Quantitative Methods 2 (PhD)

Quantitative Methods 2 (PhD)
This is the second-semester PhD course in quantitative methods for political/social science and taken by first year PhD students in political science from Trinity College’s Department of Political Science, as well as students participating in the joint PhD methods progamme between TCD and University College, Dublin (UCD). This course is also titled Quantitative Political Analysis under the TCD-UCD Graduate Research Education Programme in the Quantitative Social Sciences.
Version: March 10, 2009Trinity College Dublin, Spring 2009Wednesdays 10:00-13:00, Room College Green 4
Detailed Schedule
1. Properties of Estimators.Code from examples in class here.Required Reading:
o Kennedy Ch 1-2, “Introduction” and “Criteria for Estimators”.
o Faraway Ch 1, “Introduction”.
o (Review) Verzani Ch 10, “Linear Regression”.
Recommended Reading:
o W&W Ch 7, “Point Estimation”.
Homework: Exercise 1.Answer key for Problem Set 1.
2. The Classical Linear Regression Model.Code from examples in class here.Required Reading:
o Kennedy Ch 3, “The Classical Linear Regression Model”.
o Faraway Ch 2, “Estimation”.
o William Greene. 2003. Econometric Analysis. 5th ed. Upper Saddle River, NJ: Prentice Hall. Ch 2, “The Classical Multiple Linear Regression Model”.
Recommended Reading:
o W&W Chs 11-13, “Fitting a Line”, “Simple Regression”, and “Multiple Regression”.
Homework: Assignment here, additional data file you will need for assignment here.
3. Inference, Intervals, and Hypothesis Testing.Code from examples in class here.Required Reading:
o Kennedy Ch 4, “Interval Estimation and Hypothesis Testing”
o Faraway Ch 3, “Inference”
Recommended Reading:
o W&W Ch 8, “Confidence Intervals” and Ch. 9 “Hypothesis Testing”
Homework:Assignment 3.
4. Diagnosing problems with the CLRM.Code from examples in class here.Required Reading:
o Kennedy Ch 6, “Wrong Regressors, Nonlinearities, and Parameter Inconsistency”
o Faraway Ch 4, “Diagnostics”
Recommended Reading:
o Berry, William D. and Stanley Feldman. 1993. “Multiple Regression in Practice.” In Regression Analysis, ed. Michael Lewis-Beck. London: Sage.
o Fox, John. 1993. “Regression Diagnostics.” In Regression Analysis, ed. Michael Lewis-Beck. London: Sage.
Homework:Assignment 4.
5. Problems with predictors.Code from examples in class here.Required Reading:
o Kennedy Ch 11, “Multicollinearity”.
o Faraway Ch 5, “Problems with the Predictors”.
Recommended Reading:
o Kennedy Ch. 5, “Specification”.
Homework:None – “homework vacation” week!
6. Problems with error assumptions.Code from examples in class here.Required Reading:
o Kennedy Ch 7, “Non-zero expected disturbance”; Ch 8, “Non-spherical disturbances”; Ch 10, “Simultaneous Equations”
o Faraway Ch. 6 “Problems with the Error”.
Recommended Reading:
o Hanushek, Eric A. and John E. Jackson. 1977. Statistical Methods for Social Scientists. San Diego: Academic Press. Ch 8, “Structural Equations.”
o Kenneth Benoit and Michael Marsh. 2008. “The Campaign Value of Incumbency: A New Solution to the Puzzle of Less Effective Incumbent Spending.” American Journal of Political Science 52(4, October): 874-890.
Homework: Week 6 assignment, and the dataset needed here.
7. Models for binary data: Logit and Probit.Code from examples in class here.Required Reading:
o Faraway, Julian. Extending the Linear Model with R. Ch 2, “Binomial data” .
o King, Gary. Unifying Political Methodology: The Likelihood Theory of Statistical Inference. Cambridge, England and New York: Cambridge University Press, 1989. Reprinted, Ann Arbor: University of Michigan Press, 1998. Chapter 3, Chapter 5.1-5.4.
