Political Science 706                                                                                                                                           R.A. Francisco

Spring, 2006                                                                                                                                                Tuesday, 1:00-3:50

 

Research Methods I

Introduction

 

The graduate degree is a research degree. This course seeks to provide the basic knowledge needed to conduct quantitative research. It covers an introduction of statistics and computer methods as they relate to statistical research in political science. Political Science 707 goes far deeper into statistics. This course is its foundation.

 

Office Hours

 

Conducted in 313 Blake on Tuesdays, Wednesdays and Fridays from 10:30-11:30 and on Fridays from 1:30 to 3:30. Telephone: 864-9023; e-mail: ronfran@ku.edu; web-site: http://lark.cc.ku.edu/~ronfran; data web site: http://lark.cc.ku.edu/~ronfran/data

 

Texts and Readings

 

The following texts are available on the internet. They will be read in their entirety.

 

Richard D. De Veaux and Paul F. Velleman. Intro Stats.  New York: Boston: Pearson/Addison-Wesley, 2003 or 2004.

Micheal S. Lewis-Beck. Applied Regression: An Introduction. Beverly Hills, CA: Sage Publications, 1980.

William D. Berry and Stanley Feldman. Multiple Regression in Practice. Beverly Hills, CA: Sage Publications, 1985.

 

We will also have a relatively small packet of chapters that do not appear in our texts. Other short required readings will be handed out in class. You will find two other categories of readings. Recommended readings will provide greater depth for specific topics. Applications are examples of the technique we cover under each topic in the political science literature.

 

A Calculator and Statistical Software

 

You will learn how to do statistical analysis with computers. However, it is also important to understand how the most important statistics are actually calculated. Exercises on sample data will comprise several assignments. Almost any calculator will be adequate, but you will find a scientific calculator (i.e., one that has square root, reciprocal, and exponent keys) most useful. A graphing calculator is not necessary. Many calculators will now calculate linear and even multiple regression coefficients, but you will not need these capabilities. They are widely available in spreadsheets (e.g., Excel [be careful, Excel makes mathematical errors], Open Office and Gnumeric) as well as in computer statististical software (SAS, SPSS, STATA, Minitab, and R). We will stress R in this course, an open-source free software that has a full capacity for all statistics and graphing.

 

Assignments and Grading:

 

The following weighting system will determine the final grade: 1) participation, 10%; 2) assignments, 20%; 3) midterm examination, 20%; 4) take-home examination, 25%; 5) research paper, 25%. The research paper should use research techniques covered in the seminar. It can be based on a paper you prepare for another class, but you should notify me in such a case. Late assignments are not accepted. Incompletes are given in exceptional circumstances.

 

Course Objectives: Understanding at a good level of depth the following concepts and skills at the end of the semester.

 

  • data, distributions, and data sources
  • relationships among and between data
  • how to create new data
  • the basic ideas of probability
  • matrices: how to add, multiply and take inverses of them
  • randomness, sampling and sampling distributions
  • statistical inference
  • computer statistical programs
  • how to conduct linear and multiple regression
  • inference for regression and regression pathologies
  • nonparametric statistical tests and when to use them
  • the fundamentals of logistical regression
  • how to choose the right method of analysis from the data context

 

24 January: Welcome, Introduction to Data, their Distributions and Descriptions

 

  • Overview of the course
  • Why we go fast, then slowly
  • Innumeracy and social science research
  • Means, medians and even geometric means
  • Statistical plotting and graphing
  • Normal distributions and linear transformations

 

Required:

 

De Veaux & Velleman, Chapters 1-5.

 

Recommended:

 

John L. Phillips, Jr. How to Think about Statistics. New York: W.H. Freeman, 1992.

 

Application:

 

John Mark Hansen. “Individuals, Institutions, and Public Preferences over Public Finance.” American Political Science Review 92:3 (September 1998): 513-531.

 

31 January: Looking at Data and Relationships among Variables

 

  • Variance and standard deviation
  • The importance of scatterplots
  • The Pearson correlation coefficent (r) and its properties
  • Linear regression and ordinary least squares
  • Residual plots
  • Specification problems
  • Power & logarithmic transformations
  • Power laws and their uses

 

Required:

 

De Veaux & Velleman, Chapters 6-8.

Lewis-Beck, pp. 1-25.

Recommended:

 

Merran Evans, Nichols Hastings, and Brian Peacock. Statistical Distributions. New York: John Wiley & Sons, 1993.

Edward R. Tufte. Data Analysis for Politics and Policy. Englewood-Cliffs, NJ: Prentice-Hall, 1974.

Applications:

 

James D. Fearon, “Counterfactuals and Hypothesis Testing in Political Science.” World Politics 43:2 (January 1991):169-195.

J. Eric Oliver and Janelle Wong. “Intergroup Prejudice in Multiethnic Settings.” American Journal of Political Science 47:4 (October 2003): 567-582.

