Term Paper for STA2101f11

Students in the graduate section of this course (STA2101: Applied Statistics I) are required to write a short term paper. The topic is regression for censored data. The length of the paper is five to ten pages typeset or word processed, plus an appendix of computer output. Ten pages (exclusive of the appendix and title page, if you have a title page) is the absolute maximum. The paper is due on the day of the final exam, or before.

Find a good textbook, and read enough of it to choose a method you like with an example that includes data. The example data set must have at least two independent variables. Do the example using SAS, R or both and reproduce the main numbers given in the textbook. You need at least two parameter estimates and at least one hypothesis test.

What you write:

• Give the complete reference for the text you used. If you don't know what this looks like, see the reference section in the textbook itself.
• Briefly explain what censored data are, and why it is important to have statistical methods for such data.
• Define any other terms you use. It would be preferable to write the definitions precisely but in your own words. If you feel you absolutely must quote from the text word for word, make it clear by putting the material in quotes, and give the page number. You need to use some judgement here. The terms in question are specific to the topic at hand, not general terms like "random variable" that are part of the common statistical vocabulary.
• State the model. Here it is okay to use the notation of the text, though you are free to change it if you wish. Give one or more page references so the model can be located in the text. This is required even if you change the notation.
• Briefly describe the methods of estimation and inference. For example if it's maximum likelihood, describe how standard methods are modified to incorporate censoring.
• Describe the data set used for the example.
• Give the numerical results you obtained. Refer to both the appendix and the textbook pages on which these numbers may be found.

The appendix has your computer input and output. If you use both SAS and R, you may have two appendices.

Important: You must do the work yourself. It is academic dishonesty to present someone else's work as your own, or to allow your work to be used for this purpose. This means that if you use the exact words from the textbook or any other source, you must provide a complete reference with the page number or (if it's a website), the URL. I see no reason to quote exactly unless it's really a great quote.

I plan to mark this paper very gently, because for many students it will be the first time they have had to do something like this. But if I detect copying without a reference to the original, I will give the paper a zero. If exactly the same words are used in two different papers (unless it's a properly cited quote), both papers will receive a zero. Similar considerations apply to the computer part. Do your own work.

No late papers will be accepted. They must be handed in by the time of the final exam, or the mark is zero. In order to put this off, you would need to fake an emergency serious enough so you would miss the final exam.

More comments:

• The text you use must be a Statistics text, not a computer manual. If SAS or R code for the example is given in the text you have chosen, find another example in the book, or find another textbook.
• Online material is okay if it is equivalent to a textbook. Just give the URL.
• Again, I plan to mark these papers very generously. Just do the work yourself and hand it in on time.