Updated: 04/18/07, 10:0 pm
What is new: Nonparametric statistics (short description below, data by email).

Instructor

• Kiyoshi Igusa
• Goldsmith 305
• Office hours (subject to change): MWR 11-12, MW 3:30-4 and by appointment.
• Phone 63062.

Syllabus

• Textbook: Introduction to Mathematical Statistics and Its Applications (4th Edition) by Richard J. Larsen, Morris L. Marx
• I hope to cover the entire book starting with Chapter 5 (Estimation)
• There will be weekly math homework, 5 ``projects,'' 3 or 4 quizzes, no final exam. All assignments are due on the last day of class. See the detailed syllabus for more information.

Announcements

• Project 3 (find your own data) is due Wed, May 2.

Homework

1. (Due Wed 1/31/07) p.356 #5.2.12, p.363 #5.2.20, p.387 #5.4.2, 5.4.10 details and hints
2. (First version due Wednesday, 2/28/07) Project 1
3. (Due Wed 3/7/07) p.459 #6.4.6, p.462 #6.4.21, p.480 #7.3.7, p.486 #7.4.6, p.489 #7.4.16, p.507 #7.5.9, p.566 #9.2.6. details
4. HW3 (due Wed 3/14/07) 9.3.1, 9.3.3, 9.4.4: for this problem do it both ways (1st method is like Case 9.4.2 in the book or like the birthdate cut-off problem in class, the second method is like Case 10.5.1 in the book or like the children of pilots question in class) 10.3.8 (use the chart in problem 10.2.2) 10.3.12 10.4.3 10.5.6 details

Quizzes

• Practice Quiz 01 with answers, more comments.
• Review for Quiz 2 with answers, Part 1, answers, Part 2.
• Quiz 2 with answers
• Quiz 3 review A (what I handed out in class) with answers (revised)
• Quiz 3 review B: List of topics, more questions with answers

Notes

• Estimators
• slides on ANOVA
• Nonparametric statistics
  • Sign test. This is a one sample test for the median. You simply count the number of data points which are above the H0-value of the median and use the binomial distribution.
  • Wilcoxon sign rank test. This is also a one sample test for the median. You assign ranks according to the distance from the median then add up only the ranks of the positive signed numbers.
  • Wilcoxon Rank sum test. This is a two sample test. You assign ranks to all the numbers and add up the ranks of the smaller sample.