 

7 February: Producing and Using Data; Sampling

 

  • Samples vs. censuses
  • Where to find data
  • Content analysis, automated and otherwise
  • Experiments and experimental design in political science
  • Bias, randomization, and sampling designs
  • Survey methods

 

Assignment:

 

Do problem 17 (p. 178) in De Veaux & Velleman.  Do (1) a linear regession with killed or injured as the dependent variable (y) and assault rate as the independent variable (x) and (2) report the correlation coefficient between the two variables.

 

Required:

 

De Veaux & Velleman, Chapters 12-13.

 

14 February: Probability and the Study of Randomness

 

  • Basic probability, independence and the addition and multiplication rules
  • Random variables: discrete and continuous
  • Means and variance redux
  • The Law of Large Numbers
  • Rules for variances
  • General probability rules and conditional probability
  • An introduction to Bayesian statistics: the world of priors & posteriors

 

Required:

 

De Veaux & Velleman, Chapters 14, 15 & 16.

 

Recommended:

 

John P. Hoyt. A Brief Introduction to Probability Theory. Scranton, PA: International Textbook Co., 1967.

Richard von Mises. Probability, Statistics and Truth. New York: Dover,  1981.

 

Application:

 

Edward N. Muller, Henry A. Dietz, and Steven E. Finkel, "Discontent and the Expected Utility of Rebellion: The Case of Peru," American Political Science Review 85:4 (December 1991): 1261-1282.

 

 

 

 

21 February: Sampling Distributions

 

  • Sampling distribution for count data and proportional data
  • The binomial distribution
  • Binomial means and standard deviations
  • Means and standard deviations for sample proportions
  • Sampling distribution of a sample mean
  • The Central Limit Theorem
  • Event history and Weibull distributions

 

Required:

 

De Veaux & Velleman, Chapters 17 & 18.

 

28 February: Introduction to Statistical Inference and Significance

 

  • The concept of confidence and calculation of confidence intervals
  • Tests of statistical significance
  • The Null hypothesis
  • P-values and the probability of statistical significant variables
  • One- and two-sided tests of significance
  • Levels of significance
  • Abuse of significance
  • Power and inference; type I and type II inference errors

 

Required:

 

De Veaux and Velleman, Chapter 21.

 

Recommended:

 

E.S. Keeping. Introduction to Statistical Inference. New York: Dover, 1995.

 

7 March: Inference for Population Means; t-Distribution and Sample Tests

 

  • Inference for population means and t-distributions
  • One-sample t-tests
  • Paired two-sample t-tests as tests of hypotheses
  • The two-sample z statistic
  • Inference for small samples and the notion of resampling
  • The degrees of freedom (statistical, that is)
  • Pooled two-sample t-tests
  • F statistics and F distributions

 

Required:

 

De Veaux & Velleman, Chapters 23, 24 & 25.

 

 

Assignment:

 

Find two data sets with different sample sizes. Calculate their means, variance and standard deviations. Include the data in your report.

 

Applications

 

Ronald A. Francisco. “Coercion and Protest: An Empirical Test in Two Democratic States.” American Journal of Political Science 40:4 (November 1996): 1179-1204.

William F.S. Miles and David A. Rochefort, "Nationalism Versus Ethnic Identity in Sub-Saharan Africa," American Political Science Review 85:2 (June 1991): 393-403.

M. Stephen Weatherford, "Measuring Political Legitimacy," American Political Science Review 86:1 (March 1992): 149-166.

 

14 March: Inference for Proportions; Adding Scalars & Matrices; Regression

 

  • Inference, confidence intervals and significance tests for a single proportion
  • Inference, confidence intervals and signficance tests for two proportions
  • Introduction to linear algebra: scalars, vectors and matrices: introduction to the addition operation
  • More practice and work with regression

 

Required:

 

De Veaux & Velleman, Chapters 19 & 20.

Linear algebra handout 1

Berry & Feldman, pp. 9-37.

Crawley, Statistics: An Introduction using R, Chapter 8 (in packet)

 

Assignment:

 

Do problem 11 (p. 490) in De Veaux and Velleman and calculate a paired two-sample t-test. Is there a significant difference between the before and after?

 

Do this assignment in R. I do not encourage anyone to do it with a calculator.

 

Applications:

 

Controversy: Paul R. Abramson and Charles W. Ostrom, Jr. "Question Wording and Macropartisanship." American Political Science Review 86:2 (June 1992): 475-486.

 

21 March: Spring Break, no class

 

28 March: Cross-Tabs and Contingency Tables; Multiplying Matrices

 

  • Data analysis for two-way contingency tables
  • Inference to two-way contingency tables
  • Introduction to the Chi-Square distribution and test
  • Multiplying scalars and matrices

 

Required:

 

De Veaux & Velleman, Chapters  22 & 26.

Linear algebra handout 2

 

Midterm Examination

 

 

 

 

Applications

 

John M. Bruce, John A. Clark, and John H. Kessel, "Advocacy Politics in Presidential Primaries," American Political Science Review 85:4 (December 1991): 1089-1105.

