Introduction
Biostatistics (or biometry) is the application of
statistics to biological research. In this studio/seminar, we
will explore some common and not-so-common techniques used to
analyze biological data. Examples will be drawn principally from
data collected by students during independent projects or internships,
with additional examples from published studies.
How this course works
This course will be run as a ‘studio' seminar.
Each week we will focus on a particular topic or analytical technique.
While a small amount of class time will be devoted to introducing
the topic, the bulk of the class time will be dedicated to applying
that week's techniques to an appropriate dataset. Since most datasets
are not amenable to analysis by all techniques, it is likely that
only one or two datasets will be used during any one week. My
underlying philosophy of data analysis is reflected in the organization
of topics: first, look at your data; second, summarize your data;
only after those two steps are completed should you think about
analyzing your data statistically. In addition, this process does
not make sense in the absence of an hypothesis, which should be
stated clearly, with cognizance of all underlying assumptions.
When and where
does the class meet?
The class meets on
Friday afternoons, from 1:00-3:50 pm, in Clapp 422.
Required Textbook (available from the Odyssey Bookstore across the street):
Selvin, S. 1998. Modern applied biostatistical methods using S-Plus. Oxford University Press, New York.Optional Textbook (available from the Odyssey Bookstore across the street):
Krause, A., and M. Olson. 2000. The basics of S and S-Plus, 2nd edition. Springer-Verlag, New York
Selected readings (on electronic reserve):
Ellison, A. M. 1993. Exploratory data analysis and graphic display. Pages 14-45 in S. M. Scheiner and J. Gurevitch (editors). Design and analysis of ecological experiments. Chapman & Hall, New York, New York, USA.
Ottenbacher, K. J. 1996. The power of replications and the replications of power. The American Statistician 50: 271-275.
Peterman, R. M. 1990. The importance of reporting statistical power: the forest decline and acidic deposition example. Ecology 71: 2024-2027.
We'll start the class with an introduction to your datasets and to the S-Plus software environment. Your datasets provide real-world case-studies that can be analyzed with most of the commonly-used, biostatistical techniques. These case-studies will also introduce you to standard statistical notation. One good way of approaching this Selvin's textbook is to see if you can use your data in the same way in which his examples are presented. The four reserve articles will be used to support material not covered in Selvin.
Recommended readings (on reserve in the library):
Sokal, R. R., and F. J. Rohlf. 1995. Biometry. W. H. Freeman and Company, New York, New York, USA.
Venables, W. N., and B. D. Ripley. 1994. Modern applied statistics with S-Plus. Springer-Verlag, New York, USA.
The introductory chapters from Sokal & Rohlf provide a review of material with which you should already be familiar, having seen it in various science courses at Mount Holyoke. You should re-read them if you need a refresher on basic statistics, probability, P-values, or hypothesis testing. The book by Venables & Ripley is a good guide to efficiently using S-Plus, the statistical software we'll rely on in class. There's not much in Venables & Ripley that you can't get out of the on-line help files, but you may find the book easier to read.
Weekly Assignments
There are weekly assignments;
which are posted on this web site. Each assignment will be due
at the beginning of the following week's class. Extensions will
not be given.
Term Paper
Each of you will write a final 15-20 page paper in
which you present your analysis of your own dataset. This paper
should be written in standard scientific format: Abstract, Introduction,
Methods, Results, Discussion, Literature Cited. However, the emphasis
should be on the analytical techniques, not on the scientific
results. In all cases, you should present at least two alternative
methods of analysis that are appropriate for analyzing your data.
Examples of alternative methods are: Bayesian vs. frequentist;
parametric vs. distribution-free. There are many others. Where
appropriate (i.e., in all but the Bayesian case), statistical
power should be discussed explicitly. For those of you doing senior
independent or honors theses, this term paper may inform or complement
the analytical section of your senior paper, but is not expected
to substitute for it. The first half of your term paper is due
on March 16, 2001, and the second half, together with the revised
(if necessary) first half is due on May 10, 2001. Extensions
will not be given.
Grades
Your semester grade will be based on your weekly assignments
(40%), term-paper (50%) and class participation (10%).
Weekly schedule
of topics
Click here
for the 2001 schedule
|
|
|
|
|
|
|
|
||||