Stat 371, Fall 2012
Assignment 5
Instructor Dr. Schonlau, due Thursday November 19 at the beginning of class. I do not take late
assignments. A TA will collect the assignments at the back of the room at the beginning of class
and then leave. Feel free to
Chapter 2:Models linking
explanatory and response variates
Matthias Schonlau, Ph.D.
Stat 371
Statistics for Business I
Chapter Outline
Multiple Linear regression equation
Examples
The least squares estimator
R squared
When least squares estimator cannot b
Chapter 11 Nonresponse II
Matthias Schonlau
Stat 371
1
Outline
Response rates and bias
Nonresponse Bias
when everything is known
Correcting for Nonresponse bias:
Callbacks
Post stratification
2
Response Rates and bias
A response rate is a convenien
Chapter 11 - Nonresponse
Matthias Schonlau
Stat 371
1
RESPONSE RATES DEFINITIONS
2
3 different response rates
3 organizations did a phone survey using the
same questionnaire at roughly the same time
with the same population
They get the following respon
Stratified Random Sampling
Matthias Schonlau
Stat 371
1
Overview
Introduction
Inference for
Example: Water quality
Efficiency/Variance under stratified sampling
Optimal allocation
Various
Competing survey goals
Sample size determination
Stratification
Post stratification
Matthias Schonlau
Stat 371
1
Stratification: unfortunate
randomization outcome
conduct a SRS phone survey SRS
Afterwards we find by chance or due to response
bias the sample contained only 30% men.
Response bias: For example, in pho
Chapter 8
Simple probability samples
Part II
Matthias Schonlau
Stat 371
1
Outline
Estimation for SRS
Example: Number of provinces visited
Drawing a random sample
Estimation in stata for sample designs
Estimation of Means
Estimation of Totals
Estima
Chapter 8
Simple probability samples
Matthias Schonlau
Stat 371
1
Outline
Common sampling schemes
Simple random sampling (SRS)
How to draw a simple random sample
Example: Number of provinces
Sample size determination
2
Common Sampling schemes
Simple Rando
Chapter 7
Sample Survey Issues
Matthias Schonlau
Stat 371
Roughly following Sharon Lohrs book
Sampling: design and Analysis
1
Outline
women in love survey
Target population/ study population/ sample
Types of Convenience samples
Measurement problems
Sampli
Chapter 6
Model building
Matthias Schonlau
Stat 371
1
Outline
Model building strategies
Forward selection
Backward selection
Stepwise
All possible subsets
Criteria for All possible subsets:
Adjusted R squared
Mallows Cp
2
What do with an insignifi
Chapter 6
Model building
Matthias Schonlau
Stat 371
1
Outline
Criteria for All possible subsets:
Adjusted R squared
Mallows Cp
2
Criteria for model inclusion
To decide on a model, so far we have
considered whether variables are (jointly)
significant
Chapter 5
Assessing Model Fit
Matthias Schonlau
Stat 371
Outline
Assumptions
Diagnostic 1: Residuals vs fitted values
Diagnostic 2: Residuals vs each x-variable
Diagnostic 3: QQ plot
Diagnostic 4: Leverage vs sequence
Diagnostic 5: Studentized residuals v
Chapter 4
The analysis of variance
Stat 371
Matthias Schonlau
Outline
Review tests with one parameter
Tests with multiple parameters
Example 6 products
Example: all coefficients are jointly zero
Relationship between t and F tests
Tests with one param
Chapter 3: Making Inferences for
Regression Models
Stat371: Statistics for Business I
Matthias Schonlau, Ph.D.
Review
Chapter 2:
An estimate of beta for the linear regression
model
How much variation does this model explain?
Chapter 3:
Inference: Hyp
Chapter 3:Indicator Variables
Stat371: Statistics for Business I
Matthias Schonlau, Ph.D.
INDICATOR VARIABLES
x-variables need not be continuous
The x-variables in the regression so far have
been continuous or ordered categorical
Number of bathrooms
Y
Stat 371
Fall 2012
Statistics for Business I
Instructor: Matthias Schonlau
Email: schonlau@uwaterloo.ca
Phone: 519-888-4567 x31518
Office: M3, room 4111
Lectures:
Tuesday and Thursday 2:30-3:50 pm in MC1085
I will not use the tutorial slots on Fridays 3
Appendix: Introduction to STATA
Stata is a professional data analysis software. It is widespread in academia, particular in economics,
sociology and in the health sciences. It has a much larger following outside of academia.
1. Starting Stata
There are se
University of Waterloo
STAT 371
STATISTICS FOR BUSINESS I
Course Notes
Fall 2012
by
JOCK MACKAY
rjmackay@uwaterloo.ca
revised and updated by
STEFAN STEINER and MATTHAIS SCHONLAU
shsteiner@uwaterloo.ca
Statistics 371 R.J. MacKay, University of Waterloo 201
Review of Stat 231
Introduction to Statistics
Matthias Schonlau, Ph.D.
Stat 371
Statistics for Business I
What is statistics?
Statistics is about making decisions in the face
of uncertainty
In the simplest case , we are trying to
understand uncertainty