Recommended Reading:
o Aldrich, John H. and Forrest D. Nelson. 1984. Linear Probability, Logit, and Probit Models. Beverly Hills: Sage.
o W&W Ch 18 “Maximum Likelihood Estimation”.
o Verzani Ch. 12, “Two extensions of the linear model”.
Homework: Week 7 assignment, and the Titanic and economic bills datasets.
8. Models for count data: Poisson and Negative Binomial.Code from examples in class here, and the Benoit (1996) file weede.dta is here.Required Reading:
o Faraway ELMWR Ch 3, “Count Regression”.
o King, Gary. Unifying Political Methodology, Chapter 5.7-5.10.
Recommended Reading:
o Kirkwood, Betty. 1988. Essentials of Medical Statistics. Oxford : Blackwell Scientific. Chapter 17.
Homework: See Week 9.
9. Estimating uncertainty in inferential models.Code from examples in class here, and the data file dailprobit.dta is here.Required Reading:
o King, Gary; Michael Tomz; and Jason Wittenberg. 2000. “Making the Most of Statistical Analyses: Improving Interpretation and Presentation.” American Journal of Political Science 44(2, April): 341-355.
o Faraway LMWR, Ch 3.5-3.6, “Confidence Intervals for ….”
Recommended Reading:
o The Zelig package for R – this will allow you to implement the “CLARIFY” techniques in King, Tomz and Wittenberg (2000)
o King, Gary. “Calculating Standard Errors of Predicted Values based on Nonlinear Functional Forms.” The Political Methodologist, Vol. 4, No. 2 (Fall, 1991).
o King, Gary. 1986. “How Not to Lie with Statistics: Avoiding Common Mistakes in Quantitative Political Science.” AJPS 30(3):666-687.
Homework: Week 9 assignment, plus AFFAIRS.dta and habeas.dta datasets.

Objectives and Learning Outcomes
This course extends the analytical and theoretical background developed in Quantitative Methods I. It focuses on building a greater understanding of the methods introduced in that course, such as the workings of multiple regression and problems that arise in applying it, as well as going deeper into the theory of inference underlying regression and most other statistical methods. Quantitative Methods II also covers new classes of models for binary and count data, emphasizing the need to t appropriate models to the underlying processes generating the data being explained.
This course is primarily about data analysis and developing a deeper understanding of the generalized linear model. The focus is on practice, and this focus is reflected in the choice of texts and in the emphasis on applied coursework. While this course deals to some degree with the generalized linear model on a mathematical and theoretical level, its main focus is practical, the ability to use the techniques when faced with the need in practical research. Consequently the learning method combines lectures and reading with hands-on statistical programming exercises using real datasets.
The learning outcomes associated with this nine-week course are aimed at students being able to:
Develop a deeper understanding of the linear regression model and its limitations;
Know how to diagnose and apply corrections to some problems with the generalized linear model found in real data;
Use and understand generalizations of the linear model to binary and count data;
Develop a greater familiarity with a range of techniques and methods through a diverse set of theoretical and applied readings;
Know where to go to learn more about the techniques in this class and those called for that were not covered in this class.
Prerequisites
Quantitative Methods I for Political Science or an equivalent course. A basic knowledge of mathematics, in particular algebra and simple calculus, is beneficial but not assumed. Also since the practice component is done in R, it is assumed that students have already used this program in Quantitative Methods I.
Logistics
Meetings. Classes will meet nine weeks for one sessions per week, on Wednesdays from 10-13:00. The class will be mostly lectures and presentations by me, with the rest devoted to practical data analysis relating to weekly problem sets. For this reason I encourage students to bring their laptop computers to class, although this is not an integral requirement. (Since electrical outlets are limited in the classroom, please have your batteries charged ahead of time!)
