Data Driven Inquiry: Reforming the Teaching
of Science 101 Through
the Use of Instructional Technology
Gregory D. Bothun, Department of Physics, University of Oregon
Overview
The integration of research into the general undergraduate
curriculum can take a variety of forms. At the lowest level,
this form consists of a professor who does research being
the principal lecturer to a class full of students. While
neither interesting or creative, many faculty at the research
university actually believe that they are exposing students to
research by simply lecturing to them. The belief is that
if the students are able to see a real life researcher, they
will be molded and inspired by such a vision. This form of
"integration" remains the dominant one at the University of
Oregon, and I suspect at most any Big State Research University
(hereafter BSRU).
At the opposite extreme, actual student research with faculty
mentors would earn students course credit. Of course, this
scheme is incredibly expensive to implement and the resource
starved BSRU environment can't come close to providing this
kind of learning opportunity for all but a very small percentage
of its total students. In between these extremes lies a vast
continuum of levels.
In this contribution we will demonstrate
how technology and computer networks can be used to facilitate
the integration of research, more specifically actual data products,
into the teaching curriculum. Despite the potentially large
payoff, large scale implementation at BSRU has not yet occurred
because of a variety of barriers that change with time.
We will confront some of these barriers directly in a way that
hopefully demonstrates them to be indirect reflections
of a general inability for BSRU to create a range of learning
environments for its students. This, of course, requires an
investment in general undergraduate education as well as a
recognition that things are different in 2001 compared to 1901.
The Scale of the Problem:
Before launching into the potential that technology has to transform
the teaching and learning experience, it is important to be grounded
in reality. For the rest of the discussion we will make BSRU a
generic university of 20,000 students that is mandated by the state
to accept all reasonably qualified high school graduates. To
graduate from BSRU you must take several general education courses.
To meet the science requirement, we assume that four courses need
to be taken by each of the 20,000 students. We further assume that
the average graduation time at BSRU is 10 semesters. Since four
courses are required then, at any given time, 4/10 x 20,000 students
are enrolled in at least one general education science class.
That's 8000 students! (1/2 of which will be in Astronomy 101).
So the typical situation at BSRU is that thousands of undergrads will
be enrolled in some introductory science class for non-majors for
the sole purpose of fulfilling a general education requirement.
In general, at BSRU the class size for this exercise is in the range of
150-300 students (and can be higher). Its been my experience (at
the Universities of Washington, Michigan and Oregon) that this class
is an exercise in student seat time and generated student credit hours
for the host department.
One professor together with a few TAs, spews words or show slides
to the students for a term and the students get their credit and move on.
The opportunity to use these classes to actually teach science as
a process and to instill science literacy in the students is wasted.
As will be shown below, technology offers a way out of this dilemma
but implementation on the required scale remains a challenge.
Restructuring the Mass Lecture:
A major obstacle towards effective implementation of pedagogical changes
lies in our collective inability to admit the many deficiencies of the
mass lecture approach. This sardine can method of instruction certainly
represents the lowest unit cost to the University and so it is
administratively favorable. The suggestion that the students don't really
learn anything in this environment, somehow, seems to carry little or
no weight.
Conversations with colleagues who generally teach these classes reveals
the second obstacle to change. Approximately 80% of the professors
I have discussed this "problem" with actually believe the students are
learning the material directly from them, even when they are speaking
to 300 students at a time. Perhaps the word learn has been confused with
the word "absorb" in this case. Or, perhaps it is I that just doesn't know
what learning is. On the other hand, approximately 20% of the professors
readily acknowledge that the mass lecture is a student credit hour
generating machine for their department and that student learning is
really not a primary goal. Hence, the first step for pedagogical change
at BSRU is to be honest about the nature of these courses and the
reason for their existence. When scholarship asymptotically approaches
zero, change might result.
Unfortunately, there really is a lot at stake with respect to this issue
of pedagogical and structural change in these mass lecture science classes.
For many University students nationwide, an introductory Physics or Astronomy
course represents their last formal exposure to science.
The nearly ubiquitous pedagogy
is lecture presentation of ``factual'' material, even though this is a poor
representation of science and its practice. Not surprisingly, this
approach to teaching science tends to produce a scientifically illiterate
general population with an overall disinterest in the process of science.
