomic resources but reached an answer of saturation, or breakdown as a resource on the final exam without accessing
students who used super-atomic resources but reached an its internal structure or having to justify it further by using
answer of breakdown. And finally, it includes the few stu- sub-atomic resources.
dents who used none of these or misunderstood what the Four other students used Cluster C or D, getting saturation
question asked of them. on the pretest, and then switched to Cluster A or B on the
The grouping of these resources into clusters aided us in final, indicating that they understood the lecture and used the
finding patterns in student responses over varied context and resources given to them by instruction to correctly answer
amount of instruction. As stated before, for all three testing the final. (Note that when we use the word correctly here,
situations, using sub-atomic resources almost always coin- we are referring only to the final answer and are not com-
cided with activating the resource of breakdown. menting on the correctness of the reasoning the student
Conversely, activation of super-atomic resources typically used.) The remaining students either used the same cluster
coincided with activation of the saturation resource instead for both exams, or used a different cluster resulting in the
of breakdown. Interestingly though, on Final 2015, there was same answer for both.
not as much of a difference between the number of sub-
atomic and super-atomic students who arrived at breakdown. B. B-field
In fact, 13 out of 17 students reached breakdown on Final
2015, whereas only 8 out of 16 reached breakdown on Mid- The second part of our research was to analyze data from
Semester 2015. Below are details about why this may have the portion of the question on the two final exams about
occurred. materials in extremely high magnetic fields. This part of the
Table II summarizes the number of students who used question was not included in Mid-Semester 2015. Unlike the
each cluster on Mid-Semester 2015 and Final 2015. Four stu- polarization question, none of the students who answered the
dents switched from using Cluster A, B, or C in the pretest to magnetization question had seen the problem before their
using Cluster D (and still getting breakdown) on the final. final exam. However, they had all been taught concepts
Take the response of Zachary for Final 2015 for which he about magnetic fields and magnetization through practice
wrote, The material is bounded. Above and below the faces problems and lectures prior to the final. The purpose of this
of the material are regions of different r and ve. These question was to get students to compare and contrast what
regions do not have the same property as the linear material, happens to linear materials when they are highly polarized
and so the max P occurs at E threshold, when all the charges versus highly magnetized.
are on opposing faces. Along with his written answer, The vast majority of student responses included only
Zachary also drew a diagram of a rectangular prism in an super-atomic resources to describe and explain magnetiza-
electric field. He went on to say, To remain in equilibrium, tion and its effect on the material. These resources consisted
the dipole Ei Eext . This is achieved by highly polarizing mainly of alignment of atoms in some sense. Only two stu-
the material, so all the positive charges are on top face, and dents did not reach the correct answer of saturation, and
charges on bottom face. This occurs at E threshold. If interestingly, those students used sub-atomic resources in
Eext > Ethreshold , the charges cant balance out, and the their responses. Those students mentioned the internal struc-
charges begin to strip away from the material and end the ture of the atom during magnetization, and they concluded
linear property. with breakdown of the material.
Zachary was able to reach breakdown without activating Breakdown and saturation are the correct answers for E-
sub-atomic resources, unlike in his answer for the question field and B-field, respectively. So by switching the context,
during the pretest for which he used Cluster A. He discusses the same resource can be applied correctly or incorrectly.
only super-atomic qualities and phenomena in his response That is, resources by themselves are neither right nor wrong.
