MA764XA - Summer 2005 - Regression Analysis

Examination 2 Study Guide

For the exam, I will be supplying copies of the Z, t and F tables from the text. If you wish to bring some tables from another book, please bring a copy of the table and not the book.

Exam 2 will cover Mendenhall chapters 1 through 10. You are responsible for only those sections which you had homework problems from.

Formulas you are responsible for from Mendenhall:

All the formulas from examination 1
VIF page 349
regression residual page 366
standardized and studentized residuals page 397-8
Cook's Distance page 405
Durbin Watson d page 415
Two and Three straight line models, pages 429 and 430
WSSE page 439
3 and 4 point moving average, page 470
exponential smoothing, page 473

Questions typical of those that might appear on the examination follow:

Review the typical questions from examination 1.
What is homoscedasticity?
What is heterocedasticity?
Describe forward selection.
Describe backward elimination.
What are the assumptions associated with LS regression analysis?
Describe multicollinearity in 25 words or less.
What are some ways we can detect multicollinearity in a regression model?
Why might we apply transformation to our data when using LS regression?
Speak to extrapolating using a LS regression model.
What is a residual? Why do we care?
What are some techniques we can use to examine residuals from a LS regression?
Why might we calculate a standardized residual?
Why might we calculate a studentized residual?
Why might we calculate a Dubin-Watson d statistic?
Under what circumstances might we consider using a piecewise linear regression?
Under what circumstances might we consider using a weighted linear regression?
Under what circumstances might we consider using logistic regression?
What is a time series model?
Calculate a _ point moving average for the following sequence of values.
Exponentially smooth the following sequence of values using w = _.

You should expect that I'll be provifing you with R outputs of data analysis on several data sets and will ask you to do calculations and answer questions relative to those R outputs.