Content uploaded by Paul W. Irving
Author content
All content in this area was uploaded by Paul W. Irving on Mar 06, 2018
Content may be subject to copyright.
Examining Thematic Variation in a Phenomenographical Study
on Computational Physics
Nathaniel T. Hawkins,1Michael J. Obsniuk,1, 2 Paul W. Irving,1and Marcos D. Caballero1, 3, 4
1Department of Physics and Astronomy, Michigan State University, East Lansing, MI, USA
2Department of Physics, Kettering University, Flint, MI, USA
3CREATE for STEM Institute, Michigan State University, East Lansing, MI, USA
4Department of Physics and Center for Computing in Science Education, University of Oslo, Oslo, Norway
Projects and Practices in Physics (P3) is a transformed, first-year introductory mechanics course offered
at Michigan State University. The focus of the course is concept-based group learning implemented through
solving analytic problems and computational modeling problems using the VPython programming environment.
Interviews with students from P3were conducted to explore the variation of students’ perceptions of the utility
of solving computational physics problems in the classroom setting. A phenomenographic method is being used
to develop categories of student experience with computational physics problems based on themes emerging
across the different students’ interviews. This paper will focus on exploring the variation within the theme
of Computation Helps to Learn Physics that arose from our preliminary analysis of the data from a larger
phenomenographic study. When examined on an individual basis, this theme provides important insights into
students’ perception of the use of computation, such as the way that students can engage with computation as a
learning tool in a Physics classroom.
I. INTRODUCTION
Computation is an essential part of modern physics prac-
tice, as much of the physical world is now understood us-
ing computational modeling and analysis techniques. Help-
ing students of physics learn computational methods, tools,
and ways of thinking is likely to become a large part of fu-
ture educational efforts [1]. However, the undergraduate en-
vironments in which students use computing are understudied
[2, 3]. This expanding field of research is open to many unan-
swered questions such as: "How do students learn computa-
tional tools and methods?", "How do students learn physics
from engagement with computation?", and "How effective
are students at appropriating and using computational prac-
tices in physics?"
The study presented here focuses on the development of
the theme Computation Helps to Learn Physics, which is a
description of the ways in which students’ perceive the use of
computation in their learning of physics. Students’ percep-
tions of the tools and methods of instruction (e.g., computa-
tional activities) are important, as these perceptions can deter-
mine how a student chooses to engage in these activities and
can thus affect how and what they learn [4]. This is particu-
larly true when considering how negative views of computing
can affect student learning [5]. In this paper, we present the
first steps of a larger phenomenographic study that is aimed at
developing categories of experience around students’ percep-
tions of the utility of computing in their learning of physics.
We present an emergent theme from the initial stages of our
analysis that we feel is robust enough to provide important
insights into how students are perceiving the interaction be-
tween computation and physics in our context.
II. RESEARCH CONTEXT
The students in Projects and Practices in Physics (P3) are
primarily engineering students, with some physical sciences,
computer science, and pre-professional majors as well. The
structure of a P3semester is that students typically solve an-
alytic problems in a group setting during one class period and
computational problems in the subsequent class period. More
details about P3are available in Irving et al. [6]. For the con-
text of this study, we will be focusing on a suite of 3 problems
students engage with over the first 3 weeks of the semester.
The three computational problems in this suite are as follows:
(1) a boat moving across a running river (vector addition and
relative velocities), (2) a hovercraft accelerating off of a cliff
and trying to land safely in the water below (2D kinematics),
and (3) establishing the geosynchronous orbit of a satellite
(Newtonian gravity). The students work on what we call min-
imally working programs, which are scripts that will run and
produce a visual output in the VPython environment, but are
missing the underlying physics equations needed to make a
physically correct model. In this way, we can have the stu-
dents focus on integrating physics into computation rather
than learning a programming language. Twenty-one students
were interviewed over the course of two semesters for this
study. One-on-one interviews were conducted following the
completion of all three computational problems. Interviews
were conducted by one of the course instructors, one faculty
member, one graduate student teaching assistant, and one un-
dergraduate researcher.
III. RESEARCH DESIGN AND ANALYSIS
The larger scope of this study utilizes the framework of
phenomenography, which can be summarized as the devel-
opment of categories of description surrounding individuals’
experiences with a certain phenomenon as a whole that is
then used in facilitating the grasp of concrete cases of hu-
man understanding [7]. In utilizing this framework, we aim
to develop categories of student perceptions of the utility of
computation in an introductory mechanics course. The re-
edited by Ding, Traxler, and Cao; Peer-reviewed, doi:10.1119/perc.2017.pr.037
Published by the American Association of Physics Teachers under a Creative Commons Attribution 3.0 license.