As scientists and educators, we know
that the ability to perform unbiased and accurate observations and
experiments is crucial to scientific advancement. Yet the primary teaching
aid is a textbook which usually cannot communicate this vital aspect of science.
To teach science as a process, the student has to be
empowered to make an observation or perform an
experiment, formulate a hypothesis, then test the predictive power of
the hypothesis by performing new experiments. In this way the hypothesis
(or model) is either refuted or refined. While physical resources
are not available for the mass students to engage in this practice,
the use of robust simulation software is more than an adequate substitute
and provides a dimension of learning that is sorely missing.
Indeed, the real nature of science is that of a discovery process,
and yet science is seldom taught in this manner. We
need to put the sense of discovery back into the scientific curriculum
and that can most practically be done using software and simulation.
Five Reasons why the Mass Lecture is Broken:
There are
many deficiencies in the traditional mode of teaching a large lecture class.
These deficiencies, noted below, can be corrected by the preparation and
delivery of a more dynamic, innovative and interactive curriculum which
can be developed, given adequate resources. We will first list and comment
on the deficiencies and then move on to discussing how technology can
overcome each of these five limitations.
1. The Passivity Problem:
First, there is little direct contact between students and the instructor.
Lecturing in front of a large class encourages passivity by the students,
few of whom will typically come to office hours or otherwise seek help.
Research shows that students often have preconceptions or naive beliefs about
scientific concepts. In the standard mode of a
lecture, that of transmission from expert to novice, these prior conceptual
frameworks are rarely, if ever, challenged. Often, preconceptions prevent the
student from assimilating new concepts and knowledge.
These stereotypes are duly reinforced in introductory science
classes in which the large class size precludes
active experimentation, and the students are exposed only to traditional
classroom lectures interspersed with a few classroom demonstrations. As a
result, science is often taught as a collection of ``facts'' about the physical
world. Memorizing the facts of ``science'' and then regurgitating them at exam
time is certainly not the same as understanding and appreciating the process
of science.
2. The Facts Problem:
Second, in direct response to the passive nature of our current pedagogy,
there is a critical need to re-engage our introductory science students in
a mode of experimentation and discovery that will lead to
student-driven inquiry.
The current state of pedagogy in these large introductory science classes, in
contrast, relies heavily on static material as there is no practical mechanism
for students to interact directly with the material via experimentation. As a
consequence, most of the students in these classes remain disinterested in the
material and do not gain an adequate understanding of it.
Indeed, getting real
data and tools into the hands of students will allow them to duplicate the same
steps that the real scientist undertakes.
3. The One Size Fits All Problem
Third, the lecture mode of presentation is poorly suited to the range of ways
that students learn. Cognitive research identifies four distinct learning
styles. Concrete learners use direct
experience, doing, acting, sensing, and feeling. Abstract learners are skilled
at analysis, observation and critical thinking. Active learners are best able to
apply new information to facilitate tasks. Reflective learners prefer to reflect
and think about new information. Traditional delivery-oriented education
emphasizes procedures that best serve abstract and reflective learners. Large
lectures can deny or limit the active, hands-on, experimental approach that
many students require. A single approach to teaching and assessment will lead
to failure for many students who would be well-served by a more diversified
approach to learning. In particular, simulations can be used effectively
to allow the students to make mistakes in some experiment or situation without
feeling penalized.
4. The Syllabus Problem
Fourth, traditional classes vest the instructor as the primary information
source. The instructor leads the students linearly through the curriculum and
the textbook to reach a predetermined outcome. Student success or
failure depends on their ability to master (e.g. memorize) the body of
information presented in lectures or in the textbook.
I refer to this as the
syllabus problem (addiction). That is, the class is designed strictly
on the basis of a one page listing of topics. The goal of the course is
to get through the topics. Most students and faculty would be better
served by ripping up their syllabus and stating the learning goals of
the course instead. Because of the widespread nature of the syllabus
problem, it is no surprise that most students perceive science as a dry,
static body of facts. Ideally, the student should be free to navigate through
the course content, with the instructor as a guide and facilitator. The leading
edge of science consists of recent research results, technology being pushed to
its limits, and the cut and thrust of ideas being debated. The course material
needs a solid base of reliable knowledge, but it should reflect the richness
of active research as well as the process that leads from data acquisition to
new knowledge. Paring down the topical list so as to promote better student
engagement and debate would make for a far more powerful learning experience.
