Cross-disciplinary instruction is common in STEM programs. Physicists teach engineers; chemists teach biologists; and mathematicians teach everybody. But scientific disciplines create distinct cultures – conventions, goals, expectations, and epistemologies – and these differences can lead to serious challenges for an instructor teaching out of her discipline. In this talk, I discuss how what has been learned in decades of Physics Education Research can help us better understand these challenges. And to use that understanding to figure out how to help our non-physics students get more authentic value from our classes — and to value what they have learned.
OECD bibliometric indicators: Selected highlights, April 2024
Rethinking Physics Service Courses:The challenge of cross-disciplinary STEM instruction
1. Albert
Donut boy
Schrödinger’s cat
Macavity
Emily
Cal Tech groupWill
Sheldon
Katherine
Rethinking Physics Service Courses:
The challenge of cross-disciplinary
STEM instruction
Edward F. Redish
Department of Physics (Emeritus)
University of Maryland
Iowa State University
October 12, 2020
10/12/20 Iowa State 1
2. Outline
• Who do physicists teach?
• It’s not us!
• The problem: They often don’t like our classes.
• The challenge of interdisciplinary service instruction
• Why doesn’t standard teaching work?
• What we learn from Physics Education Research
• A case study
• Introductory Physics for Life Science Students:
NEXUS/Physics
• But what can I do if …
• Back to reality.
10/12/20 Iowa State 3
4. STEM Education: Who teaches whom?
• We care about our majors. (Of course! They are future us.)
A lot of our instructional effort focuses on them.
• But we not only teach our own majors. STEM supports STEM
• “Service courses” represent a large fraction
of students being taught in many STEM departments.
• Physicists teach engineers.
• Chemists teach biologists.
• Mathematicians teach us all.
10/12/20 Iowa State 5
5. Physics (and math)
have particular issues
• The large classes in
engineering, chemistry,
and biology contain mostly
their own majors.
• We mostly teach students
in other departments.
0
200
400
600
800
1000
1200
Engineering
Biology
Other(ABP)
M
ajors
Number of students taught
in Intro Physics at UMd F/2020
10/12/20 Iowa State 6
6. A problem:
Non-majors often don’t like our classes!
The challenge of interdisciplinary STEM education
10/12/20 Iowa State 7
7. This can’t be a problem...
• We know how to teach physics.
• Physics is physics is physics ... for everybody.
• We have a working model – and dozens of standard texts
we can use with all kinds of resources.
• This makes teaching non-majors straightforward
– even if time consuming due to
the difficulty of administering large-classes.
• No thinking required!
810/12/20 Iowa State
🤷
🤦
8. Unfortunately...
1. Our non-physics STEM students often don’t like our classes.
• They see it as a hoop to jump through
rather than something that will help them professionally.
• They often resist the best we have to offer
(learning to think using math!).
• While this is especially true of our life-science students,
many of our engineers resist doing symbolic math.
• They often deride physics for being “unrealistic and irrelevant”.
• They may do the minimum needed to “get through” and
not engage intellectually in learning to reason physically.
• As a result, they may not bring to their professional classes
what they were supposed to get from physics.
910/12/20 Iowa State
😱
9. Unfortunately...
2. Our clients have also been paying attention.
• In 2000, ABET (the engineering accreditation organization)
shifted its focus from course requirements to learning
requirements (eliminating any specific reference to physics).
• In 2009, AAMC (the medical school organization) proposed
to shift its focus from course requirements
to learning requirements.
• Some departments (especially bio)
have dropped physics as a general requirement.
1010/12/20 Iowa State
10. Why doesn’t our “tried and true”
approach always work?
What we learn from Physics Education Research (PER)
10/12/20 Iowa State 11
11. Some deep problems
• Mismatch with the way most people learn.
(Selection fallacy)
• “I teach the way that worked for me.”
• Assignment of responsibility
(Deficit fallacy)
• “They’re just not good enough / putting in enough effort /…”
• Disciplinary differences
(Failure to see cultural differences between disciplines)
• “Wait,…what?! We’re all scientists here.”
10/12/20 Iowa State 12
12. The way people learn
13
• Over the past 20-30 years, PER has increasingly
revealed a failure of large traditional physics courses to
• build good concepts
• improve student attitudes towards physics
• help students learn to think scientifically
• Research-based instructional methods
have begun to document improved learning
– at least along some dimensions.
10/12/20 Iowa State
13. What works better?
• Thinking!
