Project History
How it got started
The DOE project is a cooperative effort among three
different groups of people
representing the MISD, the MMSTC, and an industrial statistical
consultant. How did all these folks meet and manage to come
together to build the project? As with many such cases, the answer
involves both good intentions and good luck.
In 1984, Kathy and Bob Peterson participated in a special
Woodrow Wilson Foundation summer institute on Quantitative
Literacy in Princeton, NJ. A principal organizer and speaker at
that conference was the late William Hunter, a professor of
statistics at the University of Wisconsin. Prof. Hunter taught the
teachers about DOE, the Statistical
Design and Analysis Of Experiments. To reinforce the classroom
discussions, several applied statisticians working in industry
were invited to speak about the importance of statistical thinking
and practice in business and industry. One of the speakers was
Bert Gunter, then a member of the technical staff at the Sarnoff
Research Center of RCA Labs, which is right down the road from
Princeton. As a recent graduate of Wisconsin, he reinforced
Hunter's message about the practical value of DOE in industry.
However, there was only sufficient time at the Institute to pique
teachers' curiosity, but not enough for them to clearly learn how
it worked or how it could be taught in high school. So Kathy and
Bob left intrigued, but unable to proceed.
In 1986, after leaving RCA to start his own consulting
practice, one of Bert's first clients was a group at the Ford
Motor Company developing statistical training for Ford engineers.
As luck would have it, through Kathy's contacts in the Oakland
University statistics program, Kathy and Bob Peterson were also
working with the group (as training experts, of course), so Bert,
Kathy and Bob were thrown together purely by chance. Part of the
training under development included DOE. (most engineers are not
taught DOE in college, but Ford, like many other progressive
companies, wanted engineers to use the methods in their work).
Over the course of several years, Bert, Kathy and Bob discussed
the feasibility and importance of bringing DOE into the high
school curriculum. Despite the fact that it has been around for
more than 75 years and successfully applied in agriculture,
chemical processing, pharmaceutics and many other industries, very
few colleges teach it except in advanced statistics courses that
only a few statisticians take. It is also true that DOE represents
a very different approach to experimentation.
For example, standard, accepted scientific practice is:
- When one must experiment by varying several different factors
to see how they affect an outcome (also called a response)
of interest, vary only one factor at a time while holding the
others constant at a standard setting so as not to confuse
things. This is sometimes known as OFAT (One Factor
At a Time) experimentation.
- Experiment in systematic order to maximize convenience and
minimize experimental efforts.
By contrast, DOE says that one should:
- Vary all factors simultaneously to deal with interactions
and maximize efficiency (experimental precision).
- Experiment in random order to maximize accuracy and reduce
the chance of wrong conclusions.
In science and technology, as in all other endeavors, it is
difficult to change accepted practice, even when other methods are
shown to be better.
Nevertheless, they agreed that DOE presented many opportunities
for improving science and math teaching (see project
goals). Moreover, they also agreed that a carefully developed
and limited presentation could be used to teach many of the
important concepts and most useful methods of DOE without having
to deal with complicated statistical matters. As an added
incentive, many companies in the auto industry were training their
employees in DOE, so exposing high school students to the ideas
would make schools more responsive to real educational needs.
As a result, in 1992 Bob arranged for the MISD to have Bert
teach a set of five 3-hour workshops on DOE for local high school
teachers as part of its quantitative literacy and statistics
offerings. This was done in conjunction with Bert's consulting and
training duties at Ford to help defray travel costs.
These workshops included extensive hands-on activities and a
requirement that each participant conduct their own experiment
using DOE. Both Kala and Marie attended this seminar. Though quite
skeptical about these new methods that seemed to violate
well-accepted practice, Kala did a chemistry project of the kind
that she would teach to students and was amazed when the results
came out exactly as she knew they were supposed to. As a math
teacher, Marie naturally focused more on the possibilities for DOE
to reinforce and extend her curriculum with practical
applications. However, both were convinced that the possibilities
were worth exploring further, and invited Kathy, Bob and Bert to
work with them at the MMSTC both to teach teachers and to present
some initial classes on DOE to the students. This was how the
MMSTC efforts began.