5. The Isolation Problem
The final problem with large lecture classes is the way that they limit the
opportunities for cooperative learning. In an ideal world,
students would break into smaller sections that would
center around experimentation and student-driven inquiry. While
there are examples where this is done (e.g. Eric Mazure at Harvard)
these are the rare exception rather than the ubiquitous rule. Instead,
the students tend to be isolated from one another and therefore can not
benefit from peer learning and collaboration. Such isolation tends to
reinforce to the student that they should act in a competitive manner
instead of a cooperative one as their grade depends on the former and
not the latter. Of course, forming a peer learning environment is
expensive and few public
universities have the budget for these resources and rising enrollments mean
that class sizes are never likely to get smaller. With the conventional mode
of classroom lectures, there are few ways to encourage students to do research,
work on projects and papers together, or communicate with each other. The large
lecture format, in essence, precludes student collaboration and questioning of
the material which ultimately leads to decreased motivation on the part of
science educators to teach these classes.
Instructional Technology as a Means to Improve Pedagogy
Each of the five ills associated with the mass lecture can be partially,
or in some cases completely, mediated with the appropriate use of
instructional technology. In what follows, we will describe how
the mass lecture that generally defines Astronomy 101, can be transformed
into a real, data-driven learning experience, by using software
and simulations. However, to be clear, what this ultimately means is
more work for the students as the burden of learning shifts to them.
Implementation of the scheme below has most certainly resulted in
lower teacher evaluations as a result of increased workload
on the student.
However, I would argue that the scholarship of the class has greatly
increased and in the end, that should be the measure of effective
teaching.
The Passivity Problem
Most of Astronomy 101 consists of showing "pretty pictures". This
is a mirror of the passivity problem and is ironic in the sense that
when you show, say the latest and greatest image from the Hubble
Space Telescope, its not really a pretty picture, its a digital image
upon which you can measure things. Similarly, a spectrum
of star is not an image that you memorize its a digital collection
of data that you analyze. The heart of our curriculum reformation
movement at the UO is to construct interfaces to data that allows the
student to download and actually measure the data. This is all done
in the JAVA programming language so that the Web Browser, which
most all students are now familiar with, is the only piece of software
the student has to use. So we have constructed JAVA based engines that,
among other things, will allow students to:
- Identifying features in stellar spectra that come from
certain elements.
- Measure the strengths of these stellar features as a function of
temperature
- Measure intensities of the pixels that compromise and image to
construct radial light profiles
- Measure the spectra lines in a galaxy spectrum to determine its
redshift
- Fit a radial velocity curve of a star to determine the mass of the
planet in orbit about that star which is necessary to produce the
observed radial velocity variations.
- Run an N-body experiment to reproduce the observed appearances of
interacting galaxies.
These are done in the form of standard homework assignments and students
work together in groups (more on this below). In doing this, the students
are now forced (well it is an assignment) to analyze real astronomical
data as a standard part of the course. This is a significant amount of
work on the part of the student but it does immerse them in data. Much
emphasis is put on how noise effects the kinds of measurements that are
made and we have some robust JAVA applets that simulate noise on a
detector. The pedagogical goal here is for the students to learn about
signal-to-noise, noisy data, sparse sampling and what a detection is.
Success is measured by the number of e-mail questions asserting that
the virtual apparatus is "broken" because it doesn't return the same
answer as the example in the noise-free textbook.
By getting data products into the hands of the student, and an interface
for measuring and analyzing the data, instructional technology has gone
a long way in removing passivity from the learning environment. Students
now have to directly work with and think about the data. While this
requires significant development (and is therefore expensive), this is
what the technology was made for. Interactive data driven exercises are
the main reason that instructional technology should be adopted as
a standard part of this course. Spiffy lectures in Powerpoint (the current
main use of the medium) pales in comparison to what this technology
really can offer the student.