• Mental engagement! Sense-making!
Building networks of connected knowledge!
NOT memorizing isolated facts and algorithms.
• Making (and correcting) mistakes!
• Setting up situations where they get to try out ideas,
make mistakes, and work out WHY those are mistakes
• Talking!
• The process of science is a conversation!
They need to learn to talk through problem solving
and sense making.
10/12/20 Iowa State 14
14. We sometimes discourage best learning practices.
• When we give homework and exams where students
succeed by memorizing and rote reasoning,
we are sending a message that says
“physics is not about thinking or understanding.”
• Situations where only right answers are valued
are counterproductive.
10/12/20 Iowa State 15
15. What helps?
• What we as teachers do matters much less
than what the students do.
• We need to create situations
in which what they do leads to learning.
• Things we can do:
• Active engagement (when done thoughtfully)
• Give group activities (learning to talk to each other
about science and problem solving)
• Understand something about the way students think.*
10/12/20 Iowa State 16
E. F. Redish, Teaching Physics with the Physics Suite Chapter 2,
http://www2.physics.umd.edu/~redish/Book/02.pdf
17. The Biology/Medicine Reports
• Leading biologists and medical schools
have been calling for major reform
of undergraduate instruction
• In 2009, AAMC (with HHMI) published
Scientific Foundations for Future Physicians
– a call for rethinking pre-med education in the US to
• bring in more coordinated science
– biology, math, chemistry, and physics
• focus on scientific skills and competencies.
1810/12/20 Iowa State
18. NEXUS/Physics
• In 2010, my research group won a challenge grant from HHMI
and a research grant from the NSF.
• Our charge was to reform the physics class
for life science students as part of a national effort
to increase their skill development
and multi-disciplinary strengths. (National Experiment
in Undergraduate Science Education – NEXUS).
10/12/20 Iowa State 19
E. F. Redish, et al. (17!) ”NEXUS Physics,” Am. J. Phys. 82:5 (2014) 368-377. doi: 10.1119/1.4870386
19. Our context
• We teach physics to about 600 life science students
per term in a two-semester course.
• Both semesters are taught each term (including summer).
• We have moderately large lectures (~100-200)
• Classes include
• Lecture 3 hrs/wk: (3 x 1) or (2 x 1.5)
with clickers (TurningPoint, TopHat, or Canvas): 2-3 faculty
• 3 hours of recitation/lab per week
with 24 students per section: ~12 TAs
• Weekly homework, mostly online (using ExpertTA).
• Individual hour exams but a common final.
10/12/20 Iowa State 20
20. An interesting opportunity for deep research!
• We brought together a large team of physicists, biologists, chemists,
curriculum developers, and education researchers.
• We spent a year discussing what physics could do
as part of a curriculum for biologists and pre-health care students.
• We spent two years teaching the class in small classes,
videotaping everything in sight and learning as much as
we could from listening to (and interviewing) the students.
• This was a research project.
We documented it in ~30 peer-reviewed papers.*
10/12/20 Iowa State 21
https://www.compadre.org/nexusph/course/NEXUS_Physics_Publications
21. What we found: Barriers to interdisciplinarity
• Our ini]al nego]a]ons immediately ran into a brick wall.
Biologists and physicists had very different views of what to do.
• Many professional biologists saw most of tradi]onal introductory
physics class as useless and irrelevant – and our standard “we can
apply physics to biological examples” as trivial and uninteres]ng.
• This resistance was found in students as well. Many expected
physics to be useless. Most avoided taking it un]l they were seniors
(when the grade wouldn’t count for med or grad school applica]ons).
10/12/20 Iowa State 22
22. What do we have to offer?
Physics has value for STEM students!
10/12/20 Iowa State 23
23. Physics is the best place to learn to reason
about the world mathematically!
• Many non-physics STEM majors succeed
in their intro level courses by memorizing
facts, algorithms, and heuristics.
• Most (especially biologists – but often chemists too)
have never seen an example where using math
improves their professional understanding.
• We need to negotiate a change in student expectations
10/12/20 Iowa State 24
24. Keys to help students see the value of physics
• Articulation
• How does this course fit into the program of the students in the class?
• What can we do in physics that can be of value to them
in their professional classes?
• Authenticity
• There have to be examples where learning physics helps them better
understand things that they have only been told and learned by rote.
• Learning the math/physics blend
• Physics relies heavily on math, but the way math is used in physics
is dramatically different from the way it’s used in math classes.
