|
Size: 2553
Comment:
|
Size: 3346
Comment:
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 10: | Line 10: |
== Goal of this workshop == To provide sufficient background and a platform for a discussion of common issues in linear and logistic regression fitting, including applications of multilevel models. Some of the specific questions that you have sent me: * Data distributions and transformations to increase normality * How robust are LME's in the face of normality violations (short answer: relatively robust; there are worse violations) * Evaluation of multiple logistic regression model * also: "Precision measures" -- I assume this refers to tests of the quality of fit of a model? * decision criteria to include or exclude variables. Why researchers sometimes keep insignificant predictors in the model? * When to use mixed-effects models * Conceptual background |
HLP Lab Mini Course on Regression Methods
November 24 & 25, Copenhagen, Denmark
Goal of this workshop
To provide sufficient background and a platform for a discussion of common issues in linear and logistic regression fitting, including applications of multilevel models.
Some of the specific questions that you have sent me:
- Data distributions and transformations to increase normality
- How robust are LME's in the face of normality violations (short answer: relatively robust; there are worse violations)
- Evaluation of multiple logistic regression model
- also: "Precision measures" -- I assume this refers to tests of the quality of fit of a model?
- decision criteria to include or exclude variables. Why researchers sometimes keep insignificant predictors in the model?
- When to use mixed-effects models
- Conceptual background
[wiki:/Session0 The week before the tutorial] |
November 17 |
Prepare |
[wiki:/Session1 Session 1] |
November 24 |
Linear and logistic regression |
[wiki:/Session2 Session 2] |
November 25 |
multilevel models |
Readings
In the sessions, I refer to the following readings:
Obligatory:
Analyzing Linguistic Data: A Practical Introduction to Statistics using R by Harald Baayen (2008). [http://www.amazon.com/Analyzing-Linguistic-Data-Introduction-Statistics/dp/0521882591/ hardback ($97)] [http://www.amazon.com/Analyzing-Linguistic-Data-Introduction-Statistics/dp/0521709180/ paperback ($35)] [attachment:baayen_analyzing_08.pdf Complete electronic draft]. Baa08.
Recommended:
[http://www.amazon.com/Analysis-Regression-Multilevel-Hierarchical-Models/dp/0521867061/ Data Analysis Using Regression and Multilevel/Hierarchical Models] by Gelman & Hill (2007). [http://www.stat.columbia.edu/~gelman/arm/ Online resources]. G&H07.
Also useful:
[http://www.amazon.com/Introductory-Statistics-R-Peter-Dalgaard/dp/0387954759/ Introductory Statistics with R] by Peter Dalgaard (2004). [http://staff.pubhealth.ku.dk/~pd/ISwR.html Online resources]. [http://site.ebrary.com/lib/rochester/Doc?id=10047812 Electronic copy through U of R libraries]. Dal04.
[http://www.amazon.com/Categorical-Analysis-Wiley-Probability-Statistics/dp/0471360937/ Categorical Data Analysis] by Alan Agresti (2002). [http://www.stat.ufl.edu/~aa/cda/cda.html Online resources]. Agr02.
R packages
Please make sure you have the following R packages installed before you attend the workshop. Please also note that I will assume R version 2.7.1, and it's easier if you have that same version (or at least that recent a version).
[http://cran.r-project.org/web/packages/Design/index.html Design]. Linear and generalized linear regression.
[http://cran.r-project.org/web/packages/lme4/index.html lme4]. Multilevel modeling.
[http://cran.r-project.org/web/packages/arm/index.html ARM]. Companion package for Gelman & Hill (2007).
[http://cran.r-project.org/web/packages/languageR/index.html languageR]. Companion package for Baayen (2008).