Syllabus: Stat 514, Sp 2003
 

Web page: See the Stat514 web page for additional information. To get there, go to www.stat.psu.edu, then to courses, and finally to Stat 514.

Instructor: Tom Hettmansperger, 317 Thomas Bld, 5-2211, tph@stat.psu.edu,

Office Hours:  W 1:25-2:20.
 

TA: Fengjuan Xuan, 331B Thomas Bld, fxuan@stat.psu.edu, Office Hours:

 

Text: Introduction to Mathematical Statistics, 6th ed, by Hogg, McKean, Craig (HMC)

This text is in manuscript form and available from:

Kopese at 230 S. Allen St. (one block south of Beaver Ave on Allen St.)

The manuscript is in two parts: The first part, used in Stat 513, contains Chapters 1-4 and the second part contains Chapters 5-10.  The price for each part is roughly $23.

 

Optional text: Statistical Inference, 2nd ed, by Casella and Berger (CB)

Price: over $100 ( do not buy the 1st ed which was available at the university bookstore. I do, however, recommend that you purchase the 2nd ed.)

There will be a copy of the 1st edition on reserve in the Math Library (McAllister Building)

I plan to cover Chapters 4-8 and topics from 9 and 10 as time permits. Topics covered include limiting distributions (central limit theorem), maximum likelihood estimation and likelihood ratio tests, sufficiency and completeness, optimality.  Examples will include regression and analysis of variance.

 

Examinations: There will be one exam and a final. The exam will be a night exam and will be around midterm.

 

Homework: Homework will be assigned on a regular basis and is due the class period after it was assigned. It will be graded.  Solutions will be posted on the class webpage.

 

Grading: The exam will be worth 100 points, and the final will be worth 200 points. The homework will only be used in case you are on the border between two grades. (The borders can be large, however.)

 

Academic integrity: The academic integrity policy of the Eberly College of Science will apply to this course. See www.science.psu.edu/Integrity/index.html
 

Goals: We will continue to consider model building in Stat 514. The models will involve unknown parameters and the emphasis will be on estimation and testing hypotheses concerning these parameters. In addition, we will spend some time developing asymptotic methods, culminating in the central limit theorem.  We will consider inference in standard linear models (regression and designed experiments) and also in some other important models such as mixture models. Throughout, it will be important to determine the sampling distribution (at least approximately) of the statistic we are considering. To that end we use analytical methods (mathematics) and computational methods (bootstrap, simulations, em algorithms). I expect you to be able to develop your own programs in Minitab, S-plus, R, or whatever you are familiar with.