A number of new skills need to be developed.
10/12/20 Iowa State 25
🔑
25. Articulation: The problem
• The population we chose to address were biology majors
(and pre-health-care professionals).
• They were advised to take physics as seniors since it was
not a pre-req for anything and was not seen of as valuable
for almost any other classes.
• The class had been taught as a gen. ed. course assuming students
had little or no experience with college level science.
(Many had 10 or more science courses before physics.)
10/12/20 Iowa State 26
26. Articulation: A solution
• After long consultations with biologists and chemists,
we designed the class to fit in the sophomore/junior year.
• We made intro bio and chem pre-requisites.
(All biology students already had to take calculus.)
• We chose to focus on the common biology core
– cellular and molecular biology – rather than on
topics of interest to small subgroups. (Like kinesiologists)
• We added topics seen as useful
• diffusion, fluid dynamics, molecular forces, binding energy, entropy…).
• We removed or reduced topics of little interest
• projectile motion, angular momentum and rotationals, magnetism, relativity).
10/12/20 Iowa State 27
27. Authenticity: The problem
• Many students saw physics as irrelevant to biology.
• The examples in most intro physics texts
had trivial biological content. They showed physics
applied to biological systems but didn’t show that
you could learn anything that helped you understand biology.
• “Consider a spherical cow.”
• ”A greased pig slides down a ramp.”
• “Convert the units of the wavelength of photons
relevant for photosynthesis.”
• We sought examples that students would perceive
as of real value in understanding biology (or chemistry).
10/12/20 Iowa State 28Watkins, Coffee, Redish, & Cooke, Disciplinary Authenticity, Phys Rev ST-PER, 8 010112 (2012)
28. The Math: The problem
• My students often see math as inappropriate in biology.
• As a result, they often resist thinking about the math.
• My students often find that despite A’s in calculus,
they can’t do the math I ask them to do.
• Even though the math is trivial algebra? What’s up with that?
• My students are desparate to put numbers in right away
and hate doing symbolic problems without numbers.
• And they wan’t to drop the units, only putting them in at the end.
• My students want me to give lots of “practice problems”
(plug-and-chug) rather than a few “thinking problems.”
10/12/20 Iowa State 29
29. Iowa State
“Epistemological misconceptions”
• These problems on the previous slide result from students bringing in
incorrect expectations about the nature of the knowledge
they are learning and about the role of mathematics in science.
• So. What do we do?
• Give up and say, ”these students just can’t do it”?
• Figure out what their problems are and see if we can teach them how to do it?
• If the second, we need to better understand the issues. Here’s the key:
The way we do math in science is very different from the way
math is done in math. We blend physical concepts with math
and this changes the way we interpret the symbols.
10/12/20 30
🔑
Meredith & Redish, Reinvendng Physics for Life Science Majors, Physics Today 66:7 (2013) 38-43
30. Really? Consider two examples.
A. 1 inch = 2.54 cm is a legitimate equation but 3 second = 3 cm is not.
B. The Electric field is defined by 𝐸 = 𝐹/𝑞 where 𝑞 is the charge on
the test (probe) charge. If 𝑞 is replaced by 𝑞/3. 𝐸 does not change.
The numerical quantities we deal with in physics are typically NOT numbers,
but stand for something physical that can be quantified in a variety of ways
depending on choices we make.
The symbols in the equation are not arbitrary mathematical placeholders.
𝐹 is not just a symbol; it’s the electric force felt by the charge 𝑞.
Changing 𝑞implies that 𝐹 will change as well.
10/12/20 Iowa State 31Redish & Kuo, Language of physics, language of math, Science & Education (2015-03-14)
31. Many problems arise if you don’t make the blend.
• Our equations tend to have many more symbols than
in intro math classes. Without the blend, picking out
what you should focus on can be difficult.
• Which are constants, variables, and parameters can change
in the same equation depending on what question we ask.
• We use not only equations, but graphs and diagrams to code
information about the physical world. Without the blend,
you don’t see what information is coded and can’t create
or interpret them correctly.
10/12/20 Iowa State 32
32. A particular problem with not blending
• It’s not enough to recognize a concept belongs to a parfcular
symbol. The blend has to be more contextual than that.
• We ogen use the same leher to represent mulfple concepts.
