Students’ use of physical models to
experience key aspects of scientists’
knowledge creation process
Kim A. Kastens, Education Development Center
Ann Rivet, Teachers College
Cheryl Lyons, Teachers College
Alison R. Miller, Teachers College
Bridging the Gap between Tabletop Models and the Earth System
NARST
Puerto Rico
April 8, 2013
Our teachers—excellent though they
are—do not view teaching about
models as a goal of Regents Earth
Science
STANDARD 4:
Students will understand and apply scientific concepts, principles, and theories pertaining to the physical
setting and living environment and recognize the historical development of ideas in science.
In our professional development, we
discussed…
Students aren’t just learning about moon
phases, or seasons, or deposition…
At the same time, they are learning about
models and modeling.
Strategic Decision: How overt do you want to be
about discussing the scientific practice of
modeling?
Very In the middle Not at all Kinds of models
Expressed models
Physical (concrete) models
Static scale models
Diagrammatic models
Runnable physical models
Mental models
Mathematical models
Computer models
Kinds of models
Expressed Conceptual Mental models Model
models
Physical (concrete) models
Static scale models
Diagrammatic models
Runnable physical models
Mathematical models
Computer models
All are “simplified representations
of objects, structures, or systems
used in analysis, explanation,
interpretation, or design.”
“Models are simplified representations of objects, structures, or
systems used in analysis, explanation, interpretation, or
design.”
• Models help us explain our ideas more clearly to
other people.
• Models get our ideas out there in public where we
can argue about them with other people and test
their validity against data.
• Model + Brain can answer harder questions and
solve harder problems than brain alone.
Models are not toys for children; models are brain-extenders for scientists.
Expressed runnable models are not just for communicating or
demonstrating or revealing what scientists already know….
physical model
computational model
Such models are tools for creating new knowledge.
OK. So how does this brain-extending process work when scientists learn
from external runnable models?
Make Observations
Interpret observations as result of processes
Represent processes in an expressed runnable model
Make more Observations
Make predictions to challenge model
Improve model
Use model to infer system behavior at times and places where you have no observations
Compare behavior of model with behavior of earth captured in data
The problem:
Make Observations
Interpret observations as result of processes
Represent processes in an expressed runnable model
Make more Observations
Nearly
invisible to
students and
the public
Make predictions to challenge model
Improve model
Use model to infer system behavior at times and places where you have no observations
Compare behavior of model with behavior of earth captured in data
Earth Science example: Global climate model
Milankovitch model: orbital forcing 
glacial/interglacial
Seasons; Latitudinal variation;
ice ages
Represent processes in an expressed runnable model
Interpret observations as result of processes
Make Observations
Predictions about latitudinal distribution of warming with increased CO2 Make more Observations
Make predictions to challenge model
Add positive feedback loops
Improve model
Poles are warming faster than low Compare latitudes
behavior of model with behavior of earth captured in data
Model behavior matches timing but not amplitude in sediments
Use model to infer system behavior at times and places where you have no observations
Forecasts of future sealevel
and temperature
How can we make the obscured steps more salient?
Make Observations
Interpret observations as result of processes
Represent processes in an expressed runnable model
Make more Observations
Nearly
invisible to
students and
the public
Make predictions to challenge model
Improve model
Use model to infer system behavior at times and places where you have no observations
Compare behavior of model with behavior of earth captured in data
Our project developed three candidate
strategies for better teaching and
learning with models
(a) Explicitly teach and practice analogic
mapping, articulating correspondences
and non-correspondences
Familiar Unfamiliar
(b) Use models as a tool for
solving problems and
answering questions.
(c) Use models as a tool
for reasoning about
Earth data.
NSF DRL09-09982
Strategies suggested in our project:
Make Observations
Interpret observations as result of processes
Represent processes in an expressed runnable model
Make more Observations
Make predictions to challenge model
Improve model
Use model to infer system behavior at times and places where you have no observations
Strategy A: Map correspondences and non‐
Compare correspondences between behavior of model model and referent
with behavior of earth captured in data
Strategy C: Use
models as a tool
for reasoning
about Earth data.
Appendix F: Scientific & Engineering Practices
Practice 2: Developing and Using Models
NGSS Practice 2:
Interpret observations as result of processes
Make Observations
Gr 12: Use models… to predict phenomena….
Represent processes in an expressed runnable model
Gr 2: Develop … models (e.g. … physical replicas….) that represent … patterns in the natural … world. Make more Observations
Make predictions to challenge model
Gr 8: Modify models—based on their limitations—to increase detail or clarity Gr 8: Modify a model … to explore what will happen if a component is changed.. Improve model
Use model to infer system behavior at times and places where you have no observations
Compare behavior of model with behavior of earth captured in data
Gr 12: evaluate the merits and limitations of two different models of the same … system in order to select … a model that best fits the evidence. Would it possible for students to experience the entire knowledge-construction
Make lunar cycle using simple physical
calendar models? We think so….
From assorted balls, lights, sticks, string
Make Observations
Interpret observations as result of processes
Represent processes in an expressed runnable model
Make more Observations
Make predictions to challenge model
Improve model
Use model to infer system behavior at times and places where you have no observations
Compare behavior of model with behavior of earth captured in data
Ambiguity: which way does moon go?
Improve model
Compare behavior of model with behavior of earth captured in data
.
✗
North Pole
Northern Hemisphere
.
✓
North Pole
Ambiguity: Which direction does the Moon go around the Earth?
Would it possible for students to experience the entire knowledge-construction
Make lunar cycle using simple physical
calendar models? We think so….
From assorted balls, lights, sticks, string
Interpret observations as result of processes
Make Observations
Represent processes in an expressed runnable model
Make more Observations
Make predictions to challenge model
Moon must move CCW (as seen from N)
Improve model
Use model to infer system behavior at times and places where you have no observations
Compare behavior of model with behavior of earth captured in data
Ambiguity: which way does moon go?
Make predictions to challenge model
Northern Hemisphere
Southern Hemisphere
Southern Hemisphere
Would it possible for students to experience the entire knowledge-construction
Make lunar cycle using simple physical
calendar models? We think so….
From assorted balls, lights, sticks, string
Interpret observations as result of processes
Make Observations
Predict moon phases in Southern hemisphere
Make more Observations
Make predictions to challenge model
Moon must move CCW (as seen from N)
Represent processes in an expressed runnable model
Improve model
Use model to infer system behavior at times and places where you have no observations
Ambiguity: which way does moon go?
Compare behavior of model with behavior of earth captured in data
Compare S. Hemisphere data with predication
Take-home messages
• The way in which scientists use expressed runnable models to
create new knowledge (as opposed to demonstrating existing
knowledge) is obscure to most students, teachers, and the public.
• This is a problem, because expressed runnable models underlie
many of the most important and controversial advances in modern
science.
• It should be possible to build learning experiences that work
through the scientists’ model-using knowledge-creation process,
even using simple, transparent physical models.
Bridging the Gap between Tabletop Models and the Earth System
Questions?
Kim A. Kastens, EDC
Ann Rivet, Teachers College
Cheryl Lyons, Teachers College
Alison R. Miller, Teachers College
Bridging the Gap between Tabletop Models and the Earth System