Lecture 13: Coding Schemes for Regression
Reading Assignment:
Muller and Fetterman, Chapter 12: Coding Schemes for Regression (Required)
Goals for the Next Two Weeks
1. Understand various coding schemes for ANOVA and relationships between ANOVA
and multi
Lecture 12: Selecting the Best Model
Reading Assignment:
Muller and Fetterman, Chapter 11: Selecting the Best Model (Required)
Selecting the best model, while perhaps one of the most common data analysis tasks, is
an exploratory activity that is usually
Lecture 11: Transformations
Reading Assignment:
Muller and Fetterman, Chapter 10: Transformations (Required)
Transformation of the response and/or predictor variables may correct violations of
homogeneity, linearity, and Gaussian distribution of errors (
Lecture 7: Correlations
Reading Assignment:
Muller and Fetterman, Chapter 6: Correlations (Required)
For two random variables X and Y , recall that the correlation is dened as
= Corr(X, Y ) =
where 1
Cov(X, Y )
Var(X)Var(Y )
,
1. We estimate the popula
Lecture 5: Multiple Regression: General Considerations
Reading Assignment:
Muller and Fetterman, Chapter 4: Multiple Regression (Required)
Why use more than one covariate in a model?
Why not t separate models for every covariate?
Omitting an important
Lecture 6: Testing Hypotheses in Multiple Regression
Reading Assignment:
Muller and Fetterman, Chapter 5: Testing Hypotheses in Multiple Regression
(Required)
After tting a model, one seeks to draw inferences about parameters. Correlations and
condence i
Lecture 3: General Linear Model: Estimation and Testing
Reading Assignment:
Muller and Fetterman Chapter 2: Statement of the Model, Estimation, and Testing
(Required)
We will consider the case in which we observe a single response and one or more
covaria
Lecture 19: Power & Sample Size Calculation
Reading Assignment:
Muller and Fetterman, Chapter 17: Understanding and Computing Power for the
GLM (Required)
UNC Biostatistics 663, Spring 2015
1
Motivation
One of the most common questions asked of a statist
Lecture 1: Introduction and Overview
Reading
Muller and Fetterman Chapter 1: Examples and Limits of the GLM
Linear models are used to study how a quantitative response variable depends on one or
more explanatory variables. The model is called linear beca
Lecture 16: ANCOVA and the Full Model
Reading Assignment:
Muller and Fetterman, Chapter 16: The Full Model in Every Cell (ANCOVA as a
Special Case) (Required)
Understanding this chapter allows you to see all of regression and ANOVA as special
cases of th
Lecture 17: Logistic Regression
Often, the response of interest in a scientic study is a binary variable, such as
DISEASED/NOT DISEASED or DEAD/ALIVE. In this case, what are the problems with
the linear regression model?
When studying linear regression, o
Lecture 14: One-Way ANOVA
Reading Assignment:
Muller and Fetterman, Chapter 13: One-Way ANOVA (Required)
We use analysis of variance (ANOVA) to answer questions like the following.
Do two or more groups differ in mean response?
Does the new drug reduce
Lecture 18a: Mixed effect model
Reading Assignment:
Muller and Fetterman, Chapter 15: Special Cases of Two-Way ANOVA and Random
Effects (Required)
Mixed effects = xed effects + random effects. Mixed effects models are an extension of
the GLM for correlat
Lecture 15: Two-Way ANOVA
Reading Assignment:
Muller and Fetterman, Chapter 14: Complete, Two-Way Factorial ANOVA
(Required)
In two-way analysis of variance (ANOVA), we wish to evaluate the importance of all
combinations of two categorical variables in p