We figure out which is which by context and understanding
the blend as describing a parfcular physical situafon.
p = pressure
momentum
dipole moment
Q = charge
heat flow
volume flow
V = velocity
volume
voltage T = tension
temperature
period
10/12/20 Iowa State 33
33. Iowa State
We can teach the blend through
general purpose tools / strategies
Dimensional analysis
Reading the physics
in a graph
Estimation Telling the story
Functional dependence Diagrams
Anchor equations Repackaging
Toy models
Building equations
from the physics
10/12/20 34https://www.compadre.org/nexusph/course/Building_your_mathematical_toolbelt
34. Skills apply to all topics.
• Adding a focus
on developing math skills
doesn’t add new topics.
• It modifies the way we teach
topics we already include.
• It does require us to be “meta”
— to be explicit about tools
since students don’t easily
generalize.
10/12/20 Iowa State 35Redish, Using math in physics: Overview, to be published in The Physics Teacher (2021)
35. Dimensional analysis / functional dependence
• Most quantities we use in physics are quantified as a result
of measurement. They therefore have dimensions
(mass, length, time, charge, temperature) and units.
• Learning to see the dimensionality of symbols is a first step
in building the blend of physical concept with math.
• Learning to pay attention to the functional dependence
of a symbol (how it enters an equation) is a powerful tool
to build student perceptions of authenticity
of symbolic math.
10/12/20 Iowa State 36
36. An electric dipole consists of a pair of equal and opposite charges
separated by a small distance. Consider a dipole consisting of
charges +𝑞 and −𝑞 separated by
a distance 2𝑑. The dipole moment
is defined as 𝑝 = 2𝑞𝑑.
Which of the formulas at the right is
the correct law for the force between
𝑝 and 𝑄?
Evaluate if the dimensions of each
expression are correct and whether
the functional dependence is
physically plausible.
(For example, since we know the
force gets weaker as 𝑄 gets farther
away, the distance 𝑟 can’t be in the
numerator.)
10/12/20 Iowa State 37
37. Reading the physics in a graph
• Students often frame the creation of a graph
as an answer (something the teacher has asked them to
do) rather than as a tool
(a way of understanding something physical).
• The fact that each graph codes many different
bits of information and we use multiple graphs
to describe the same physical situation makes it
a GREAT way of helping students learn to build the blend.
10/12/20 Iowa State 38
38. A tilted airtrack has a spring at one end. A cart on it is pressed
against the spring and released. The cart bounces up and down.
The graph represents the velocity of the cart as a function of time
starting at the moment of release. Positive is to the left.
Which of the letters on the graph can
identify an instant of time
at which the cart is
1. instantaneously not moving
2. in contact with the spring
3. moving down the track
4. at its highest point on the track
5. has an acceleration of zero.
10/12/20 Iowa State 39
39. A superball is dropped from a height of 1 m and bounces
a number of times before it is caught. Below are shown
graphs of some of the physical variables of the problem.
Which graph could represent
1. The velocity of the ball
2. The kinetic energy of the ball
3. The gravitational potential energy
of the ball
4. The momentum of the ball
5. The total mechanical energy
of the ball.
6. The position (height) of the ball
10/12/20 Iowa State 40
40. Toy models
• One of the most powerful tools we use in physics in learning to make
sense of the world with math are toy models.
• We consider the simplest possible example that shows a phenomenon
and beat it into submission — until we understand it completely.
• These are extremely valuable starting points for building more complex
models of realistic situations.
• The problem is that biologists
(and engineers) tend to focus
on realistic situations
and see toy models
as bogus and irrelevant.
10/12/20 Iowa State 41
41. We need to be explicit about toy models,
motivate them, and consider corrections.
• Using toy models is a great start
for learning how to model more
realistic situations.
• But even a toy model involves
multiple steps and “the blend”.
• If we only give problems that
involve the top leg, students
think “it’s just math” and ignore
the critical steps of deciding
“What’s important?”
and “Is that good enough?”
10/12/20 Iowa State 42
42. Real springs only follow the Hooke's law model
for small displacements around their rest length.
Which graph might represent T vs L for a real spring?
A real spring behaves as follows:
• stretching from its rest length,
it obeys Hooke's law for a small stretch.
• As you stretch it further, it gets stiffer
as the coils begin to bend.
• Eventually it straightens out into a long
straight wire which is very hard to stretch.
• If you keep pulling harder, the wire suddenly
stretches easily and breaks.
• If you try to compress it, the coils get pushed
together and you can squeeze very hard
without getting much compression.
