Stat 431, Fall 2007
Statistical Inference
People
Additional help
Visit the StatLab: location and schedule at
http://stat.wharton.upenn.edu/~sivana/statlab.html.
Lectures
109 Steinberg Hall - Dietrich Hall. Monday/Wednesday, 1:30
– 3:00 pm.
Course homepage
Refer to
http://stat.wharton.upenn.edu/~mtraskin/courses/stat431/fall07/index.html
(this page) for announcements, handouts, homework assignments and other
materials.
Course description
This course is about making decisions under uncertainty using
statistical methods. The topics include estimation, confidence
intervals, hypothesis testing, single and multiple linear regression,
one-way and two-way analysis of variance, variable selection, logistic
regression and categorical data analysis.
Interpretation of the results and analysis of assumptions is an
important part of the course.
Statistical computing package will be
extensively used to carry out the computations. However no special
emphasis will be made on the details of computations.
Prerequisites
Familiarity with basic probability theory is assumed. Stat 430 or
equivalent should provide sufficient background. Otherwise chapters 2,
3, 4 and 5 of the Devore's book will be a good
substitute together with a short review
of fundamental concepts used by the course, written by Prof. David
Freedman, University of California, Berkeley.
JMP version 7, available in the Wharton Computer Labs,
F75/F80 Jon M. Huntsman Hall (Wharton account required: see
http://accounts.wharton.upenn.edu). Individual copies are also available
for purchase at
http://estore.e-academy.com. A
six-month license costs $29.95 and twelve-month is $49.95.
J. L. Devore, Probability
and Statistics for Engineering and the Sciences, 7th ed.
Assignments
Grading
- Homework assignments: 20% (lowest score excluded)
- Midterm: 30%
- Final: 50%
Weekly assignments will be due at the beginning
of lecture each Monday. Problems involving computer calculations should
be worked using JMP. No extensions to
the due date will be given. However, the lowest homework assignment
score will be omitted from the final grade calculation. Unsubmitted work
counts as a zero score.
You may work with and help each other, however you must submit your own
solutions, with your own writeup, unless otherwise noted.
- Homework 1, due 1:30pm, Monday September 24: homework: pdf
- Homework 2, due 1:30pm, Monday October 1: homework: pdf
- Homework 3, due 1:30pm, Monday October 8: homework: pdf
- Homework 4, due 1:30pm, Wednesday October 17: homework: pdf
- Homework 5, due 1:30pm, Monday October 29: homework: pdf
- Homework 6, due 1:30pm, Monday November 5: homework: pdf
- Homework 7, due 1:30pm, Monday November 12: homework: pdf
- Homework 8, due 1:30pm, Monday November 19: homework: pdf
- Homework 9, due 1:30pm, Wednesday November 28: homework: pdf
- Homework 10, due 1:30pm, Wednesday December 5: homework: pdf
Exams
In-class midterm: Wednesday, October 24th, 1:30
– 3:00 pm, 109 Steinberg Hall - Dietrich Hall. No make-up.
- Practice midterm is available: pdf.
Final: Wednesday, December 12th, 12:00 – 2:00 pm, 351 SHDH.
Both exams are open book and open notes. Calculators may be used but
no laptops are allowed.
Lectures will closely follow the text which will be occasionally
supplemented with handouts on topics going beyond those covered in the
book. A review of basic probability theory and common distributions,
e.g. chapters 2, 3, 4 and 5 in Devore, might be
useful. Topics discussed in sections 3.4 (binomial distribution), 4.3
(normal distribution) and 5.4 (central limit theorem) are of special
interest.
