Download Abel Rodriguez Ph.D. candidate, Duke University Teaching

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts
no text concepts found
Transcript
Abel Rodriguez
Ph.D. candidate, Duke University
Teaching Statement
When I started teaching probability and statistics to undergraduates, first as a teaching assistant and
eventually as a visiting professor at Universidad Simon Bolivar in Venezuela, I was shocked to find how
little importance students attributed to statistics and how much prejudice they harbored. Even in a
school filled with “mathematically literate” engineers and computer scientists, most people would shrink
away from statistics courses, arguing that they were hard, boring and of little applicability in their fields. I
discovered that the most challenging element of teaching statistics was not the material per se, but
rather the students’ attitudes.
Through my struggles to improve the content of my classes and to effectively challenge students’
attitudes regarding the subject, I have come to believe that the methodology for teaching statistics is in
no small part responsible for fostering these negative attitudes.
Instructors often underestimate the role of examples and in-class problems in capturing and maintaining
the attention of students. I have learned to introduce new concepts through carefully chosen examples
that relate to the student’s major field. This helps me maintain the students interested and demonstrate
that probability and statistics can be as important to them as lab skills. I believe that the careful choice
of examples is important even for students from pure fields: they need to understand that probability and
statistics uses the language of mathematics, but is not only mathematics, and therefore, it is not enough
for them to learn how to prove results. They also need to learn how to analyze data and how to convey
those results through written or oral means.
I also believe that using simulations and games is an excellent way to demonstrate the concepts and
start building on the intuition necessary for our students to be able to apply statistical concepts in real
problems. There are plenty of on-line simulation resources, like the “Chance” website, the GNU book
Introduction to Probability by Charles Grinstead and Laurie Snell or James O. Berger’s website at Duke
University, that contain material that can be readily adapted for most undergraduate and master level
courses. This material is not only free and accessible to everyone, but allows the students to explore the
subject on their own, exciting their curiosity. Even graduate students can benefit from numerical
experimentation when learning more advanced material or in their own research. Having them
rediscover paradoxes as part of homework and projects is an excellent way to reinforce key theoretical
concepts.
However, the use of computers in class should not be limited to running pre-selected simulations and
games. Statistics is about data analysis, and current computers are powerful tools for data management,
computation and visualization. My undergraduate courses always include lab sessions where students use
computers to deal with real data and generate mini-reports that are an integral part of their evaluation.
Statistical methods are regularly abused or misused by the media, and even by specialized scientific
publications. Training our students to read critically, translate statistical results into plain English, and be
aware of ethically unacceptable situations is an important part of teaching statistics, both at the
undergraduate and the graduate level.
I typically require my undergraduate students to read the book “How to Lie with Statistics” by Darrel
Huff, and present a journal by the end of the semester where they comment on articles appearing in
newspapers or non-specialized magazines, following the style in the book. Another integral part of any of
my undergraduate classes is a small-group project where the students must apply the tools they have
learned during the course to a concrete problem of their choice. I require each group to meet with me at
least twice: one time to approve the topic, and a second one as a follow-up. By letting them choose the
topic I ensure that the project remains interesting and help them link statistical techniques with their own
applied fields. I have successfully applied this methodology in courses for industrial engineering, biology
and urbanism. One particularly interesting example came out of an urbanism course, where students
used data from a survey collected for a different class and generated a sort of joint project that was
evaluated jointly by the instructor from the urbanism department and me.
Teaching at the graduate level presents other challenges in this regard: with more tools at their disposal,
students need to understand the ethical implications of their actions. I think that involving my students
in discussion about topics like multiple comparisons, experimental design, frequentist vs. Bayesian or
objective vs. subjective methods, can be not only an excellent way to motivate a specific topic in a class,
but can be used to emphasize the decisions that they will have to face when participating in collaborative
research.
Teaching statistics is hard: On one hand, most people are used to thinking in deterministic rather
probabilistic terms and therefore underestimate the importance (or, in some cases, lack of importance) of
quantifying and understanding the variability in the data to understand the problem at hand. On the
other hand, frequentist reasoning underlying hypothesis testing procedures is widely used and taught,
but counterintuitive. Teaching statistics requires a certain degree of specialized abilities besides the
knowledge of the field and general teaching skills. As a argued above, it is not enough just to simply
teach in the same way we were taught. Innovative and creative thinking is indispensable if the discipline
is to take its role in promoting the advancement of knowledge.