 

4 April: Inference for Regression; BLUE; F-tests for Regression

 

Simple linear regression, residuals, and BLUE

F-tests for regression

Regression analysis in alternative statistical software

 

Required:

 

De Veaux & Velleman, Chapters 9, 10 & 27.

Lewis-Beck, pp. 26-47.

 

Applications:

 

Enrique A. Baloyra, "Criticism, Cynicism, and Political Evaluation: A Venezuelan Example," American Political Science Review 73:4 (December 1979): 987-1002.

Martin Gilens, "Gender and Support for Reagan: A Comprehensive Model of Presidential Approval," American Journal of Political Science 32:1 (February 1988): 19-49.

Alain Noël and J.P. Thérien, “From Domestic to International Justice.” International Organization 49:3 (Summer 1995): 523-553.

 

11 April: Inference for Multiple Regression; Pathologies of Multiple Regression; Time-Series

 

  • Inference for multiple regression
  • Pathologies: multicollinearity, autocorrelation, heteroscedasticity and other data problems
  • Introduction to vital considerations for time-series analysis

 

Required:

 

Lewsi-Beck, pp. 47-74.

Berry & Feldman, pp. 37-72

 

Assignment:

 

Find a data set with two potentially related variables. Do a linear regression. Show the scatterplot, the intercept and the parameter as well as the residual plot and the signficance test and its probability.

 

Recommended:

 

Christian Gourieroux and Alain Monfort. Time Series and Dynamic Models. New York: Cambridge University Press, 1997.

Richard F. Gunst and Robert L. Mason. Regression Analysis and its Application: A Data-Oreinted Approach. New York, Marcel Dekker, 1980.

James D. Hamilton. Time Series Analysis. Princeton, NJ: Princeton University Press, 1994.

Gary King, "How Not to Lie With Statistics." American Journal of Political Science 30:3 (August 1986): 666-687.

 

18 April: Sociological Statistics: One-Way Analysis of Variance; identity matrices and inverses

 

  • Inference for one-way analysis of variance
  • The two-sample t-test redux
  • What is an identity matrix?
  • How to find the inverse of a square matrix; hip matrices do not have inverses
  • We marry a matrix to its inverse and out comes an identity matrix!
  • The Jacobian matrix as a treasure trove of information

 

Required:

 

De Veaux & Velleman, Chapter 28 (on CD)

Linear algebra handout 3

 

Recommended:

 

M.J.R. Healy. Matrices for Statistics. New York: Oxford University Press, 2000.

Irving Reiner. Introduction to Matrix Theory and Linear Algebra. New York: Holt, Rinehart & Winston, 1971.

 

Application:

 

Benjamin Radcliff, "The Welfare State, Turnout, and the Economy: A Comparative Analysis." American Political Science Review 86:2 (June 1992): 444-454.

 

25 April: Two-Way Analysis of Variance; More Multiple Regression

 

  • The two-way ANOVA
  • Inference for two-way ANOVA
  • Heteroscedasticity and autocorrelation redux

 

Required:

 

De Veaux & Velleman, Chapter 29 (on CD).

Berry & Feldman, pp. 73-89.

 

Applications

 

Markus M.L. Crepaz, "Corporatism in Decline?" Comparative Political Studies 25:2 (July 1992): 139-168.

Dennis M. Simon, Charles W. Ostrom, Jr., and Robin F. Marra, "The President, Referendum Elections in the United States," American Political Science Review 85:4 (December 1991): 1177-1192.

 

2 May: Nonparametric Statistics

 

  • What is an nonparametric statistic?
  • Wilcoxon rank sum test
  • Wicoxon signed rank test
  • Runs tests
  • Kruskal-Wallis Test

 

Required:

 

Moore & McCabe, Introduction the Practice of Statistics, Chapter 14 (in packet).

 

Recommended:

 

Charles A. Lave and James G. March, An Introduction to Models in the Social Sciences New York: Harper and Row, 1975.

 

 

 

Applications

 

Viktor J. Vanberg and Roger D. Congelton, "Rationality, Morality, and Exit," American Political Science Review 86:2 (June 1992): 418-431.

 

9 May: An Introduction to Logistic Regression

 

  • Why logistic regression?
  • The logistic regression model
  • Inference for logistic regression
  • Alternative tests for non-interval data

 

Required:

 

Moore and McCabe, Introduction to the Practice of Statistics, Chapter 15 (in packet)

 

Application:

 

Althaus, Scott L. “Information Effects in Collective Preference.” American Political Science Review 92:3 (September 2003): 545-558.

 

Assignment:

 

Find two categorical variables and do a runs test for them.

 

Recommended:

 

Gary King, Unifying Political Methodology. New York: Cambridge University Press, 1989.

Bruce Western, “Concepts and Suggestions for Robust Regression Analyis,” American Journal of Political Science 39 (August 1995): 786-817.

 

8 May: Research papers due

 

14 May: Final examinations due