Stat 6337 Advanced Statistical Methods. Part I

MW 530 pm - 645 pm in CBW 1.206

Fall 2008


Instructor: Michael Baron
Office: ECSN 3.912
Phone: 972-UTD-6874
Texts: (1) Applied Linear Statistical Models by Kutner, Nachtseim, Neter, and Li
5-th edition, McGraw-Hill, 2004 (required; also used for STAT 6338 in Spring 09)
(2) SAS and SPSS Program Solutions for Use with Applied Linear Statistical Models
by Johnson and Replogle (not required)

Data sets and codes

  • Pulse data that is used by SAS program simplestat.sas. If there is any problem connecting with the data set, we can always use this other method.
  • Presidents data for Homework 3
  • GPA data for homework 4
  • Restaurant data for Project 2 - ASCII file. Restaurant data for Project 2 - SAS data file
    Reading the sas7bdat file is very simple: DATA restaurants; SET 'C:\your path\rest.sas7bdat';
    To see what's in the file, you can either print it or use PROC CONTENTS DATA=restaurants;
  • Moisture and sweetness data for Homework 5
  • Homes data

    Related articles and data

    More examples of Simpson's paradox
    You can fit regression models to the US population and the world population and compute your prediction for years 2010, 2015, 2020.
    For your regression comments: Women may outsprint men by year 2156 - from BBC News

    SAS corner

  • SAS manual
  • Review of SAS literature

    Computers, Unix, connections, NetID

  • New users: Set a NetID and create an account
  • Computer labs
  • Computer support
  • Two ways to connect to Apache from your home computer:
    (1) There is a fast connection with only one text window. Download putty.exe from PuTTY, save it on your home computer, open, and choose SSH and Host Name = apache.utdallas.edu. Save this choices for future use.
    (2) Here is a slower connection, but you can have a graphical user interface with it. Install Xmanager from Xmanager 3.0, start "Xstart", choose any session name, Host = apache.utdallas.edu, Protocol = SSH, User Name = your NetID, your Unix password, and choose Execution Command = 4 (Solaris) on the little black arrow. When you get a white window, type "sas" or "sas &".
  • Detailed instructions on remote connection to Apache.
  • New to Unix? Not a problem. You will find many useful links here
  • The University of Texas distributor of SAS

    Course outline

    1. Basic statistical methods most often used in the analysis of data. Estimation, confidence intervals, hypothesis testing. Understanding and use of P-values.

    2. Linear regression and regression diagnostics. Analysis of variance. Goodness of fit. Analysis of residuals.

    3. Multivariate regression. Model selection techniques. Stepwise regression. Multicollinearity. Variance inflation factors.

    4. Nonlinear models. Logistic regresion.

    Throughout the semester, students will practice data analysis and implementation of the introduced methods using Statistical Analysis Software (SAS) and will gain experience of solving practical statistical problems involving realistic data sets.

    Prerequisites: Multivariable Calculus and STAT 5352 or STAT 6331
    Any questions/comments/suggestions? Write to mbaron@utdallas.edu