Simplifying Learning Environments
For Introductory Students
David Wheat
System Dynamics Group, University of Bergen (Norway)
Virginia Western Community College
P.O. Box 19234 Roanoke, Virginia 24019
dwheat@wheatresources.com
Abstract
An interactive learning environment (ILE) appropriate for intermediate macroeconomics students
has been simplified for students in introductory college and high school courses. Without
changing the underlying model or the “flight simulator” options, the simplified instructional
approach relies on feedback loop diagramming more than stock-and-flow diagramming.
Interactive Vensim causal loop diagrams are embedded in a STELLA interface, using slide show
and video software. In addition, students appear to learn more as model-users if they engage in
preliminary model-building activities using simple word-and-arrow diagrams.
MacroLab is a STELLA model and interactive learning environment (ILE) for teaching
macroeconomics. Developed for college-level courses, MacroLab is currently used to teach
distance education students at Virginia Western Community College. The ILE uses the STELLA
“story-telling” feature to unfold a model of a national economy, one-sector-at-a-time, throughout
the course. A recent modification enables feedback loop diagramming! to be used, in addition to
stock-and-flow diagramming. When explaining a complex system in an introductory course,
loop diagrams appear to be more effective in shaping mental models than the stock-and-flow
approach. Similar benefits may be obtained in K-12 classrooms.
The feedback loop method is not always superior to the stock-and-flow method for
communicating the structure of a system dynamics model. No feedback loop diagram, for
example, can convey the “bathtub” imagery of a stock with an inflow and outflow. For complex
' Alternatively called “causal loop diagramming.”
lof7
models, however, the simplicity of loops has some advantages. An ILE that can tell the “story”
of a model using both stock-and-flow diagrams and feedback loop diagrams is a useful tool.
Feedback Stories in STELLA. As the model economy takes shape over a 16-week period,
the stock-and-flow structure of MacroLab becomes complex. STELLA’s excellent story-telling
feature does facilitate the presentation, but complex models require several stories. Moreover,
the detail in some stock-and-flow maps can distract beginners, particularly as a model evolves
during a semester. This is particularly challenging for distance education students who have to
learn both system dynamics concepts and economics at the same time, without any classroom
contact with their instructor. Moreover, in a face-to-face course taught last fall, I witnessed
similar difficulties even while physically present to help the students.’
For introductory students, I now rely on an alternative approach that emphasizes
feedback structure, using causal loop diagrams. While a formal assessment of the relative
effectiveness of the two approaches has not been done, my impression is that current students
find the loop diagramming easier to use, and they are developing stronger mental models of a
national economy than did my previous students.
Feedback loop stories can be added to the STELLA interface, using a combination of
simulation, presentation, and video software. Vensim is used to create the feedback loop
diagrams. Then PowerPoint and QuickTime are used to convert the static feedback loops into
annotated, interactive video that can be imported into a STELLA learning environment.’ That
enables the feedback structure of the model—complete with annotations—to be presented in a
series of building-block stories.
2
~ Intermediate economics students would have less difficulty, of course, and I have maintained the stock-and-flow
stories in MacroLab for that level of instruction.
Apple’s Keynote presentation software also comes in handy, as explained below.
20f7
Procedure.
1. Use Vensim to create links and loops, such as
figure 1. Copy and paste into presentation software,
such as PowerPoint or Keynote.
2. Use PowerPoint (or Keynote) to separate the
links and loops into pages that “build” the story.
Ungroup the pasted picture of the links and loops,
so that individual pieces can be selected and
deleted. If you want to tell the story of this feedback loop in four pages, then make four copies
of this slide. Then selectively delete links so that each “page” has just the information that you
want. Then add text to explain each slide. When you finish, it will look like the four slides on
the next page (figures 2.1-2.4). The slide show will provide annotated images of an unfolding
“story” of the links and loops of interest.
3. Use Keynote to convert the slide show into a QuickTime movie.’ In Keynote, open the
wages
—[+
unemployment
rate
GDP
labor
Figure 1
PowerPoint file. Then export the file in QuickTime format, choosing “interactive” and “full
quality” when presented with those options.
4. Import the movie into a STELLA model. Create an information button, and select the “import
movie” option. Then browse and select the file exported from Keynote. Clicking on the button
will activate the movie. Click on a slide to advance to the next one. °
* PowerPoint files can also be saved in QuickTime format, but the resulting movie is not easily controlled after
importing into STELLA, If PowerPoint movies imported into STELLA run automatically when opened, use Keynote
to make the conversion to QuickTime. Another hint is to create a duplicate title slide because the movie “jumps”
past the opening slide when opened in STELLA.
Unfortunately, it is not possible to review previous slides.
3 of 7
wages wages
4
+ +
unemployment
rate
GDP — GDP
Na Na
labor labor
Figure 2.1 Figure 2.2
A segment from the main reinforcing loop, As more labor is employed, the unemployment
showing that more employed labor increases rate falls.
GDP, and higher GDP means higher wages.
wages wages
4 4
—[+ —|+
unemployment unemployment
rate rate
— GDP — GDP
Na Ne
labor . labor
Figure 2.3 Figure 2.4
The link from the unemployment rate to As wages increase, less labor is employed.
wages is negative. If the unemployment As wages decrease, more labor is
rate rises, starting wages tend to fall, and employed. Thus, the link from wages to
that makes average wages lower than they labor completes a counteracting loop.
otherwise would be.