after instruction about materials in electric fields and the
internal process of polarization. This could be due to the fact C. Comparison between class activities and student
that the instructor facilitated activation of those sub-atomic responses
resources in a way that allowed them to go from being plas- We recorded the class meetings in 2015 about magnetiza-
tic to solid in Zacharys mind. The instructor reviewed the tion as part of a broader project on student reasoning during
exact question after the pretest, and used sub-atomic descrip- class. Upon reviewing the lectures, we discovered how dif-
tions in a Mid-Semester lecture including the fact that each ferent the magnetization lectures were from the polarization
atom is made up of a little bit of negative and positive, and lectures discussed in Sec. VI. The instructor used almost no
the positive center is surrounded by the negative outside. She conceptual or sub-atomic description of magnetization, and
also describes breakdown as when electrons get ripped instead focused on bolstering the students mathematical
away from atoms, and contrasts separation and alignment. understanding of the way material behaves in high magnetic
This lecturing could have increased Zacharys ability to use fields. She mentioned ways different materials hold magneti-
zation and for how long (such as the ferromagnets ability to
Table II. Comparing the cluster of resources during Mid-Semester and Final become a permanent magnet), and several students included
2015. that information in their responses. However, the only con-
ceptual explanation on a sub-atomic level was that atoms can
Mid-semester test Final test
align with or against the field based on the thing that is
Cluster A 4 7 doing the spinning to make the current in the dipole. She
Cluster B 3 2 says that this thing is all kinds of quantum-y things and
Cluster C 7 2 has to do with how electrons are paired in atoms.
Cluster D 2 6 This is a contrast from the lecture on polarization which
focused on a sub-atomic description of breakdown. The

709 Am. J. Phys., Vol. 85, No. 9, September 2017 Mitchem, Zohrabi Alaee, and Sayre 709
student responses reflect this difference. Although only two VI. SUMMARY AND CONCLUSION
students said that something happens to the material besides
maximum alignment (and therefore saturation), the things We discussed the clusters of resources that were activated
that students said were aligning varied widely. Over twenty together when upper division E&M students expounded on
different words were used for this thing, and most of the the behavior of materials in fields. Some examples show that
pictures drawn were arrows lining up with each other, indi- those clusters can change with time and context, specifically
cating a super-atomic model of magnetization. The words with increased instruction of the concepts in question. We
have offered insight to which mental models are most helpful
used included correct terminology such as magnetic
for fostering complete conceptual understanding of polariza-
moment, dipole, and electron spin, and incorrect or
tion and magnetization.
incomplete terminology such as electrons, atoms, and
The data indicate that use of resources related to the inter-
little pieces. All answers including alignment and satura-
nal structure of the atom during polarization increased likeli-
tion were given credit. This lack of a complete group of
hood of activating breakdown. Students who did not activate
resources to make up a mental model of magnetization could
these sub-atomic resources tended to activate saturation
be due to the instruction the students were given. The stu-
instead. This predictive relationship was not as strong on
dents lacked a link between the idea that material is made of
Final 2015, where we argue that because the students had
atoms and that something about those atoms aligns during
seen the question and received instruction, the resource clus-
magnetization. A similar result was found in Borges and ters they used changed from Mid-Semester 2015. The break-
Gilberts research in which they describe student use of a down resource became more solid and required less
causal agent (in this case, the thing that is aligning) to justification for use, making the students answers less sub-
fill in logical gaps left by the question.28 Furthermore, previ- atomic but not less correct.
ous work on how students understand electric and magnetic We also assert that wording of a question has a role in acti-
fields together suggest that students often confuse these fields vating certain resources in the student and can induce students
or their effects with each other.3133 For example, Scaife and to think mechanistically.34 By reading the question, students
Heckler found that interference between electric and mag- knew that some aspect of a material becomes saturated when
netic concepts can make students confused. They also stated it is in an extreme magnetic field, and they filled in objects to
that students responses depend on whether electric or mag- link their knowledge that material is made up of things, and
netic force questions are posed first, and this effect depends that these things align in the presence of the field.
on whether electric or magnetic force was most recently This research yields multiple implications for instruction
taught.33 of E&M. After comparing the answers in Part A and B of the
The wording of the question may have also played a role final question and watching videos of the lectures corre-
in the high number of saturation responses. Although both sponding to them, we hope to encourage instructors to put
questions about the electric and magnetic fields are worded more emphasis on sub-atomic conceptual models when
the same, the data should not be analyzed independently teaching polarization. These implications for instruction can
from the questions asked. This is because the nature of the also apply to lower and higher level courses than upper divi-
response expected of the student for each question was dif- sion E&M, and conceptual understanding of physical phe-
ferent, and the question itself can activate certain resources. nomena on a sub-atomic scale could be a helpful supplement
The phrase cannot become more magnetized in the ques- to math based curricula as well.