The Facts Problem:
In this case, while instructional technology can facilitate a move away from
facts and towards inquiry, the barrier of student attitude may be the
most difficult to overcome. Students expect to "learn" material simply
by memorizing the facts spewed by the professor. They are quite
uncomfortable with the notion that they are supposed to discover the
"facts" themselves via some guided, interactive journey through data.
Many students, when confronted with the scientific reality that there
is no clear answer, become frustrated and confused. Their brains are
hardwire to equate learning with fact memorization because, afterall,
that's what is on the test! All that instructional technology can really
do in this case is to offer a facilitated avenue of inquiry as another
mode of learning. Re-programming the students to get away from fact
memorization as the principal mode of learning is considerably more
difficult. Of course, it is precisely this programming which is at the
heart of the science literacy issue. Hopefully, improved teaching
methodology at the K12 level will eventually produce a population of
undergrads that would rather question than memorize.
On the other hand, even in the fact memorization mode, instructional
technology can play an improving role. This interface allows the professor
to customize the curriculum around their own material and thereby
inject their research directly into the classroom. Even though the
students will not have inquiry based modules centered around that
research, they will at least be exposed to it so that the subject matter
might be come more real and less abstract.
The One Size Problem:
This is where instructional technology is ideally suited. It is clear
that the best learning occurs in a one-on-one environment. Indeed,
the greatest amount of learning that occurs during the entire
term in the typical 300+
student class, likely happens when 2 or 3 of them (accidentally)
wander into your office and ask a real question. Clearly, BSRU does
not have any where near the resources to replace classroom learning
with one-on-one learning via physical means. However, software is
generally used in a one-on-one basis. In principle, it would be possible
to design and implement intelligent agent software that would allow the
individual user to customize their journey through the resource base
that comprises the curriculum. The "game-like" nature of such a journey
might even motivate the students more than the traditional lecture.
Such curriculum flexibility would help to overcome the one-to-many and
one style of teaching that plagues the classroom environment.
Indeed,
nowadays it is customary for students to make mental models of their
professor and guess (assume) what pleases them. This kind of performance
is unrelated to actual course material. Why students do this is anyone's
guess but it clearly does happen. How do I know? Well, recently I taught
a course that was mostly configured to operate in a self-paced discovery mode. For the first
three weeks of the term the students were extremely confused. They did not
know what to do. One student directly told me that she felt "incapable" of
learning on her own and depended entirely on the professor telling her what
to do. This kind of student attitude is a clear
warning that the learning process has been effectively paralyzed by our
traditional approach and therefore structural change is required if we
want to achieve science literacy as the principal learning outcome.
The Syllabus Problem
One of the potential teaching resources that the Internet offers is the ability
for the student to interact with a diverse and distributed expert knowledge base.
The sage on the stage is no longer the sole keeper and presenter of the information.
If used subversively, the student learning experience can be greatly enhanced.
As just one example: I give a series of lectures on the dwindling nature of the
fossil fuel reserve and the likely "fact" that we only have 50 years worth of
of fossil fuel production on the planet. I then assign the students a number of
other expert or knowledgeable web sites to go research and find credible
information to prove me wrong. The students are generally motivated by this kind
of assignment. If well crafted, certain classes, particularly those in Environmental
Science, can be a sequence of research experiences for the students who then have
to put together disparate sources of information to synthesize a conclusion. Wow -
you just taught the scientific method instead of the syllabus. Which one is more
important?
To be explicitly redundant, the syllabus problem exists because we collectively fail to define the learning
goals for these general education classes. In my own case, I am no longer concerned
that my students are exposed to "less" astronomical topics (to memorize, of course)
because I know they are actually doing science. My fundamental pedagogical
goal is for them to learn the pivotal role of uncertainty in scientific
inference. Making the students work with noisy data or
noisy (virtual) detectors emphasizes this point far more elegantly than I could ever
do in a lecture.
Collaborative Learning
If you disagree that instructional technology can offer a better learning experience
for large lecture classes then at least consider the following. One aspect of
undergraduate education that is making it irrelevant to the real world is its
continual failure to promote collaborative learning environments. These days, the
real world is an interdisciplinary team oriented environment in which individual
team members are expected to make real contributions to the overall project. But
we don't teach Collaboration 101 at BSRU! Higher education really needs to
examine this issue and to ask itself whether a focus on collaborative learning would
make undergraduate education a more valuable experience.