The end result is an annotated feedback loop story embedded in a STELLA interactive
learning environment. It can be used as a stand-alone instructional device for beginners or it can
be used in conjunction with the STELLA’s “story-telling” feature that explains stock-and-flow
diagrams.
4o0f7
Model-Builder vs. Model-User. Using MacroLab to teach economics raises the perennial
question of whether students can learn much from a model without actually building it. Clearly,
they would learn more if they could build it, but that is not a feasible option in an introductory
economics course. The challenge, therefore, is to provide students with experiences during the
course that are similar to some of the experiences they would have if they actually built the
model. I have addressed this issue with small “hypotheses” assignments.
In the distance learning course, the hypotheses assignments are implemented via the
online discussion board. Each hypothesis assignment asks students to draw a link between two
variables. For example, the question might be: “What causes GDP to change?” Students have to
draw a link such as “X > (+) GDP” and speculate what X might be (e.g., “productivity’),
accompanied by a brief explanation of their hypotheses. The ensuing threaded discussion
provides an opportunity for applauding insights, correcting misunderstandings, or (frequently)
clarifying how to draw causal links. The benefits are two-fold. First, when the students later see
those same variables added to a larger feedback loop structure, they already have first-hand
experience with the diagramming method. Second, they have already been involved in
discussions about the ceteris paribus relationships involving those variables. This preparation
makes it easier to move on to consideration of larger, more complex structures.
Incidentally, such hypothesis-building early in the course coincides with textbook
chapters that are introducing concepts (e.g., GDP, productivity, unemployment, and inflation)
and focusing more on measurement and historical reference modes than theory. The hypotheses
assignments require students to draw initially on their own experience and insights regarding the
causes and effects of macroeconomic behavior. Most students are probably unaware that they
are engaged in the early stages of model-building; they think they are “merely” offering educated
opinions. Model building emerges, then, from the clash of simple ideas expressed by students
5 of7
and instructor, rather than as a complex pre-ordained picture to be memorized. The table below
illustrates an actual threaded discussion among students and instructor on the topic of
productivity during a distance education course in macroeconomics, using Blackboard’s
Discussion Board feature.
Instructor: What do you think "causes" the productivity of labor to change? As you did in the
previous exercise, use the causal link language to summarize your hypothesis, and then offer a
short explanation.
Kim: technology ----- >(+) productivity
As technology improves companies can implement new equipment to increase productivity.
Ruth: absenteeism ------ > (-) productivity
When there is an increase in absenteeism it causes a decrease in productivity.
Instructor: Good, Ruth. This could occur for two reasons. The absentee is still employed but not
producing, and that brings down the average productivity for all workers. Second, the absentee
(when at work!) may be required to interact with others to get work done, and being absent may
cause the others to be less productive, too.
Tom: new techniques ---> + labor productivity
As new techniques for completing tasks are completed, productivity will increase since these
new techniques take less time than the old ones.
Jenny: product demand------ >(+) labor productivity
When a product is in greater demand, for whatever reason (new technology, fads, price
reduction, etc.) production will rise to meet the demand.
Instructor: You're misunderstanding productivity, Jenny. It is not the same as production.
Production is output per year. Productivity is output per worker per year.
Jenny: human capital------------ >(+) productivity
When companies send workers to learn new skills or improve on skills, then workers’
increases. Is this more what you mean?
: Absolutely, Jenny. By a "change in productivity," we mean a change in "output per
worker" during some time period (such as year, month, or hour). Productivity improvements
usually result from improvements in tools & equipment due to improvements in technology, or in
employee knowledge and skills. It could also result, as another student said, by reorganizing the
work more efficiently.
Bill: capital ---> (+) productivity
When a company invests in new equipment, it makes the job easier, thereby productivity is
increased as the worker can produce faster and with more ease.
This particular threaded discussion had many more entries. The same process could be
used in a regular classroom brainstorming session. By the end of the exercise, the basic reasons
6 of7
for productivity growth had been identified: the quantity and quality (technology) of capital
equipment and the skills of workers, including the way that work is organized.
Later, when “labor productivity” was added to a labor figure 3
productivity
#k
productivity to grow” had already been discussed (figure 3). skill
feedback loop version of the model, the issue of “what causes
technology capital
It was, therefore, much easier to move ahead with the layout
of the larger model.
More importantly, however, the students seem to develop a better intuition about the
whole process of “theory-building” when this approach is used. It doesn’t substitute for actually
building a model, of course. But building a model is often not a practical component of a
classroom exercise, particularly if it involves a complex model that takes an entire semester to
“build.” Students can participate in laying one-brick-at-a-time, however, and thereby have a
better chance of claiming some ownership in the final structure.
Conclusion. Educators may be interested in the simplified instructional approach that
relies more on feedback loop diagramming than stock-and-flow diagramming, plus the method
for getting interactive Vensim loops into STELLA learning environments. They may also be
interested in ways to provide at least some model-building experience to students who are
primarily expected to be model-users when learning about a large complex system. This paper
provides an overview of both efforts aimed at simplifying learning environments for introductory
students. The author welcomes comments, suggestions, and questions.
7Tof7
CC BY-NC-SA 4.0