Course Syllabus
Spring 2002
Wednesday 4:10pm - 7:00pm
455 Baker Hall

Course web page:  http://www.msu.edu/~dejongc/cj907

Instructor:         Dr. Christina DeJong
Office:              528 Baker Hall
Office Hours:    Monday, Tuesday 3:00-4:00pm
                        or by appointment
Telephone:        432-1998
Email:             dejongc@msu.edu

Course web page:  http://www.msu.edu/~dejongc/cj907

Course Description:

This course will focus on the analysis of categorical dependent variables, commonly found in criminal justice research. We will also spend some time at the end of the semester on special types of regression models (i.e., survival models, hierarchical models). We will focus on analytical techniques, software applications, and policy relevance for categorical research.

The prerequisite for this course is successful completion of CJ906 (Advanced Quantitative Methods in Criminal Justice Data Research) or an equivalent course.


Your semester grade will consist of two technical papers and one review paper (each worth 20% of your grade), and one final paper and presentation (40% of your grade). The technical papers will each utilize a special technique learned in the course (binary, nominal, ordinal, and count models).  Students should be able to discuss their rationale for analysis, and any special technical issues surrounding the analysis.  The third paper will be a review of your peers' second technical paper.

The final paper will be presented to the class, and should focus on a categorical dependent variable and the correct methods for analysis. Relevant criminal justice implications should be discussed, as well as directions for future research. This paper is expected to be of publication quality, and students will be encouraged to submit these papers for publication.

Students must supply their own data sets for this course. Data are readily available on-line from the Inter-University Consortium for Political and Social Research (http://www.icpsr.umich.edu/).  Data from a Master's thesis may also be used.  I strongly recommend that you acquire a dataset before the first day of class.  Appropriate data sets will have either (1) a relevant continuous dependent variable that can be easily recoded into categories for analysis, or (2) a choice of several different dependent variables that can be used with different methods (ie, binary, multinomial, count, truncated or censored).  Option #2 is preferred, since analysis using count data is an integral part of the course, and those variables typically cannot be created from continuous variables.


Students may choose to use either LIMDEP or SAS to complete the analysis in their papers.  Unfortunately, SPSS is not able to analyze categorical dependent variables.

Textbooks & materials:

Aldrich, John H. (1984). Linear Probability, Logit, and Probit Models. Thousand Oaks, CA: Sage.
Breen, Richard (1996).  Regression Models: Censored, Sample Selected, or Truncated Data.  Thousand Oaks, CA: Sage.
DeMaris, Alfred (1992).  Logit Modeling: Practical Applications.  Thousand Oaks, CA: Sage.
Long, J. Scott (1997). Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks, CA: Sage.
Menard, Scott (1995).  Applied Logistic Regression Analysis.  Thousand Oaks, CA: Sage.

Assorted readings: These will be contained in a reading packet, location to be announced. (Articles marked with * are available on Proquest).  All other articles will be provided.

Logistic Regression, Logit & Probit
*Allison, Paul D. (1999).  Comparing logit and probit coefficients across groups.  Sociological Methods & Research 28(2):186.

Ordered Models
Chandek, Meghan Stroshine (1999).  Race, expectations and evaluations of police performance: An empirical assessment.  Policing 22(4):675.

Multinomial Models
*Perna, Laura W. (2001).  The relationship between family responsibilities and employment status among college and university faculty.  The Journal of Higher Education 72(5):584.

Count Models:
*Ousey, Graham C. (2001).  Young guns: Examining alternative explanations of juvenile firearm homicide rates.  Criminology 39(4):933.

*Welki, Andrew M. and Thomas J. Zlatoper (1999).  U.S. professional football game-day attendance.  Atlantic Economic Journal 27(3):285.

Event History Models
Chung, Ching-Fan, Peter Schmidt, and Ann D. Witte (1991).  Survival analysis: A survey.  Journal of Quantitative Criminology 7(1):59.