Instructional technology can greatly facilitate student-student communication.
Team exercises around data projects can be given and network tools can be made
available to assist the teams with their data analysis and reduction. This is
a powerful learning environment with peer-to-peer learning constantly occurring.
Combined with group presentations in class, the students learn good written and
oral communication skills as well as develop their organizational skills. Although
the students initially do not like to work in groups, they generally evolve
through the term. The beginning stages are highly dysfunctional but if one is
patient, the end product is worth the wait and is very satisfying because you have
actually accomplished something. Generally, student reluctance to participate
in the group process is related to its perceived
effect on their grade (i.e. "what if I am in
a bad group ...; everyone in my group is a slacker ...). Once the grade
perception hurdle is overcome,
students develop a better attitude and generally have a pretty good time in
the group learning environment.
I strongly believe that this aspect of the deployment of instructional technology
completely justifies the investment. The real world of scientific research is
collaborative in nature. Any university department functions by collaboration.
Collaboration is what we do in our professional lives all of the time. Why are
we so reluctant to teach in that way? Again, there is a significant development
effort involved in constructing group exercises around data but it can be done
and in the end is highly worthwhile.
For example, we have developed a web based content organizational tool
for groups to
construct presentations from available resources. Specifically, for a group
project on global climate change, a wide array of data sets, images, and charts
were provided for all teams. The teams could then select which data products best
supported their particular perspective and craft a presentation around them.
At all times, individual group members could see the project evolving and to make
their own contributions. While inefficient the first time this system is used,
over the course of the term the students learn how to use the tool and presentations
improve considerably. The major learning outcome lies in the newly acquired
collaborative skills of the students. They are likely to retain this skill
set long after the syllabus topics have been forgotten.
Barriers to Implementation
I have outlined some of the potential that instructional technology has in facilitating
the integration of research into teaching and how it can be used as a tool to
modify, or even greatly change, the nature of the general education science class.
Despite this potential, the majority of these classes are taught the same way in 2001
as they were in 1951 (perhaps even using the same textbooks). While inertia on glacial
timescales is part of the higher education landscape, there must be additional reasons
for why systemic reform in general undergraduate education is so difficult to implement.
While its hard to quantify the reasons, qualitatively, based on conversations and
interviews with the relevant players, the following barriers have come to light:
- Faculty entrenchment: At BSRU the faculty are there to do research. They are
not there to teach. They have never been rewarded in the past for quality teaching
but know that the failure to produce quality research will terminate their appointment.
Without the understanding that there can be scholarship associated with teaching, this
problem is unlikely to diminish. The political motivation for the
RAIRE program was
clearly to get university administrators engaged in thinking about ways to balance
the research/teaching portfolio in more productive ways when it comes to tenure and
promotion. To a limited extent this has happened. On the other hand, the wholesale
failure to appreciate that the development of interactive electronic curriculum resources,
which are actually based on real data, is the same as research, is the root of the problem.
We are way too locked into peer review as the coin of the realm when it comes to
the evaluation of scholarly materials and therefore are unable to come with the
correct rubric under which to evaluate teaching scholarship. This must change.
- Administrative reluctance: From the fiscal point of view, the mass lecture is
wonderful. It represents the lowest unit cost instruction and any alternative approach
would cost more, either in terms of course development time and money or in terms
of human resources. These courses more than pay for themselves with the collective
tuition dollars of the enrolled students. But, so what? Our mission is to provide
the students with a quality education. The defense of the mass lecture as a
wise fiscal
instrument is tantamount to declaring that we are not interested in investing to
improve the quality of education (although the reality is that BSRU probably doesn't
have the resources to do this). But how much longer can this approach continue -
10 years, 100 years, 1000 years? Given that education is now a market commodity and
that schools must compete (which usually means they must differentiate themselves
to the market) I am betting that the BSRU systems that are on the 1000 year timescale
will become extinct. They should.