Over the course of the last few years, Bert has given several
workshops at the MMSTC on DOE to both teachers and students. He
has also served in the capacity of an informal consultant to their
efforts. Part of this work has been funded by Eisenhower grants to
the MISD, and part has also been funded by a grant from the Ellis
R. Ott Foundation of New Jersey. Ott was a leading proponent of
the use of DOE in industry in the 1950's and 60's (and a professor
of statistics at Rutgers University thereafter).
With Dave Walsh's enthusiastic support, the MMSTC worked to
gradually integrate DOE throughout their math and science
curriculum. Ninth graders received graduated
"doses" in their math classes and used it in selected
biology, chemistry, environmental science, and physics classes.
Marie and Kala reported that the students had less trouble
learning and digesting the ideas than teachers, perhaps because
they had no previous paradigms to overturn! In any case, it is
encouraging to note that many students voluntarily choose to use
DOE as part of their senior science projects.
A difficulty that confronted this work is that that no suitable
instructional materials existed at the high school level. Bert
wrote a set of 25 modules to
serve as teacher training materials and source for classroom ideas
(click on the highlighted text to order). Since then, classroom
materials have been developed.
Please check out Info for the
latest information on available materials and activities.
We are also anxious for feedback and comments.
Send us email at:
-
Especially, contact Bob Peterson to be put on our mailing list
and become part of our DOE teachers network.
A Report on the Macomb, Michigan
DOE Project:
Linking High School Mathematics
and Science through Statistical Design of
Experiments
By Bob and Kathy Peterson
Introduction by Bert Gunter
To American Statistician readers and
members of the American Statistical Association:
This report briefly chronicles the history and
gives the current status of the Macomb County Michigan DOE
Project. It was written for the Board of Advisors of the project,
whose members are listed in the appendix. However, when I received
the report it was obvious to me that readers of this journal and
members of the American Statistical Association as a whole should
know of this remarkable work. To my knowledge, it is the only
successful effort to teach statistical design of experiments in
the pre-college curriculum. As such, it represents a radically
different vision of how statistical thinking should be integrated
into basic math and science education from the current
mathematics-oriented advanced placement training. I believe it is
a vital and exciting vision that we professional statisticians
need to think about.
This effort is largely the work of an
extraordinary husband/wife team: Bob Peterson, Mathematics
Education Consultant for the Macomb Intermediate School District (MISD)
and his wife Kathy Peterson, a former high school teacher, and for
the past ten years, a teacher and Department Coordinator
Administrator in the Math Department of the Detroit College of
Business. As the report mentions, they were inspired in 1984 by
the late Bill Hunter and others at a Woodrow Wilson Summer Program
for outstanding national high school math teachers for which Bob
had been selected.
Although Bob, whose undergraduate degree is in
chemistry, not math, was intrigued by these ideas back then (which
were so at odds with his traditional one-factor-at-a-time
training, for doing scientific experimentation), it wasn't until a
few years later in 1987 when the opportunity arose to really do
something. It was serendipity. At the time, I was an independent
statistical consultant working with Steve Zayac at the Ford Motor
Company to develop and teach statistical methods to Ford engineers
and technicians. As fate would have it, Steve had also contracted
with Bob to help develop this training. The thinking being that
many of the technicians targeted would have the equivalent of high
school backgrounds and so a person experienced in teaching such an
audience would be very helpful.
Of course, Steve was right, and out of our work
together, came the DOE effort. Partially supported by Ford and
borrowing time from my trips to Detroit, Bob managed to arrange
for me to give workshops to math and science teachers in 1990.
These workshops were practice oriented - all participants had to
do and report on their own experiment as part of the
"homework" - and required no math beyond basic algebra.
Still, it was pitched too abstractly (I had much to learn) for
most of the audience to connect with their teaching. Fortunately,
as the report mentions, two brave teachers were intrigued enough
to give it a try: Kala Smith, a chemistry teacher, and Marie
Copeland, a math teacher, both of whom worked at a Magnet
Math/Science Center (the Macomb Mathematics, Science and
Technology Center) in the Warren Consolidated School System, one
of the twenty one school districts in Macomb County serviced by
the MISD.
