|
Spring 2007 Course
- Advanced
Quantitative Methods—Maximum Likelihood (Graduate) January 2nd-January
11th
This
course is primarily concerned with regression analysis of political
and social phenomena via maximum-likelihood estimation. In many situations, Ordinary Least
Squares is unsuitable for analyzing data that are of interest to
political scientists. This
course will introduce students to a set of methods and techniques
for handling situations in which the assumptions of OLS are
violated. In particular, the
methods employed will be concerned with techniques for dealing with
categorical and limited dependent variables. Models to be covered include: logit, probit,
multinomial logit, ordered probit, and models for event counts. We will also consider models for
dealing with sample selection and event history. Although there will be emphasis on
the theoretical and technical aspects, most of our attention will be
directed to the application of these methods and the presentation of
the results.
The class will meet in the SPA
Computer Lab from 10am – 4pm on January 2nd-4th and
January 7th-11th. There will be a one hour break
for lunch each day. Assignments will be due during the course, but a
final paper will be due at the end of the Spring semester. A separate
class meeting will be scheduled in late April or early May for students
to present these papers.
The following books will be
required for this class. You may wish to purchase these in advance in
case the bookstore does not stock them before class begins on January 2nd.
Author : Long, J. Scott
Title : Regression Models for
Categorical and Limited Dependent Variables
Publisher : Thousand Oaks,
CA: Sage. 1997
Author : Freese,
Jeremy and J. Scott Long
Title : Regression Models for
Categorical Variables Using Stata
Publisher : College Station,
TX: Stata
Press. 2006
Author : Allison, Paul D.
Title : Event History Analysis:
Regression for Longitudinal Event Data
Publisher : Thousand Oaks,
CA: Sage. 1984
Author : Eliason, Scott R.
Title : Maximum Likelihood
Estimation: Logic and Practice
Publisher : Thousand Oaks,
CA: Sage. 1993
Syllabus
(From previous course…topics will remain the same, but scheduling will
obviously change).
Class Resources for Conduct I and
Conduct II
·
Stata Guides for Conduct II
o Bivariate Tests
o Rules for
Creating Tables
o Regression,
Recoding, and Labeling
o Putting
Regression Output into Tables
o Merging
and Appending Datasets
o Omitted
Variable Bias Illustration
o Durbin
Watson Test and Residual Plots
o
Creating Predicted
Values Using Clarify
|