Further distribution must maintain attribution to the article’s authors, title, proceedings citation, and DOI.
2017 PERC Proceedings,
168
2
Label Variation of Theme N
A Computation Helps Me to Build a Conceptual Understanding of Physics 7
B Engaging in the Practices of Computation Helps Me to Learn Physics 14
C Computation Helps Me to Learn Physics When Interacting with Problems in Set Contexts 5
D Computation Doesn’t Help Me to Learn Physics, But Could Be Helpful to Others 3
E Computation Does Not Help to Learn Physics in General 4
TABLE I. Table summarizing the thematic variation present with the theme Computation Helps to Learn Physics.Nis a count of how many
students (out of 21 total) exhibit this variation at some point during their interview. Note, some students exhibit multiple variations. Students
may exhibit multiple varitaions throughout the course of the interview depending on the context in which they are describing their perceptions
of computation.
sults and analysis presented in this paper are the first stages
of the phenomenographic analysis process which is akin to
qualitative research grounded in thematic analysis [8]. In
semi-structured interviews students reflected on whether the
computational activities they engaged in were helpful in their
learning of physics. Analysis then focuses on reoccurring
themes that emerge from the dataset with the expectation
for phenomenography that several themes will emerge that
are related to each other. Upon analysis of the transcribed
dataset as a whole, there were many instances where sim-
ilarities in the ways that students discussed computation as
it relates to their learning of physics appeared. These were
then reviewed to differentiate the modes of variation across
this theme, which is discussed below. We acknowledge the
analysis in this paper is not phenomenographic, as that would
involve an investigation of their complete perceptions, but we
believe this analysis at this stage can still offer valuable in-
sights that can guide those who intend to integrate computa-
tion into their pedagogy. The larger study that this paper is a
subset of will entail a full phenomenographic analysis.
IV. FINDINGS
The theme discussed here is Computation Helps to Learn
Physics. While the analysis of the dataset is still in a pre-
liminary stage, exploring the variation within this theme is
crucial as a standalone discussion because understanding stu-
dents’ perceptions of learning is important to the discussion
of improving pedagogy [7, 9].
Table I shows a summary of the variations discussed in this
section as well as a count of the number of students who dis-
play each variation in their interview data. It is important
to note that students can display more than one variation of
a theme during the course of their interview. As a result of
this, variations within the theme are not mutually exclusive.
In a phenomenographical study, we expect there to be mul-
tiple themes present across the dataset, each with its associ-
ated variation. Here, we are only looking at the presence of
one theme, so individual instances of variations are counted
as such, instead of being part of a larger result. The varia-
tions of the theme Computation Helps to Learn Physics are
outlined below.
A. Computation Helps Me to Build a Conceptual
Understanding of Physics
This variation can be identified in the student data in in-
stances where the interviewees discussed experiences with
the computational activities that helped them make a concep-
tual connection with physics. These instances are typically
initiated by a student talking about their struggles with a par-
ticular physics concept going into a computational problem,
but they then transition to a discussion about how the compu-
tational exercise helped them to understand the concept. One
example that demonstrates this variation comes from Macku:
Macku: They were trying to explain that centrifugal force
is not really an actual force. We just use it as a concept to
explain things that are bigger. That was a big part, because
we kept trying to say, ‘oh, what kind of force are you going
to need on it?’ We were like, ‘we only really need that force
towards earth, we don’t need the centrifugal force.’ The cen-
trifugal force wasn’t in the code. I think it was not adding it
to the code, but it was more understanding that we don’t need
it.
Macku mentions a specific example during her discussion
of the geosynchronous orbit problem. The students use cen-
trifugal force and equate that to the Newtonian gravitation
equation to solve for the needed velocity to make a circular
orbit. In coding the net force into the program, she noticed
that it was not necessary to include the centrifugal force in
her net force equation. Thus, for her it was a moment of re-
alization that centrifugal force was not a physical force, but
rather a method for performing certain approximations and
calculations. Here, we can see a conceptual connection being
made through the implementation of computation. The dis-
tinct characteristic of this theme is a direct mention to how
engaging in computation helped the student to gain an under-
standing of a physics concept rather than a general statement
about computation being helpful.
B. Engaging in the Practice of Computation Helps Me to
Learn Physics
The next variation is where students refer to aspects of the
computational program that can be summarized as computa-
tional practices. Students’ descriptions often focused around
practices nested in the overall practice of computational mod-
eling like analyzing a visualization produced by VPython, or
169
3
constructing the written lines of code that go into building the
simulation. Students also perceive the act of iterating through
the computational model itself and making decisions about
variables as manifestations of the practice of “thinking like
a Physicist.” These are common aspects to programming in
this setting [10, 11], and students identify how participating
in these practices helps them to learn physics. This variation
is demonstrated by Alan and Dota:
Alan: It allows me to model, model quickly, change quickly,
and see the results of the changes to the model quickly. Be-
cause in these models, I can change specific variables and
when I see the results, it allows me to draw some sort of vi-
sual picture of how I can manipulate real phenomenon with
these variable manipulations.
