Project suggestions for a Chance course Spring 1996
General remarks. You might like to go through recent issues of Chance News for current
events that interest you. These are kept on the Chance Web site at (http://www.geom.umn/locate/chance).
If you find an interesting topic you can search for articles on this topic using the search mechanism you will find there. You can also look through
"Chance Magazine" on reserve in Baker Library Library. There are many interesting
chance topics there that could get you started on a topic of interest to you.
You may find an interesting chance project connected with a course you have taken.
For example, many economic courses discuss the role of chance in investments, behavior
of the stock market -- is it random, why choose a portfolio of stocks instead of
a single stock etc? Don't take on too broad a topic. It is better to say something interesting
about a specific aspect of the problem. Try to make your own contribution by giving
your analysis, insights, and, when practical, your own experiments.
ETHICS:
What are the ethical problems involved in the data bases for DNA fingerprinting? What
are the ethical issues in the genome project? There are a number of interesting ethical
issues in medical trials.
STATISTICAL ISSUES IN EVIRONMENTAL HEALTH.
The spring l991 issue of Chance magazine has a good discussion of the general problem.
Find some particular issue that interests you -- for example, are you increasing
the chance of getting cancer by sitting in front of your computer? There was a long
series of articles in the New Yorker on this topic.
STREAKS IN SPORTS
You could try to get data for Dartmouth basketball games to analyze. Alternatively,
you could give a critical survey of the work done so far. This is the Zaleskas study
done at Harvard and the three others reported in Chance magazine.
DO RIGHT HANDERS LIVE LESS LONG THAN LEFT HANDERS?
Studies have claimed that right handers live nine years longer on average than left-handers.
Review these studies and say what you think about this.
SAT ISSUES
We will consider a number of issues relating to the SAT exams: predictability, gender
issues, use in admission etc. You could look into one of these topics in more depth
and write up your own views on the matter. You could use data already obtained or
probably get additional data from the registrar.
We will see that different studies give different results on how much coaching for
SAT exams helps. review these studies and give your impressions of the weakness and
the strength of the studies.
EXTRA SENSORY PERCEPTION(ESP)
Polls suggest that about fifty percent of the people believe in some form of extra
sensory perception. What has statistics to say about this? You could arrange to do
your own experimenting and analyze the
BASEBALL SALARIES
Are baseball players paid according to their performance abilities? Data relating
to answer this question is given in the baseball-salaries folder in the Chance folder
on Pubic File Server. This data was for the year 1986 and was the basis for a contest
to see who could do the best job of answering this question using modern graphical methods.
An important tip is to look at simple functions of the data. For example, for some
data such as salaries it often helps to take the logarithm of the salaries before
looking at correlation's. For more details about this problem read the file baseball_salaries_read_me
in the baseball-salaries folder.
WEATHER PREDICTIONS.
What does the weather reporter mean by "there is a forty percent chance of rain today?"
It would be interesting to interview local weather predictors or the regional weather
office to find out what they think about this and related weather prediction problems.
You might see if there is a statistical difference between the predictions of two
different local weather predictors.
How should the weather predictor be rewarded to make his predictions as informative
as possible? For example if we just reward him for the number of times he is right
he will probably make very bland predictions that can hardly be wrong but which are
not very useful.
STATISTICS IN ADVERTISING:
Many advertisements say that "studies show" etc. Collect some of these and comment
on how honest and well documented these claims are.
EARTHQUAKE PREDICTION:
How are earthquakes predicted and how reliable have the predictions been? The current
issue (November 1996) of Discover Magazine has an article on this topic.
OPINION POLLS.
Carry out and analyze the results of an opinion poll. Discuss some of the issues in
opinion polls. For example, ways in which they effect the political process.
Economist have entirely different ways of predicting the outcome of the elections
that makes very little use of public opinion. How does there prediction records compare
with pollsters?
HOW SAFE IS SAFE SEX?
The safety of condoms is often estimated from previous estimates of the reliability
of the condom for birth control. What do they mean when they say it is 85 percent
reliable in birth control?
GAMES
There are many interesting probability problems in board games such as risk where
outcomes depends upon chance. Choose your favorite game and see if you can identify
probability problems whose solutions might help you in the game. The same is true
for casino games. Obviously, you cannot give a complete analysis of blackjack but you can
study some particular options to see what the best strategy is.
Gambling is about to appear on the web. How will they handle the blackjack problem.
Will they do a truly random shuffle? If so is it still a favorable player for
the dealer? Will you be able to right an applet to implement the basic strategy for
you etc.?
THE ROLE OF CHANCE IN LITERATURE.
A famous and fascinating example is Dostoyevsky's "The gambler" but there are many
other examples.
GOOD AND BAD GRAPHICS DISPLAYS
The classic book on this subject is The Visual Display of Quantitative Information
by Edward R. Tufte. You could make your own collection of graphs from newspapers,
journals etc. and give your analysis of which are effective and which are not and
why. An interesting project would be to use the scanner to make a graphics program that would
provide good and bad examples and your comments on them taken from Tufte and your
own research.
THE ROLE OF CHANCE IN MUSIC.
Mozart composed a piece of music by rolling dice. It would be interesting to implement
this piece on the Macintosh and see how it sounds.
Modern composers use chance on a regular basis. How do they use it? Are there statistical
regularities in their works etc?
STATISTICS IN MEDICINE
It is hard to open any newspaper and not read about a new medical breakthrough that
usual involves a statistical study. Choose an area of interest to you and give a
critical discussion of the study.
QUALITY CONTROL--THE DEMING STORY.
The story of how Japan won the industrial war is a fascinating statistical story.
You can find it written up in Chance magazine. You could interview a local business
that uses quality control and write up there experience with it. The Dartmouth Press
would be a good possibility for this. Another interesting project would be to figure out
a nice way to do the famous Deming bead experiment and try it out.
DNA FINGERPRINTING IN THE COURTS.
There are many issues legal, statistical and ethical on the use of DNA fingerprinting
in the courts.
STEVEN JAY GOULD'S "FULL HOUSE"
There are a number of aspects of this book that could be analyzed more and critiqued.
In addition you could compare it with the other recent book "Climbing Mount Improbably"
that takes a quite different view of evolution.
LIBRARY SEARCHES
You will almost certainly want to use the impressive literature search capabilities
including Lexis-Nexis that you can do at Baker Library.
The Web is a gold mine of information. Many governement agencies have reports and
data on their web sites. Last year a student wanted to do a project on weather prediction
so she e-mailed a number of meteorogists who had home pages on the web. She asked them a few simple questions about weather prediction and got back a wonderful assortment
of answers. Another student wanted to see if people's choices for numbers on the
lottery were random or if certain numbers were preferred. He was able to get a
set of numbers to look at from a friendly lottery expert found on the web.