WSU Engineering Stat 360 Suggested Projects Autumn 2000

Our syllabus called for each student to complete a project during the term (tentatively at least). These serve a very important function. I cannot see how anyone can learn to apply statistics to engineering practice purely through working problems. Projects that are defined only in the barest of form will provide students with some practice in defining an issue, laying out a course of research, hunting for data, deciding how to analyze data, and summarizing results. These are not meant to be difficult assignments. They should be very brief--four pages maximum each.

Suggested projects:

1. Computer Project #1 in the text Demonstrate the validity of the central limit theorem for samples drawn from the normal, uniform, and gamma distributions. Refer to problems 3.1, 3.2, and 3.3 in the textbook for details. You can generate the random samples using Excel, or you can write the entire simulation in BASIC or C.
2. Simulation Since I view simulation as very important, and being one of the focal points of this class, write brief papers on the generation of random numbers, monte carlo techniques, or case histories of simulations. Use the General Science Index or Reader's Guide to Periodical Literature as a resource to begin looking for articles on these subjects.
3. Approximating the Binomial Distribution Demonstrate that the binomial distribution is almost a normal distribution with mean value of np, as long as np is greater than 5. See the textbook and my polio example on the web page for details.
4. The Poisson Distribution Use Stirling's approximation to simplify computing Poisson probabilities for large values of expected rate. See my exposition on Stirling's approximation for the binomial distribution on the web site for an idea on how to proceed.
5. A Class Resource Assemble a class resource for WWW links to web sites with Statistical software, tutorials, or examples.
6. Wind/Temperature Data Examine data, which is available from me, to answer questions about system availability in a wind power generation project or the heating load on a building.
7. Epidemiology In epidemiology risks are often stated in terms of the odds ratio and its confidence intervals. These measures have recently caused lots of confusion in the general media. Please assemble a short paper on how odds ratios are calculated, what they mean, and how to calculate confidence intervals for them.
8. Software Review Many of you have been downloading evaluation software and using it for homework. Write a review that compares and contrasts two commercial software packages. Address each package's capabilities concerning graphics, ease of use, price, and help.
9. Analyze Lotto Is it true, for example, that the winning numbers are uniformly distributed? Can you verify the advertized odds of winning? Can you estimate, from the reported number of 3,4,and 5 matches, how many tickets were sold?
10. Analyze Quinto Are the card faces drawn uniformly? Can you verify the stated odds of winning?
11. Work Related Problems Use any problem you have found at work that is related in some way to statistics, regression, modeling or simulation.
12. Find data of your own Find some data that is statistical (i.e. it contains noise. Graph the data, discuss the significance of it, find population attributes like mean and standard deviation.
13. Probability plots One of the topics that we glossed over is that of probability plots and Q-Q plots. Generate random deviates from a normal distribution, a uniform distribution, or an exponential distribution. Plot these on a normal probability plot. Discuss your findings. Plot the normal deviates on a Student's t plot (ah! You'll have to think about this a bit). Discuss your results.
14. Analyze a magazine article that relies heavily on statistics for its argument and conclusions. Did the author fully explain how the data were collected? Did the author derive a valid conclusion? Are there questions regarding the data or how it was presented?
15. Compare and/or contrast two magazine or newspaper articles--one that does a very good job, statistically, and one that does a very poor job. Be prepared to defend your classification.
16. Other suggested projects in the book Most of these projects appear to be pretty contrived (i.e. "...go get data from a catapult!"). However, I have not checked all of them carefully. Maybe one or two will appeal to you.

I will add more suggestions to this page as I think of them. However, we are getting well toward the middle of the term. If students do not have a project or two of their own by Friday, October 10, 2000, then I will begin to assign them.