tion is essentially the definition of saturation, and this may More broadly, when considering how resources are acti-
have activated resources related to saturation. The students vated and whether these resources are more solid or more
could have then activated resources to link their knowledge plastic, the general goal of instruction is to help resources
of magnetization with the idea of saturation. In conclusion, become more solid. The reasoning attached to solid resour-
wording of this part of the problem may aid students to find ces is more complete and more well-connected than that of
the correct answer, saturation, and help them to obtain full plastic resources. However, there are two caveats. First, as
credit. resources become more solid, students do not need to look
Similarly, the polarization question uses the phrase inside anymore. This is not necessarily a negative effect, but
cannot become more polarized. This phrase may have acti- it could be negative in a circumstance where accessing inter-
vated saturation resources in some of the students as well. nal structure would benefit the student. Second, occasionally
However, for the polarization part, the instructor expected resources are misapplied and become solid in their incorrect
students to take their mental models of materials in electric usage and this may be where deep-seated misconceptions
fields and extrapolate them to explain a special case, break- come from.19
down. They are expected to reason further than the default of In conclusion, the goals of this study were to identify the
saturation set by the question. In other words, although the clusters of resources that successful students activated while
wording of the question was identical for both magnetization answering an upper-level conceptual problem in E&M, and
and polarization, it favored saturation as opposed to break- to highlight the role of the instructor to facilitate activation
down. Borges and Gilberts research also supports the claim of those resources. These results should improve our under-
that student responses and explanations should not be ana- standing of how students reason about fields in materials and
lyzed separately from the question they were asked.28 could yield insight into instructional strategies to improve
Though the wording is analogous for both questions, in the learning and teaching of physics at upper division level.
the magnetization case it cues students to think appropri- We know that active learning environments paired with
ately, but in the polarization case it cues them to think inap- conceptual curricula improve student understanding of phys-
propriately. This interaction effect between question ics, but improved instructional methods are still needed to
phrasing and students resource use is important to note support students reasoning in these areas, especially for
when designing questions and analyzing students responses. upper division courses. Finally, our work highlights the need

710 Am. J. Phys., Vol. 85, No. 9, September 2017 Mitchem, Zohrabi Alaee, and Sayre 710
12
for future research on students use of resources to solve Charles Baily, Laurens Bollen, Andrew Pattie, Paul Van Kampen, Mieke
problems in these upper division physics courses. We believe De Cock, North Haugh, St Andrews, Fife K Y Scotland, P. van Kampen,
and M. De Cock, Student thinking about the divergence and curl in math-
that thinking about polarization inside the atom seems to
ematics and physics contexts, 2015 PERC Proceedings (2015), pp. 36.
increase understanding and can give students better intuition 13
Bethany R. Wilcox and Steven J. Pollock, Coupled multiple-response
about special cases such as dielectric breakdown. Materials versus free-response conceptual assessment: An example from upper-
with dielectric properties are crucial to modern science, tech- division physics, Phys. Rev. Spec. Top. Phys. Educ. Res. 10, 020124
nology, and manufacturing, so improving student under- (2014); e-print arXiv:arXiv:1407.6310v2.
14
standing of them is important. Rachel E. Pepper, Stephanie V. Chasteen, Steven J. Pollock, and
Katherine K. Perkins, Observations of student difficulties with mathemat-
ics in upper-division electricity and magnetism, Phys. Rev. Spec. Top.
ACKNOWLEDGMENTS Phys. Educ. Res. 8, 010111 (pp. 115) (2012).
15
Eleanor C. Sayre and Paul W. Irving, Brief, embedded, spontaneous
The authors would like to thank the Kansas State metacognitive talk indicates thinking like a physicist, Phys. Rev. Spec.
