Methods IV: Advanced Quantitative Analysis - Matt Golder

1 Methods IV: Advanced Quantitative AnalysisCourse InformationCourse Number: POS 5747:01 Class Time: Monday : 113 Bellamy BuildingCourse Website: ~mrg217 Information for ProfessorName: Matt GolderE-mail: (preferred method of contact)Tel: 850-644-7302 Office: 540 Bellamy BuildingOffice Hours: TBAC ontact Information for Teaching AssistantName: Justin ConradE-mail: (preferred method of contact)Office Hours: TBAC ourse Objectives:This course is a continuation of Methods III and takes you beyond simple OLS models. We will focus onmaximum likelihood estimation and Advanced linear models. The central objective of this course is to learnhow to identify and correctly apply the statistical techniques appropriate to the social science question beingexamined by the analyst.

2 We will also devote considerable time to computing and conveying quantitiesof interest based on estimates of model parameters. Students will be required to calculate all quantitiesof interest manually (not using pen and paper!) before employing the automatic commands common tostatistical packages. By the end of the semester you should be quite adept at programming in STATA andusing LATEX; you will also obtain some experience :We begin the course by looking at the theory and mechanics of maximum likelihood estimation (MLE).We then turn to specific applications of MLE. We will begin the course by taking a detailed look at variousdiscrete choice models. Specifically, we will look at models to deal with binary, multichotomous, ordered,and count dependent variables.

3 We will then turn to various continuous and discrete time duration will involve looking at different non-parametric, parametric, and semi-parametric duration will then cover issues of selection bias and censored data. If time allows, we will also look at modelsdealing with longitudinal data and instrumental variables. Notes and homework assignments can be foundat the course webpage: ~mrg217 Requirements:Attendance:Attendance and participation is mandatory. There will also be a lab session for an hour eachweek run by Justin. Attendance at the lab session is voluntary but highly :There will be weekly homework assignments, which will be a mix of analytical and computer-based problems based on the techniques covered in class. These assignments will be made available onThursday and MUST be handed in by Wednesday at Assignments that are handed in after thisdeadline will not be graded.

4 I assume that you will work together on these assignments and I encouragethis. However, I recommend that you do not work in groups larger than three - I do not want the classas a whole to be doing the assignments together! Although you will be working in small groups, each ofyou should write up your own answer key (don t all use the same sentences etc.) and each of you shouldrun any computer code that is written. Some homeworks will ask you to critically evaluate how one of thetechniques that we have learned is used in some article of interest to you - this should be done individuallyand not as a :There will be a final paper in which you apply one or more of the techniques that you have learnedto a real social science question. This papermustbe written in LATEX.

5 It is hoped that you will go beyondthe Methods covered in the weekly homeworks for this final paper. It is expected that this paper will havethe potential to be published with additional work. I firmly believe that any paper that you write from thispoint on should be written with the goal of getting it published; in this regard, you may want to continueworking on this paper in the spring semester. Given the time constraints involved at this stage of yourgraduate careers, I am amenable to the possibility of you handing in a similar paper to one that you arewriting for another , you must first discuss this with me and get the permission of theother professor should also note that I will take into account the fact that you are writing a paperfor more than one class when grading it I expect it to be of a higher quality.

6 You should start thinkingabout a possible paper topic as early as possible in the semester and discuss this with :Towards the end of the semester, you will do a poster presentation of your final paper that will beviewed by the faculty and other graduate students. This postermustalso be done in LATEX. Dates and timeswill be arranged :Grading is based on the weekly assignments (60%), the final paper (35%), and the poster (5%).No incomplete grades will be given unless there is an agreement between the instructor and the student priorto the end of the course. The instructor retains the right to determine legitimate reasons for an Texts:Choose one of the following three books as your main reference book. Greene, William H. Analysis , 6th Edition. Prentice Hall.

7 Wooldridge, Jeffrey M. Analysis of Cross Section and Panel Data. MIT Press. Cameron, A. Colin & Pravin K. Trivedi. : Methods and Applications. NewYork: Cambridge University should buy all of the following books - they are all very good: Long, J. Scott. Models for Categorical and Limited Dependent Variables. SagePublications. Box-Steffensmeier, Janet and Bradford S. Jones. History Modeling: A Guide for SocialScientists. Cambridge University Press. Train, Kenneth E. Choice Methods with Simulation. New York: Cambridge : During the course, we will primarily be using STATA as our statistical package. There are variousbooks on STATA that you might find helpful. Hamilton, Lawrence C. with STATA: Updated for Version 10. STATA Press. Long, Scott & Jeremy Freese.

8 Models for Categorical Dependent VariablesUsing STATA. STATA Press. Cleves, Mario A., William W. Gould & Roberto G. Gutierrez. Introduction to SurvivalAnalysis Using STATA: Second Edition. STATA Press. Cameron, A. Colin & Pravin K. Trivedi. Using STATA. STATA Press. You can also find more information about STATA at: : During the course, you will be introduced to LATEX. I recommend that you download Ghostscript,GSView, Winedt, and MikTeX to your own computer. Instructions on how to do this are directions and download links for installing the MikTeXpackage on your computer. If you have the hard drive space, I suggest that you do a total installation; this will give you access to all available libraries and prevent you from having todownload missing packages at a later time when you need them.

9 Do not download the basic system at the top of the page; install the complete system a bit further ~ghost/ Download two and by clicking on their link and then selecting a download servernear you. Install ghostscript by double-clicking the Follow the instructions in theinstaller to complete installation; generally, clicking ok to all the dialog screens Install GSView by double-clicking the file. Follow the instructions in theinstaller to complete installation; generally, clicking ok to all the dialog screens Download the latest version (at themoment, ) onto your computer. Double-click on the installer file you havedownloaded to install the program. Once installed, you will need to register WinEdt using the Register dialog in the rightmost menu bar item at the top of the screen to avoid annoyingreminder dialogs.

10 WinEdt is relatively inexpensive to buy and you can purchase a code online computers in the Political Science data lab already have these programs. There are various bookson LATEX that you might find helpful. The Not So Short Introduction to LATEX2 . This is available online at: Mittelbach, Frank & Michel Goosens. LATEX Companion. Second Edition. Addison-Wesley. This book is available in the Political Science data lab. Lamport, Leslie. : A Documentation Preparation System User s Guide and Refer-ence book is available in the Political Science data lab. You can also find more information about LATEX at: Web Sites:Many other professors teach courses similar to this one - some have websites that you may find extremelyuseful. If you are interested, you may want to examine those of Nathaniel Beck (NYU), Chris Zorn (Penn-sylvania State University), Janet Box-Steffensmeier (OSU), Charles Franklin (Wisconsin), Gary King (Har-vard), Jeff Gill (UC Davis), and Simon Jackman (Stanford).