| Lec# |
Date |
Topic |
Text |
| 01 |
Wed 05 Sep |
Introduction/overview |
|
| 02 |
Mon 10 Sep |
Normality; boxplots; QQ plots (probability plots) |
1.4, 4.6 |
| 03 |
Wed 12 Sep |
Confidence intervals: known variance; Large-sample CIs |
7.1, 7.2 |
| 04 |
Mon 17 Sep |
Confidence intervals for population proportion; CIs: unknown variance;
Non-normal population distribution |
7.2 – 7.4 |
| 05 |
Wed 19 Sep |
One-sample hypothesis testing |
8.1, 8.2 |
| 06 |
Mon 24 Sep |
One-sample hypothesis testing |
8.3 – 8.5 |
| 07 |
Wed 26 Sep |
Two-sample inference: testing and intervals |
9.1, 9.2 |
| 08 |
Mon 01 Oct |
Two-sample inference: testing and intervals |
9.3 – 9.5 |
| 09 |
Wed 03 Oct |
Single factor (one-way) ANOVA |
10.1, 10.2 |
| 10 |
Mon 08 Oct |
Single factor (one-way) ANOVA |
10.2, 10.3 |
| 11 |
Wed 10 Oct |
Two factor (two-way) ANOVA |
11.1 |
| |
Mon 15 Oct |
Fall break, no class |
|
| 12 |
Wed 17 Oct |
Two factor (two-way) ANOVA |
11.2 |
| 13 |
Mon 22 Oct |
Simple linear least-squares regression |
12.1, 12.2 |
| |
Wed 24 Oct |
Midterm |
|
| 14 |
Mon 29 Oct |
Simple linear least-squares regression; Correlation coefficient |
12.3 – 12.5 |
| 15 |
Wed 31 Oct |
Simple regression and variable transformation |
13.1, 13.2 |
| 16 |
Mon 05 Nov |
Simple nonlinear regression; polynomial regression |
13.3 |
| 17 |
Wed 07 Nov |
Multiple linear least-squares regression |
13.4 |
| 18 |
Mon 12 Nov |
Multiple linear least-squares regression |
13.4, 13.5 |
| 19 |
Wed 14 Nov |
Variable selection |
13.5 |
| 20 |
Mon 19 Nov |
Logistic regression |
13.2, slides |
| |
Wed 21 Nov |
More on logistic regression. Q&A. |
|
| 21 |
Mon 26 Nov |
Categorical data analysis: goodness of fit |
14.1, 14.2 |
| 22 |
Wed 28 Nov |
Categorical data analysis: goodness of fit |
14.2, 14.3 |
| 23 |
Mon 03 Dec |
Bootstrap |
|
| 24 |
Wed 05 Dec |
Distribution-free procedures |
15.1, 15.2 |
- [Sun, Dec 9] Office hours on Monday, December 10 will be held from 11:00
am instead of 10:30 am.
- [Fri, Dec 7] The final examination will be Wednesday,
December 12, 12:00 – 2:00 pm in 351 SHDH.
- [Wed, Dec 5]
- Practice final is available: pdf.
- I'll hold office hours on Friday, December 7, 11:00 am – 12:00 pm
and 1:30 pm – 3:00 pm.
- Office hours on Monday, December 10, will be held on a regular schedule.
- If you cannot come during time indicated above, send me an email.
- [Thu, Nov 22] Office hours on Monday, November 26 will be held
3:00 – 4:30 pm instead of 10:30 am – 12:00 pm.
- [Thu, Oct 18] For the week of Monday, October 22, Xu will hold his
office hours on Tuesday, October 23, 3:00–5:00pm instead of
Thursday, October 25. His office hours will be on a regular schedule
the following week.
- [Wed, Oct 17]
- Homework #5 has been posted: pdf.
- Practice midterm is available: pdf.
- [Mon, Oct 08]
- Homework #4 has been posted: pdf. Notice unusual due date: it is due at the lecture (1:30pm) on Wednesday October 17.
- Reminder: homework is due at the beginning of the lecture. If you
are late to class, do not try to turn it in immediately, wait until
the end of class and turn it in then, your homework will be graded
just as if it was turned in on time, though there will be a record
that it was late.
- [Mon, Sep 17] Homework #1 has been posted: pdf. It is due at the lecture (1:30pm) on Monday September 24.
- [Thu, Sep 13] The first homework assignment will be posted on Monday, September 17.