Syllabus for Regression
Psychology
830:591

Professor Lee Jussim
Fridays, 10-12:30, Room 105, Busch Psychology Building


This is an intermediate course on regression.  You must have had at least
a year of graduate statistics, including multiple regression, to take it.
We will briefly review the basics, but quickly get into issues of
causality, problems or limitations of regression, regression and ANOVA,
mediation, moderation, and path analysis. 

READINGS SHOULD BE READ BY THE CLASS ON THE DATE INDICATED.
1/4-1/2 PAGE SUMMARIES ARE DUE AT THE START OF CLASS.

Most readings are on electronic reserve, under my name.  Longer articles
are sometimes separated into multiple links to multiple pdfs. These articles
are, unfortunately, listed alphabetically BY TITLE.  Sorry, not my idea;
this is how the RU Library, in its infinite wisdom, does it.

Also, articles/chapters with multiple authors often only have the first
author listed on electronic reserve.  So, my syllabus may say "et al"
and the reserves only list the first.  Again, don't know why the library does this.

The readings from Cohen et al, and from Pedhazur, are more for techniques,
demonstrations, and ideas.  Both have lots of examples, with lots of formulas
and computations.  Feel free to skim all that.  Please, however, read the idea
and concept sections very carefully.  "Idea and concept sections" are almost
every part that is not dense with computations and data.

Do, however, read through the sections on basic computations for fundamental
statistics, like b's, betas, r's, r-squareds, etc.

I. Intro and Overview

A. Basic Review
Week 1, 9/5/08
Correlation
Bivariate Regression
Regression with 2 predictors


B. Causality and Regression
Week 2, 9/12/08
  1. Cohen, et al, Multiple Regression/Correlation With Two or More Independent Variables
  2. http://www.socialresearchmethods.net/kb/positvsm.php
  3. http://www.socialresearchmethods.net/kb/introval.php
  4. http://256.com/gray/thoughts/2004/20040511.html
  5. http://www.jerrydallal.com/LHSP/cause.htm
  6. http://crab.rutgers.edu/~goertzel/econojunk.doc (open this document and read it).

C. bs, betas, partials, semi-partials & stepwise (hierarchical) regression, its evils, and a simple application
Week 3, 9/19/08
  1. Cohen et al, Data-Analytic Strategies Using Multiple Regression/Correlation
  2. Pedhazur, Statistical Control: Partial and Semipartial Correlation
  3. <>Madon et al, The Accuracy and Power of Sex, Social Class, and Ethnic Stereotypes: A Naturalistic Study in Person Perception
D.  Problems & Limitations
Week 4, 9/26/08:
  1. Dawes, The Superiority of Simple Alternatives to Regression for Social Science Predictions
  2. Cohen et al, Outliers and Multicollinearity: Diagnosing and Solving Regression Problems II

II. Regression & ANOVA
Week 5, 10/3/08
Cohen et al, Categorical or Nominal Independent Variables
Tricks to simplify ANOVAs done via regression (lecture).


III. Mediation & Path Analysis

A. Data Analysis
  1. Baron & Kenny, The Moderator-Mediator Variable Distinction in Social Psychological Research
  2. Cohen et al, Multiple Regression/Correlation and Causal Models (JUST UP TO PAGE 467).
B. Path Models as Theory
(and a return to experiments versus correlational research as a basis for causal inferences)
  1. Jussim, Social Perception and Social Reality: A Reflection-Construction Model
  2. Jussim, Teacher Expectations and Self Fulfilling Prophecies: Knowns and Unknowns, Resolved and Unresolved Controversies

IV. Moderation
A. Classic Moderation
Cohen et al, Interactions Among Continuous Variables

B. Curvilinear Relations
Cohen et al, Quantitative Scales, Curvilinear Relationships, and Transformations (193-221)

C. Quadratic Interactions
Madon et al, In Search of the Powerful Self-Fulfilling Prophecy