Computer Software. As in Quantitative Methods I, the statistical package will be used for all exercises.
Grading. Grading will be based on three components.
1. Problem sets: 50%. Problem sets will be handed out each Wednesday and must be submitted to the class page at http://turnitin.com before class the following Wednesday. Each problem set will consist of a number of problems combining computer analysis with interpretation and analytical problems. Computer output, when supplied, should include both the commands used as well as results. Computer results should be indicated clearly. You are encouraged to work in groups on the problem sets, although work should be submitted individually. If you have any problems that you wrote by hand, then you can use our department’s excellent scanner to convert them easily to pdf.
2. Replication project: 50%. This project will be quantative reanalysis of a published quantitative work. Your job will be to obtain the data from the original author (or obtain the same data she used for the original piece), replicate her analysis, and extend the analysis using a new model or new variables whenever possible. If done properly this replication may be suitable for publication, which should be your objective. This project will require you to begin searching immediately for an article to replicate, including contacting the author or taking equivalent steps to obtain the data for your replication. The article you replicate may be from any field in political science, but must be an empirical application using inferential statistics. The only other restriction is that you may not replicate any article assigned for class reading or exercises. Your project must be submitted with your own replication dataset, so that someone else could replicate your analysis. You must also include a copy of the article whose analysis you have replicated. The deadline for the replication project will be the same as for the paper deadline associated with the non-quantitative courses. Examples of published replications may be seen in the volumes 41 and 42 of the American Journal of Political Science, available for browsing through JSTOR.
3. Examination: 0%. There is neither an exam nor a traditional research paper for this course. The problem sets substitute for the exam and the replication project replaces the research paper.
Texts
This course will assign a variety of reading materials, some essential and some supplementary. Readings are absolutely central to this course and you will not learn anything if you attempt to rely on the lectures alone.
This year I have tried an approach a bit different from previous years: I have selected two main texts that we will go through chapter-by-chapter. The first of these, Kennedy, provides a three-level discussion of each topic: first a general discussion, then a technical discussion, and then a very technical discussion. Most students find this quite useful since it permits them to dig as deep as their abilities let them or as their need allows. The second text is a text about the linear model but very focused on demonstrating each lesson in R. The discussion and the demonstration in R is very tightly connected, and this will enhance students’ practical ability to implement the models while also improving their knowledge of R.
A text that is a very basic, very accessible but thorough introduction to statistics, written by statisticians, is Wonnacott and Wonnacott. I have listed chapters on the recommended lists for the first several weeks that will provide a very useful counterpoint to the applied, social science-oriented readings represented by the Kennedy and Faraway texts.
You may find some of the readings difficult or uncomfortable. This is completely normal. Your response should not be avoidance but rather a renewed effort to understand the material by (1) reading it with even greater care, (2) rereading it several times, (3) seeking other readings that might make the primary texts more comprehensible, and (4) working with other students in study groups. It is also perfectly normal in methods classes that you do not absorb all a text has to offer upon the first reading, but rather return to it several times over the years and learn new things as your knowledge accumulates.
There are two texts I would recommend that you purchase:
Kennedy, Peter. 1998. A Guide to Econometrics. 5th ed. Oxford: Blackwell. You can order this from Amazon.co.uk.
Julian J. Faraway. 2005. Linear Models with R. Boca Raton: Chapman & Hall.
The other texts which you should consider purchasing, but which I will make photocopies of, are:
Wonnacott, Thomas H. and Ronald J. Wonaccott. 1990. Introductory Statistics. 5th Ed. New York: Wiley.
Verzani, John. 2005. Using R for Introductory Statistics. Boca Raton: Chapman & Hall. Most students will already have a copy of this from Quantitative Methods I.
Other sources will be available on-line through the web page for this course, including the HTML version of this document. In the HTML syllabus, for instance, many of the articles may be viewed or downloaded from the web site. (Class exercises will also be on-line.)