- Student passivity: Students are most definitely part of the problem here as they
put up with, and probably prefer, the spoon fed knowledge approach to lower undergraduate
courses. Attempts by myself to mobilize the students into a protest against spoon
fed knowledge have utterly failed. Students themselves have been allowed to become
so passive that they no longer are engaged in the learning process, let alone taking
responsibility for it. I firmly believe that if the students spoke up and said that
a) they wanted more collaborative work and b) they wanted better deployment of technology
as part of their course curriculum, then the glaciers would move faster.
As an aside; one of the reasons that there are often large gaps between the
RAIRE and AIRE schools is precisely this point. The kind of passive student described above has to be
admitted (by state law) to BSRU but not necessarily to smaller, private schools. Indeed,
much of what I have written above must sound, well, idiotic, to those faculty that
are at such schools. The reality is that the passive nature of the BSRU students largely
defeats any attempt at active and collaborative learning. Its a real effort
to overcome this (and such effort is
never rewarded by the teaching evaluation process) but in the end the
spent effort is most definitely worth it.
- Infrastructure Limited: Finally there are important issues of infrastructure. The
primary reason that the University of Oregon was able to experiment with the use of
network technology as a means of integrating research into teaching was due to its
excellent network infrastructure. By 1993 there was Ethernet to every faculty desktop.
Two years later the classrooms were all wired as were the dorms. So the access issue
was solved before the instructional development effort began. By so doing, we were
able to avoid making fatal mistakes that would have turned faculty and students off
forever. You can't assign the students to use a networked data reduction tool if they
don't have easy access to the network. Fortunately, the UO was in a good position
in this regard so that this approach could be implemented.
- Go away - it costs too much: Of course, the real barrier is that the development
of quality interactive electronic materials is time consuming and expensive .
The industry rate is $2-3,000 per hour for interactive multimedia. Traditional
textbook publishing companies consistently underestimate (by at least a factor of 10)
the true development costs. As a result, the commercial world offers us course management
tools but very little in the way of content development. The money is out there but
it requires new organizational partnerships and new ways of looking at what a course
is to capitalize on them. Ideally development would be done by giving faculty members
course release time to work with graphic artists, programmers and instructional
designers to convert their static content into a dynamic, interactive and multi-faceted
journey. To date, few, if any Universities, have invested in their own
faculty expertise and creativity for the purpose of superior curriculum
development. History will reflect that this was opportunity lost.
Beware of Shovelware
Which brings us to the last issue: the triumph of style over substance.
Universities are clearly under some pressure to become more "hi-tech". The
general response of the University is the creation (sometimes mandated)
of the ON line syllabus. What is the ON line syllabus? Well, professor X
takes their paper syllabus (because they are embedded in the syllabus
problem) and converts it to digital form and posts in on the Web.
Student Y then accesses it and prints it thereby using the technology
as a Xerox machine. This is shovelware . Shovelware
can be defined as converting the content from one medium to another
medium without taking any advantage of the unique characteristics of the
second medium. If you can go backwards (e.g. WEB to paper),
then its shovelware.
This brings me to what I perceive as the fundamental problem in the current
use of instructional technology. That is individual faculty
usually try to stuff the existing course framework and infrastructure into
the technology box rather than to use the technology to teach in a new
manner . Are we really this un-creative as a faculty?
Are we really just relying on instructional technology to provide us
better course management tools? The unique aspect of the Web lies in
its inherent interactivity and that's where the development effort
needs to be centered.
This inherent interactivity opens up whole new avenues and styles of
teaching. It provides the potential to customize a curriculum
around the field's data products that the students can interact with.
It provides the potential to create a simulation environment in which
the students can perform virtual experiments or analysis and make
mistakes along the way without those mistakes being documented so as to
demoralize the student. It allows the students to interact with the
material in new and different ways (that are quite difficult to
properly assess). Most importantly, it can be used to facilitate
group collaborative projects that build written and presentation skills
and gets the students to work together on a variety of different data
driven projects.
Such uses of instructional technology are, therefore,
not shovelware. Instead, they represent a robust and responsible
use of the new opportunities that network technologies have provided us.
We can either embrace these new tools as a means for changing the learning
environment or we can comfortably keep our heads in the sand assuring ourself
that the way we have always done it at BSRU is the best.
way.