*DeJong, Christina (1997).  Specific deterrence and survival analysis: Integrating theoretical and empirical models of recidivism 35(4):561.

*Avakame, Edem (1998).  How different is violence in the home?  An examination of some correlates of stranger and intimate homicide. Criminology 36(3):601.

*Schwartz, Jennifer and Jeff Ackerman (2001).  In search of a dependent variable: Comment on Avakame, 1998.  Criminology 39(4):969.

(Other suggested readings listed below schedule of lectures).

Greene, William H. (1995). LIMDEP Version 7.0 User's Manual. Plainview, NY: Econometric Software.
This manual is not mandatory for the course, because much of it is available with the software and on-line at http://www.limdep.com.

Norusis, M. J. (2000). SPSS 10.0 Guide to Data Analysis. Englewood Cliffs, NJ: Prentice-Hall.
A self-directed guide to using SPSS version 8.0. Students who have never used SPSS before, or who need a helpful guide should buy this book. May not be helpful for students already familiar with SPSS.

Delwiche, L.D. and S.J. Slaughter (1998).  The Little SAS Book: A Primer, Second Edition.  SAS Publishing.

Schedule of Classes and Due Dates:

The chapters listed for each lecture should be read before attending that class. Readings are subject to change over the course of the semester.

I.  Introduction and Review
1/9/02 Linear and non-linear models
Review of linear regression models
Maximum Likelihood estimation

Read: Long, Chapter 1 & 2

II. Binary Outcome Models 
1/16/02 Linear probability and logistic regression models 

Read: Long, Chapter 3; LIMDEP, Chapter 21; Aldrich & Nelson

1/23/02 Logit and probit models

Read: Long, Chapter 4; DeMaris

1/30/02 Applications of binary choice models

Read: Menard, Allison, Spohn & Holleran (2001)

Paper #1 due (Binary Choice Models)

III. Models for ordinal variables
2/6/02 Ordered logit and ordered probit

Read: Long, Chapter 5; LIMDEP, Chapter 23

2/13/02 Applications & Issues for Ordered Models

Read: Chandek

IV. Models for nominal variables
2/20/02 Multinomial logit and probit 

Read: Long, Chapter 6; LIMDEP, Chapter 24

2/27/02 Applications & Issues for Multinomial Models 

Read: Perna

Paper #2 due (Ordered and multinomial models)

V. Count models, censored and truncated data
3/6/02 Spring break - class not in session
3/13/02 Tobit, Poisson and Negative binomial regression models 

Read: Long, Chapter 7 & 8; LIMDEP, Chapter 26 27; Breen

3/20/02 Applications & Issues for Count Models

Read: Ousey, Welki

VI. Overview of other special regression models
3/27/02 Class Canceled
4/03/02 Latent variable models, GLM, event history analysis, log-linear models 

Read: Long, Chapter 9; Chung et al., DeJong

Paper #3 due (Term Papers)

4/10/02 Introduction to Hierarchical Linear Modeling 
4/17/02 Applications of HLM & Introduction to the HLM program.

Read: Avakame, Schwartz & Ackerman

4/24/02 Review Paper Due

Student presentations begin

5/1/02 Student presentations (continued)

Additional suggested readings:

Grogger, J.T. and R.T. Carson (1991).  Models for truncated counts. Journal of Applied Econometrics 6:225.

Brame, R., R. Paternoster, P. Mazerolle and A. Piquero (1998).  Testing for the Equality of Maximum-Likelihood Regression Coefficients Between Two Independent Equations.  Journal of Quantitative Criminology 14:245.

Paternoster, R., R. Brame, P. Mazerolle and A. Piquero (1998).  Using the Correct Statistical Test for the Equality of Regression Coefficients. Criminology 36: 859.

Sherman, L. W., D. A. Smith, J. D. Schmidt and D. P. Rogan (1992).  Crime, Punishment, and Stake in Conformity: Legal and Informal Control of Domestic Violence.  American Sociological Review 57:680-690.