Dota: Then we had to use friction to find acceleration,
which we needed the old velocity for to find out how it
changes, and so on. In the code, we would see equations
call variables or reference values to do this same task. You’re
actually thinking like a Physicist would and applying what
you’re learning like this, which in turn helps me to learn.
The two examples show two different practices referenced
by the students. Alan refers to the creation of the visual model
and how changing specific variables in his code allows him
to see the changes in his model, or as he says, allowing him
to “manipulate real phenomenon.” The ability to manipulate
real phenomena through the practice of computational model-
ing is what Alan indicates is primarily how he learns physics.
Dota references some work with using variables in different
contexts and how in a program, she would see this unfold in
instances where equations “call variables or reference val-
ues,” which she associates with “thinking like a Physicist.”
This practice of thinking like a Physicist is what she identi-
fies is the mechanism by which she learns. In these two dis-
cussions, these students indicate practices that each feels is
helpful in his/her learning of physics. But the key distinction
here is that by doing computation and engaging in the activity
itself, it helps the student to learn physics.
C. Computation Helps Me to Learn Physics When Interacting
with Problems In Set Contexts
This variation is seen in discussions where students de-
scribe some computational activities as helpful to their learn-
ing, while other problems as not beneficial. Based on state-
ments made throughout interviews, students talk about it in
ways that range from not getting additional information by
solving a particular problem, not expanding upon their previ-
ous knowledge of the physics content, or not building a con-
ceptual connection through that specific computational prob-
lem. Students identify some particular aspect or aspects of
the computational problem as helping them to learn physics,
but if those aspects were not present in the next problem or in-
dicated to no longer be helpful, then the next problem would
be viewed as not helping them to learn physics. Evidence for
this can be seen in a statement made by Australia:
Australia: Sometimes I think [computation] helps and
sometimes I think it doesn’t help. The other [problems]
helped me to grasp what happened when I changed something
in the equation or added something to the equations. Rather,
with this one I changed stuff and it didn’t really do anything.
It either made a dramatic difference or a slight one. I can
picture a satellite moving around the earth, I didn’t need the
[code].
Australia is talking about the differences between the first
two and the third of the computational problems. She talks
about how the third problem, the geosynchronous orbit of a
satellite, was something she could “picture” in her mind with-
out the code. In the first two problems, changing some of
the underlying equations in her program lead to a meaningful
change in the visual output, which seems to be important to
her learning of physics. The fact that the third problem either
“made a dramatic difference or a slight one” establishes that
the visualization aspect is important to her learning, but the
key point to this variation is that the computational problems
are only sometimes useful to her based on this criterion. This
variation suggests that the value a student places on an aspect
of computation in the current problem can depend on their
experience with that aspect in previous problems.
D. Computation Does Not Help Me to Learn Physics, But
Could Be Helpful to Others
There arose instances in the interviews where students
would discuss how the computational portions of the prob-
lems were not helpful to their learning. This could be based
on several factors, such as prior experience/knowledge with
computation or physics or the lack thereof. What sets this
apart from other variations is that these students then con-
tinue to talk about how it should or would help someone else
to learn. This is evident in the following quote from Captain:
Captain: Two people in my group haven’t taken physics, so
I think that helps a lot of them understand what they’re doing.
They actually see a model of what they [coded]. It helps a lot
of the understanding what you did, when you actually get to
see what you did. I would say no in my case, but it should
have helped someone else. Most of the things we are seeing
in physics now I already learned. So far it’s more like getting
better at physics, not so much at learning physics for me. I
would say what I did so far is something I already did in high
school.
Captain mainly focuses his discussion on the fact that this
is mostly material covered in his high school physics class,
and thus he is not learning any new material. He does men-
tion “getting better at physics,” but this idea was not fur-
ther explored. What is of interest in his discussion is that
he talks about two of his group members who had not taken
physics previously. Being able to “model” what they are cod-
ing should, as he says, help a lot with understanding what
they had done by actually “seeing what they did.” This
idea of “seeing” could refer to the visualization produced by
VPython or the written-out lines of code. In other cases, stu-
dents have said that the computational problems were useful
to other students who had some prior computational experi-
ence, but not to them because they had never coded before.
