Nancy Heckman
Information on Expectations
Process
The Stat 548 process consists of you writing a report on a paper suggested by a faculty member. You submit
the report to the faculty member but typically do not revise it. The faculty member assesses the report.
Some faculty members might ask you to make a presentation of the report. I like to meet with you to discuss
your report. Below I outline my process.
-
We have an initial meeting to discuss the paper and to outline a plan for your report, to make sure
that expectations are clear. We also set a timetable for the completion of the report and our
meeting to discuss the report. For this meeting, you should have read the paper. You can also see if
there is any code readily available, if that is the type of paper you have chosen.
- You write the report. Feel free to contact me if you have questions or concerns.
-
You submit the report to me. I read it and come up with questions related the paper and the report.
Here are some examples of questions.
- What is a hidden Markov model and why is it useful?
- I don't understand what you did in this part of your simulation study. Coulyou please explain?
[this means your report writing wasn't clear enough for me!]
- What were the biggest challenges in the computation?
- Your proposed modification to the method seems effective. However, it is somewhat ad hoc, being
a two-stageprocedure. Can you make this a one-stage procedure by coming up with a model and
writing down a likelihood?
- I provide you with my questions from part (3) and give you some time to think about them.
- We meet to chat about my questions from part 3. You don't need to hand anything in or revise the
paper, but you can come with whatever notes you want. You probably won't be able to answer all of
the questions, but I hope you have something intelligent to say about them. And I think we will have
fun discussing the questions and we will both learn some things.
Expectations
In the content of your report and in our discussion:
- Understand the material. Be able to say what the researchers have accomplished, how they have
accomplished it, and why this might be important. Typically, this requires understanding the
statistical model developed and the conclusions of data analyses and simulation studies. You may not
be able to do this for the entire paper. In that case, clearly state what you have focussed on.
- Do something "extra". This can be applying the methodology to a new data set or to simulations and
discussing what you have found, what are the possible limitations of the method. It's great if you
can also suggest how to improve the method. This can be filling in proofs and discussing the
importance of the assumptions, how some assumptions might be weakened. These are just a couple of
ideas, and what you do depends on the paper and your interests. It can also depend on the skills you
feel you need to develop.
In your Writing
- Please include a reference to the article in the introductory paragraph or put the full name and
year of publication in the title of your report.
- Your report should be self-contained. For instance, you should define your notation. Don't assume
that I have the paper sitting near me when I read your report.
- Use the notation of the article.
-
Make sure your equations, symbols, etc are correctly used and make sense. That is, if you define
f(t) there needs to be a t in the definition.
- Give proper references, with a bibliography, if appropriate. This includes referencing R libraries.
Return to Nancy's STAT 548 page
Return to STAT 548 page