FALL, 2011 – COURSE INFORMATION – E10.2003
Instructor: Sharon L. Weinberg
Email: [email protected]
Phone/Fax: 212-998-2373/212-995-4832
Office: Kimball Hall, 246 Greene Street, 316 East (Elevator Button: 3E)
Office Hours: Tuesdays, 10 am to noon and by appointment
TA: Meghan McCormick
Email: [email protected]
Office: 726 Broadway, 5th Floor
Office Hours: Tuesdays 6:30-8:00pm and by appointment
Prerequisites: E10.2001-2002 or at least one semester of another introductory statistics course.
Website: The course uses Blackboard for posting lecture notes, handouts, readings, homework
assignments, and general information.
Class Meeting Time/Room: Tuesdays, 3:30 pm to 6:10 pm in 194 Mercer, Room 305.
Lab Section Meeting Times/Rooms: Attendance in lab is strongly recommended and encouraged.
Lab meets on Thursdays from 3:30 pm to 4:45 pm in Tisch Lab, LC19. The lab provides
additional SPSS demonstrations of what is discussed in class, and hands-on guidance for
homework assignments.
Course Goals: This course extends the material covered in E10.2001-2 by examining more deeply
multiple regression/correlation as a general and flexible system for analyzing data in the behavioral,
social, and health sciences. In addition to covering more advanced topics related to traditional multiple
regression/correlation, the course examines ANOVA and ANCOVA as special cases of the general linear
model. It also examines logistic regression, a situation in which the generalized linear model is
appropriate given that the dependent variable is not continuous and does not meet the assumption of
normally distributed errors. The software package, SPSS version 20, is used to give students hands-on
experience with topics covered. In so doing, the course provides skills and knowledge critical to those
graduate students whose research relies on the analysis of quantitative data.
Course Orientation: This course provides a conceptually oriented, nonmathematical approach to
learning applied statistics. It is not appropriate for students seeking to learn the mathematical theory
underlying the general or generalized linear model.
Text: The course lecture notes serve as the primary text for the course, although the following textbook is
highly useful and is required as an alternative, more in-depth source of information:
Cohen, P., Cohen, J., West, S.G., & Aiken, L.S. Applied Multiple Regression/Correlation Analysis for the
Behavioral Sciences, third edition, Lawrence Erlbaum Associates.
Lecture Notes Posted on Blackboard: Notes will be posted on our class website for each lecture under
Course Documents. You are advised to download these notes prior to each relevant lecture and bring a
hard copy of them to class to facilitate note taking. Along with lecture notes, all data sets to be accessed in
that lecture also will be posted under Course Documents so that you may review and replicate on your
own whatever analyses have been carried out in class.
Supplementary Readings: As posted on the Blackboard website.
Homework: Practicing what has been covered in class is essential to learning statistics. Homework will
be assigned, collected, and graded each week. All students are responsible for completing all homework
assignments on time and raising related questions in class.
Grading: 10% Class attendance and participation; 90% Ten or so weekly computer-based homework sets
Syllabus:
Month
September
October
November
FA2011 -- E10.2003 Syllabus -- Intermediate Quantitative Methods
Day/Lecture #
Topic
6 -- #1
Statistical Procedures: A Conceptual Map; Univariate &
Bivariate Statistics -- A Review
13 -- #2
Statistical Control: The Two-Predictor Case
20 -- #3
The k Predictor Case: Model Building Strategies:
Simultaneous, Hierarchical, and Stepwise Approaches;
Statistical Inference in Multiple Regression;
27 -- #4
Nonlinear Transformations and Regression Diagnostics -Checking and Addressing Underlying Assumptions;
Estimating Heteroscedasticity-Consistent Standard Errors
4 -- #5
Critiquing an Article that uses Hierarchical Regression
11
NO CLASS -- MINI FALL BREAK
18 -- #6
Interactions -- The Case of a Dichotomous and Quantitative
Variable; the Case of Two Quantitative Variables; Post Hoc
Probing of Interactions (using MODPROBE)
25 -- #7
Omitted Variables, and Simple & Multiple Mediation (using
INDIRECT)
1 -- #8
8 -- #9
15 -- #10
22 -- #11
29 -- #12
December
6 -- #13
13 -- #14
From Single Predictors to Sets of Predictors: Qualitative
Scales using Dummy Coding (Indicator Variables),
Quantitative Scales, Analytic Strategies, Proportion of VAF,
Tests of Inference
Power Analysis using GPOWER; Multiple Regression as a
General Linear Model -- ANOVA as a Special Case of GLM
Multiple Regression as a General Linear Model -ANCOVA as a Special Case of GLM; Lord's Paradox
Nonlinear Relationships: Transforming the DV and Using
Power Polynomials
Generalized Linear Equations -- Intro to Binary Logistic
Regression & Maximum Likelihood Estimation
Binary Logistic Regression with Multiple IVs and
Interactions
Characterizing Differences among Methods Covered;
Wrapping Up.