This was the opportunity we needed. We worked with
them over the next several years to give more workshops to
teachers and students and to develop DOE materials for their
science and math curricula. While this work has not achieved
complete acceptance, two pieces of "data" are worth
mentioning. First, from the outset, focus has been on integrating
the ideas within both math and science courses: basic algebra,
geometry, graphical data analysis, and statistical ideas (without
any formal inference) are covered in the math portion, while
design and performance concepts like good experimental planning,
choice of design factors, randomization, and replication are
covered in the science portion. All students are required to
design, perform, analyze, and report on their own experiments.
Second, it has become clear that the students find the DOE ideas -
especially factorial experiments - natural and easy-to-grasp;
indeed, after even the first basic exposure in 9h grade, most find
the notion of one-factor-a-time experiments limiting and
incomprehensible. It is the teachers who have the greatest
difficulty making the conceptual leaps.
In 1996, I gave up my consulting practice to join
Merck. Since that time, my contact with the project has been
limited, so that all the recent progress described here has been
entirely due to the hard work and inspiration of the Petersons. I
am thrilled by their success and am pleased to have been able to
play a part in it. I believe that there are many lessons that
those in the ASA interested in statistical thinking and
statistical education can learn from this work. I believe it is
time that the Petersons, and their many hard-working and
innovative colleagues receive the recognition they deserve for
what is, I believe, a unique and remarkable effort. I hope that
tangible support and real efforts to replicate their efforts
elsewhere will follow. They have developed practical experience
and concrete materials: it would be a shame if they were not more
widely disseminated and used. Indeed, I believe that their work
could serve as the core of a wider project. I hope this report
will stimulate efforts to this end.
The Report: Current Status of the Project
The inspiration for the program came from Dr. Bill
Hunter back in 1984. He lectured on factorial experiments to the
Woodrow Wilson Summer Program group which was instrumental in
launching the Quantitative Literacy Program. He said then that he
knew that the DOE could be taught to high school students at an
appropriate level.
Bert Gunter was also one of the speakers at the
Woodrow Wilson Program that summer. We were fortunate to continue
our association with Bert as he made frequent trips to Detroit
because of consulting work at Ford Motor Co. About five years ago,
by extending several of his frequent trips to Detroit, he was able
to conduct a DOE workshop for the teachers at a local math-science
magnet school, the Macomb Mathematics, Science and Technology
Center. Several teachers, such as Kala Smith and Marie Copeland,
really got excited about the material and began to introduce it to
their students. Kala has since retired, but Marie is still working
to get the material into the center's curriculum.
The project got a real boost when Bert wrote a 25
module manual for teachers. This did not provide classroom-ready
materials for students, but it did provide a useful, readable
resource manual for teachers, many of whom had no background in
statistics, or at most a long forgotten undergraduate course.
In the summer of 1998 Kathy Peterson
"retired" from her job as Mathematics Coordinator at
Detroit College of Business to work with the Utica Community
Schools Mathematics, Science and Technology Center. Ken Dupuis, a
math teacher at the center, wanted to introduce some statistics
into the curriculum, but wanted some help. Bob Peterson, in his
position as Mathematics Education Consultant for the Macomb
Intermediate School District, suggested a DOE course and offered
the support of the MISD.
A half-year foundation course for ninth graders
has been developed. This class has been named Data Analysis and
Research Methods, a title more understandable by the general
public than DOE. Kathy and Bob developed lessons. A retired chemistry teacher,
Larry Crossen, helped develop equipment kits. (The class is taught
in a regular classroom.) Bi-weekly study sessions were held in
which these four people went through the lessons just as kids
would. Details were hashed out, modifications were made, etc.
Then, Ken and Kathy team-taught the classes, and further revisions
of the lessons were done.
The class consists of two basic parts:
- Descriptive statistics, graphics, and comparing data sets.
The following series of activities in this first half of the
course are designed to stress the need for control,
randomization, and replication when collecting data.
- Summary statistics for central location and spread.
- Dotplots, stem-and-leaf plots, boxplots, time plots and
scatterplots.
- Comparing data sets and making decisions in the presence
of variability; studying relationships between two data sets
(correlation).
B. Experiments
- What is an experiment? How do you show cause and effect?
What is the difference between an experiment and an
observational study?
- Experiments with one independent variable are introduced.
The experiments are designed and run with control,
randomization, and replication in mind. Emphasis is on
graphical representation of the design space. Graphical
techniques are used to analye the adequacy of the proposed
model and predictions are made.