10/12/20 Iowa State 43
I teach Hooke’s law as
𝑇 = 𝑘Δ𝐿 rather than as
“𝐹 = −𝑘𝑥” to focus on
the physical blend.
43. But what can I do if…
Back to the real world
10/12/20 Iowa State 44
44. But what can I do if…?
•… I don’t have a PER group?
•… I have 300 students in my lecture class?
•… I have to teach online?
•… I have to use autograded homework?
•… I have to have scantron-graded exams?
10/12/20 Iowa State 45
45. If you don’t have a PER group…
you can learn from the research base
• You can contact PER researchers
who have learned a lot about engineers.
• University of Illinois (Gladding, Stetzer, Kuo)
• University of Colorado (Dubson, Wilcox, Finklestein)
• Colorado School of Mines (Kuo, Kohl)
• MIT (Belcher)
• …
• And explore some of the literature
• AJP Review articles
• Reviews in PER @ PER Central (on ComPADRE)
• PER Books (Mazur, Knight, Redish, Arons, Viennot,…)10/12/20 Iowa State 46
46. If you have 300 student (online) lectures…
you can still adopt PER-based active-learning tools
• Many good ones are available: See PhysPort
• Many of these ideas can be implemented in a large online lecture.
• Peer Instruction:
(1) Good clicker question, credit for participation
(2) Consult for 2 minutes in randomly assigned breakout groups
(3) Answer again, credit for correct answer
(4) Analysis and clarification
• PhET Sims:
(1) Pre-lecture reading
(2) Sim guided exploration, questions in LMS, participation credit.
(3) In class clicker quiz, credit for correct answer
• The key is thoughtful questions that require thinking
— not just knowing the answer.
10/12/20 Iowa State 47
And!
• UW Tutorials
• UI FlipIt Physics
• …
47. If you need to use autograded homework…
there are lots with good questions.
• Multiple-choice/short answer questions can be challenging
if they don’t just expect memorized or plug-and-chug answers.
• Many homework environments can handle questions
with symbolic answers. (WebAssign, ExpertTA,…)
• I create triples of questions that consider a single physical situation
from three points of view: qualitative, symbolic, and quantitative,
and give them in succeeding weeks.
• We reduce the TA grading in labs by having two-week labs and reports
written in teams of 4. This frees enough grading time that
we can have 1 hand-graded HW problem/week. (We give 3 “explain”
questions but don’t tell them beforehand which will be graded.)
10/12/20 Iowa State 48
48. If you have to use ScanTron exams…multiple-choice
questions can still test skills rather than just facts.
• Previously done questions (clickers, homework, quizzes) can be
presented with small modifications that change the answers.
• Set them up so if the answers are worked out by reasoning from principle
they’re easy, but the answers are different so a quick “I remember this”
will get it wrong. BUT YOU HAVE TO TELL THEM YOU’RE DOING THIS!
• You can create forked, multi-tiered multiple choice questions
that ask them to choose reasons after they have chosen the answer.
• You can ask “epistemological questions” that say “if you know this…
what can you conclude” where there is not enough information
to conclude the correct answer
(in which case the ”correct” answer may be wrong).
10/12/20 Iowa State 49
49. Whatever constraints you are operating under
• There are lots of opportunities to be creative
and to change things in ways that will improve
student learning and their appreciation of our classes.
• I was surprised and delighted by how many deep insights
I gained into physics that I thought I knew
— and how much new and interesting physics I learned
in order to reach my students
in a place that was meaningful to them.
• Working with biophysics colleagues was a great help.
I expect you can find engineering physics colleagues as well.
10/12/20 Iowa State 50
50. The path forward
is through
interdisciplinary
communication,
both with
colleagues and
students.
10/12/20 Iowa State 51
51. My biologist colleague, Todd Cooke and I,
summarized our interdisciplinary experiences
in a paper in CBE-LSE. Here is our conclusion.
• While we have detailed our experience, each interdisciplinary
exploration is bound to be unique to the individuals and
institutions involved. But what we expect will be common to
such activities will include: showing respect for each other's
discipline and insights; a willingness to reconsider one's own
discipline from a different point of view; and finally, patience,
persistence, and humor.
10/12/20 Iowa State 52
Learning Each Other's Ropes: Negotiating interdisciplinary authenticity, E. F. Redish and T. J. Cooke,
Cell Biology Education - Life Science Education, 12 (June 3, 2013) 175-186. doi:10.1187/cbe.12-09-0147.