170
4
E. Computation Does Not Help to Learn Physics In General
Different from the variation above, students here talk about
how they do not believe that computation helps to learn
physics in general. Instead of being unhelpful only to the in-
dividual’s learning of physics or not helpful in solving only a
few specific problems, they consider it to not be helpful in any
case. Some evidence to demonstrate this variation is given by
Vandermeer:
Vandermeer: I don’t really learn the physics. It takes away
from the physics when I’m still trying to figure out how to do
an exponent, like two asterisks, parentheses, little things like
that. I feel like having a programming class as pre-req would
be quite helpful. I think the [analytic] labs are better at that.
I enjoy pen and paper just because I don’t have any coding
experience.
Vandermeer’s discussion revolves around the idea that
computation isn’t helpful for learning physics based on her
experiences, because to her, the programming “takes away
from the physics.” The reasoning she provides is that she
doesn’t have any coding experience and the syntax is diffi-
cult enough for her that her focus is shifted from the physics
content to writing a working program. She indicates that the
analytic problems are better at teaching her the material.
V. DISCUSSION AND CONCLUSIONS
We presented five major variations present in this emer-
gent theme. Three of the variations point towards a positive
impact on student learning via computation and students hav-
ing some positive computational experience in the class (A,
B, & C), while two are associated with students not view-
ing their computational experiences as having a positive im-
pact on their learning (D & E). Only four students provided
discussions that identify computation as not being helpful to
learning physics. However, the variation in their reasoning
provides a platform to address their concerns through alter-
ations to the design of activities and messaging in the class-
room.
Delving into the quantitative counts, the variation with the
largest number was variation B. Variation B’s focus on en-
gaging in the practices of computation and the perception
that "thinking like a physicist" is working with computational
models demonstrates a positive alignment with the learning
environment’s design focus on practices and the instructor’s
meta-messaging around the role of computation in physics.
We can also see from Table I that seven students displayed
variation A at some point during their interview, which is a
positive indicator that the computational activities are play-
ing a role in the development of student understanding in P3.
A deeper investigation into the specific aspects of compu-
tational activities that supported the development of under-
standing could again support future curriculum design.
From a methodology perspective, these variations include
some deeper nuances that we expect to understand further
when combined with other emergent themes during the com-
plete phenomenographic analysis. For instance, several of the
themes explored in the larger study focus on the role of the
visualizations produced by the computational models in stu-
dents’ learning. A future theme or variation that may emerge
in the complete analysis of our dataset could be focused on
student use of visualization. This could then connect to the
theme presented in this paper in evidence such as the quote
displayed by Captain in Variation D. This student references
the utility of "seeing" in learning physics. These connnec-
tions between themes will help fill out our outcome space
of students’ perceptions of the utility of computation where
multiple themes are used to reinforce one another [9].
Continued analysis of this dataset will yield additional sup-
porting information to strengthen the variations within this
theme and a greater understanding of the effects an integrated
computational approach to learning physics can have on stu-
dents. Due to the relatively short time period over which this
study occurs, our future work will also focus on understand-
ing how student perceptions of utility of computation change
over longer periods of time.
[1] AAPT Undergraduate Curriculum Task Force. "AAPT Recom-
mendations for Computational Physics in the Undergraduate
Physics Curriculum." 2016.
[2] Aiken, J. et al. “Understanding student computational thinking
with computational modeling.” Physics Education Research
Conference. Volume 1513. 2012.
[3] Caballero, M., Kohlmyer, M., Schatz, M. “Implementing and
assessing computational modeling in introductory mechanics.”
Physical Review Physics Education Research. Volume 8, Issue
2. 2012.
[4] Landau, R. "Computational Physics: A Better Model for
Physics Education." IEEE Computer Society and American In-
stitute of Physics. 2006.
[5] Tsai, M., Chin-Chung T. "Student computer achievement, atti-
tude, and anxiety: The role of learning strategies." Journal of
Educational Computing Research 28.1, 2003.
[6] Irving, P., Obsniuk M., Caballero, D. "P3: a practice focused
learning environment." European Journal of Physics. Volume
38, No. 5. 2017.
[7] Marton, F. "Phenomenography-Describing conceptions of the
world around us." Instructional Science. Volume 2. July, 1981.
[8] Braun, V. and Clarke, V. (2006) Using thematic analysis in psy-
chology. Qualitative Research in Psychology. Volume 3 (2).
[9] Svensson, L. "Theoretical Foundations of Phenomenography."
Higher Education Research & Development. Volume 16, No.
2. 1997.
[10] Backer, A. "Computational Physics Education with Python."
IEEE Computer Society and American Institute of Physics.
2007.
[11] Esquembre, F. "Computers in physics education." Computer
Physics Communications. No. 147. 2002.
171