University Physics Education Research group (KSUPER) for Top. Phys. Educ. Res. 020121, 117 (2015).
assistance on this paper and for helpful feedback. In 16
Hai D. Nguyen, Deepi N. Chari, Dean A. Zollman, and Eleanor C. Sayre,
particular, the authors would like to thank Dean Zollman for Dynamics of students epistemological framing in group problem sol-
reviewing the manuscript and providing feedback. The ving, Eur. J. Phys. 37, 19 (2016).
17
authors would also like to thank Hai Nguyen for his David Hammer, Student resources for learning introductory physics,
Am. J. Phys. 68, S52S59 (2000).
assistance with the IRR. Portions of this research were 18
A. A. DiSessa, Between brain and behavior: Response to marton, Cogn.
supported by the KSU Department of Physics, NSF grants Instruct. 10, 261280 (1993).
DUE-1430967 and PHYS-1461251, and the Air Force Office 19
Eleanor C. Sayre and Michael C. Wittmann, Plasticity of intermediate
of Scientific Research. mechanics students coordinate system choice, Phys. Rev. Spec. Top.
Phys. Educ. Res. 4, 20105 (pp. 114) (2008).
20
a)
S. L. Mitchem and D. Z. Alaee contributed equally to the work of this Michael C. Wittmann, Using resource graphs to represent conceptual
manuscript and are listed alphabetically. change, Phys. Rev. Spec. Top. Phys. Educ. Res. 2, 20105 (pp. 117)
b)
Electronic mail: esayre@ksu.edu (2006).
21
1
David P. Maloney, Thomas L. OKuma, Curtis J. Hieggelke, and Alan Eleanor C. Sayre, Michael C. Wittmann, and John R. Thompson,
Van Heuvelen, Surveying students conceptual knowledge of electricity Resource selection in nearly-novel situations, in Physics Education
and magnetism, Am. J. Phys. 69, S12S23 (2001). Research Conference Proceedings 2003, Physics Education Research
2
Chandralekha Singh, Student understanding of Symmetry and Gausss Conferences, edited by Jeffrey Marx, Scott Franklin, and Karen
law, in 2004 PER Conference Proceedings - AIP Conference Cummings (Springer, New York, LLC, Secaucus, NJ, 2003), pp. 101104.
22
Proceedings 790, Physics Education Research Conferences, edited by Michael C. Wittmann and Katrina E. Black, Mathematical actions as pro-
Jeffrey Marx, Paula Heron, and Scott Franklin, American Association of cedural resources: An example from the separation of variables, Phys.
Physics Teachers (American Institute of Physics, Melville, NY, Rev. Spec. Top. Phys. Res. 11, 020114 (2015).
23
Sacramento, CA, 2005), pp. 6568. Katrina E. Black and Michael C. Wittmann, Procedural resource creation
3
Lillian C. Mcdermott and Edward F. Redish, Resource letter PER-1: in intermediate mechanics, in Physics Education Research Conference
Physics education research, Am. J. Phys. 67, 755767 (1999). 2009, Physics Education Research Conferences, Vol. 1179, edited by Mel
4 Sabella, Chandralekha Singh, and Charles Henderson, American
Ruth Chabay and Bruce Sherwood, Restructuring the introductory elec-
tricity and magnetism course, Am. J. Phys. 74, 329336 (2006). Association of Physics Teachers (American Institute of Physics, Melville,
5 NY, Ann Arbor, MI, 2009), pp. 97100.
Corinne A. Manogue, Kerry Browne, Tevian Dray, and Barbara Edwards,
24
Why is Amperes law so hard?, Am. J. Phys. 74, 344350 (2006). Elizabeth Gire and Corinne A. Manogue, Resources students use to
6 understand quantum mechanical operators, in 2008 Physics Education
Joshua Bilak and Chandralekha Singh, Improving students conceptual
understanding of conductors and insulators, in 2007 PER Conference Research Conference, AIP Conference Proceedings, edited by Charles
Proceedings - AIP Conference Proceedings 951, Physics Education Henderson, Mel Sabella, and Leon Hsu, American Association of Physics
Research Conferences, edited by Leon Hsu, Charles Henderson, and Laura Teachers (American Institute of Physics, Melville, NY, Edmonton,
McCullough, American Association of Physics Teachers (American Canada, 2008), pp. 115118.