- Factorial experiments with two independent variables are
introduced. Again, control, randomization, and replication
are stressed. Acrylic cubes are used for a three-dimensional
representation of the experiment. interactions are explained
and studied interpretation and analysis are heavily
graphical. Predictions are made.
- Higher order factorial experiments are conducted. The
analysis moves from a graphical approach to an algorithmic
approach.
A more complete description is available at the
Website listed later in this report.
There were 63 students involved in last year's
program at Utica. Parents' Night was a very gratifying experience.
Many of the parents expressed their pleasure that their children
were being exposed to this information, and mentioned how
important it is in their professions.
In November of that year (1998), the Macomb
Mathematics, Science and Technology Center (Marie Copeland's
school), approached Bob and said that they also would like to be
involved in the project in a more formal way, using the materials
and methods being developed.
This fall (1999), a new magnet school opened to
serve the northern end of Macomb County. The Macomb Academy of the
Arts and Sciences has incorporated the DOE course into their
curriculum. Scott Weidner is teaching it as part of his ninth
grade chemistry class. In total, we have about 200 ninth graders
who are taking this course in Macomb County. The mathematics
director of the Grand Traverse Region Mathematics and Technology
Center has expressed a desire to incorporate the course in their
program, too.
The magnet schools, in addition to providing
programs for gifted motivated students, have a mission to be agents
of change for their school systems. The MMSTC and the Utica MST
Center teachers are engaged in talks with system administrators
about implementing the course in the regular high schools in their
districts. (These are huge districts, with about 30,000 and 20,000
students respectively.) To help in their efforts, we are planning
a two-week workshop next June, along with continuing support
throughout the school year. We are also developing a teacher's
guide to accompany the student lessons that have been developed.
Macomb County is a big county, with typical urban
traffic problems. We do a lot of traveling in our support of the
three centers. We hope to eliminate some of that travel soon, as
the MISD supports the introduction of interactive TV facilities
into the three centers. We hope that some of the teacher support
efforts can be done using this technology. We also hope to be able
to use the technology to give the students a chance to see
research in action. We would appreciate your help in this area. If
you know of organizations or businesses which would be willing to
spend some time talking to young people and explaining their work
to them, we would appreciate that information. In particular, if
you would like to share experiences with the children, we would be
thrilled to host such a meeting. (If you would like to communicate
with the teacher group, either by interactive TV or any other
means, we would welcome your input).
http://www.misd.net/mathematics
We can also be reached by e-mail at: kabob41@comcast.net
Update 2001
This year the course
“Data Analysis and Research Methods” was taught in several regular
high schools to non-college bound students.
The teachers found that the material was appropriate for the
students, and the student evaluations
indicated that they both enjoyed the hands-on nature of the class and
felt that the material was useful.
Training of additional
teachers who wish to implement the course at their high schools will
be conducted in the fall, for implementation in the
2002 school year.
Additional training at the Science Centers is taking place,
with teachers of upper level courses.
They will not be teaching the material (in the science center
the students study the course in 9th grade), but these
teachers will be in a better position to incorporate the DOE
techniques into their
advanced science courses.
A writing project will
take place in the summer of 2001.
The writing team will be composed of teachers who have already
taught the DOE course, Data Analysis and Research Methods.
The purpose will be to add additional exercises and strengthen
the assessment component of the course.
Appendix: Board of Advisors for the DOE Project:
for further
information on the Board of Advisors click
here.
Charles R. Allen, Mathematics Education Consultant, Michigan
Department of Education
Dr. Kent Voigt, Interim Assistant Superintendent, Macomb
lntermediate School District
Professor George Box, Professor Emeritus, University of
Wisconsin
Dr. A Blanton Godfrey, CEO of the Juran Institute
Bert Gunter, Merck Research Labs
Dr. Lynne B. Hare, Nabisco Research Center
Professor J. Stuart Hunter, Professor Emeritus, Princeton
University
Dr. Ronald Iman, Statistical Consultant
Dr. Manert Kennedy, Executive Director, Colorado Alliance for
Science
Dr. Gary McDonald General Motors R& D Center
Professor David S. Moore, Purdue University
Professor Richard Scheaffer, University of Florida
|
