HLP Lab Mini Course on Regression Methods

May 27 2008 - June 19 2008

Session0	May 27	Basics and R primer (attendance optional, reading required)
Session1	May 29	Linear regression
Session2	June 5	Issues in linear regression
Session 3	June 10	Multilevel linear regression
Session 4	June 12	Logistic regression, GLM
Session 5	June 17	Multilevel logistic regression, GLMM
Session 6	June 19	Computational methods for model fitting
Session 7	???	R wrap-up

Texts

Master copies of the texts are available in the HLP lab (Meliora 123).

Data Analysis Using Regression and Multilevel/Hierarchical Models by Gelman & Hill (2007). Online resources. G&H07.
Analyzing Linguistic Data: A Practical Introduction to Statistics using R by Harald Baayen (2008). hardback ($97) paperback ($35) Complete electronic draft. Baa08.
Introductory Statistics with R by Peter Dalgaard (2004). Online resources. Electronic copy through U of R libraries. Dal04.
Categorical Data Analysis by Alan Agresti (2002). Online resources. Agr02.

R packages

Design. Linear and generalized linear regression.
lme4. Multilevel modeling.
ARM. Companion package for Gelman & Hill (2007).
languageR. Companion package for Baayen (2008).

Datasets

attention-r-data.csv

How to read

One goal of this course is to make sure we're all comfortable with the same terminology and methods. Another goal is to make sure that as new people enter the community, we can bring them up to speed pretty quickly. To help with both of these goals, we're asking that you take some additional steps when you're doing the reading for this class.

Keep an eye out for redundancy. If multiple pieces of assigned reading cover the same topic, and you find a single one of the treatments to be superior and sufficient, please make a note describing the nature of the redundant content, which source you preferred, and why. This will help us develop a set of "canonical" readings on these topics.
Record and investigate unexplained or unclear terminology. Because we're cherry picking chapters from multiple sources, it's likely that at some point an author will use a term that was originally presented in some (unread by us) earlier section of the text. Alternatively, an author might just assume knowledge that we don't have. In any case, when you come across a term in the reading that you believe is not explained well enough, please make a note of the term and where you found it. Then, please go one step further. Do your best to find a simple definition of the term, and record it for others to use (Wikipedia and MathWorld are likely to be good resources for this, but also feel free to consult your favorite stats text books).