25
Institute of Physics, Melville, NY, Greensboro, NC, 2007), pp. 4952. Bahar Modir and Eleanor C. Sayre, Pulling out as a procedural resource
7 when solving partial differential equations, in Proceedings of the 2016
Thomas J. Bing and Edward F. Redish, The cognitive blending of mathe-
matics and physics knowledge, in 2006 PER Conference Proceedings - Physics Education Research Conference (2016).
26
AIP Conference Proceedings 883, Physics Education Research David Hammer and Andrew Elby, Tapping epistemological resources for
Conferences, edited by Laura McCullough, Leon Hsu, and Paula Heron, learning physics, J. Learn. Sci. 12, 5390 (2003).
27
American Association of Physics Teachers (American Institute of Physics, A. Tarciso Borges and John K. Gilbert, Mental models of electricity, Int.
Melville, NY, Syracuse, NY, 2007), pp. 2629. J. Sci. Educ. 21, 95117 (1999).
8 28
Rachel E. Pepper, Stephanie V. Chasteen, Steven J. Pollock, and A. Tarciso Borges, Colegio Tecnico, and John K. Gilbert, Models of
Katherine K. Perkins, Our best juniors still struggle with Gausss Law: magnetism, Int. J. Sci. Educ. 20, 361378 (1998).
29
Characterizing their difficulties, in 2010 PER Conference Proceedings - David J. Griffiths, Introduction to Electrodynamics, Vol. 93 (Prentice
AIP Conference Proceedings 1289, Physics Education Research Hall, NJ, 1999), p. 95.
30
Conferences, edited by Chandralekha Singh, N. Sanjay Rebello, and Mel All student names are pseudonyms.
31
Sabella, American Association of Physics Teachers (American Institute of Andrew F. Heckler and Eleanor C. Sayre, What happens between pre-
Physics, Melville, NY, Portland, OR, 2010), pp. 245248. and post-tests: Multiple measurements of student understanding during an
9
Adrienne Traxler, Katrina E. Black, and John R. Thompson, Students introductory physics course, Am. J. Phys. 78, 768777 (2010).
32
use of symmetry with Gausss law, in 2006 PER Conference Proceedings Eleanor C. Sayre and Andrew F. Heckler, Peaks and decays of student
- AIP Conference Proceedings 883, Physics Education Research knowledge in an introductory E&M course, Phys. Rev. ST Phys. Educ.
Conferences, edited by Laura McCullough, Leon Hsu, and Paula Heron, Res. 5, 13101 (pp. 15) (2009).
33
American Association of Physics Teachers (American Institute of Physics, Thomas Scaife and Andrew Heckler, Interference between electric and
Melville, NY, Syracuse, NY, 2007), pp. 173176. magnetic concepts in introductory physics, Phys. Rev. Spec. Top. Phys.
10
Colin S. Wallace and Janelle M. Bailey, Do concept inventories actually Educ. Res. 7, 111 (2011).
34
measure anything?, Astron. Educ. Rev. 9, 010116 (2010). Rosemary S. Russ, Rachel E. Scherr, David Hammer, and Jamie Mikeska,
11
Charles Baily, Michael Dubson, and Steven J. Pollock, Research-based Recognizing mechanistic reasoning in student scientific inquiry: A frame-
course materials and assessments for upper-division electrodynamics work for discourse analysis developed from philosophy of science, Sci.
(E&M II), AIP Conf. Proc. 1513, 5457 (2013). Educ. 92, 499525 (2008).

711 Am. J. Phys., Vol. 85, No. 9, September 2017 Mitchem, Zohrabi Alaee, and Sayre 711