Levin College of Urban Affairs Instructor: Dr. William M. Bowen
Quantitative Research Methods, UST803 Office: UB107, 687-9226
Spring Semester, 1999 Hours: M-W 1 3 or by appt.
This course provides basic quantitative analysis and empirical research skills for Urban Studies and Public Administration doctoral students. Though the emphasis is on solving problems to obtain specific and accurate numerical answers, the larger objective is to enhance the students ability to use and interpret linear models through improved understanding of the logic of and justification for the estimation procedures.
The informal prerequisites for the course include competence at college algebra and elementary statistics, as well as ability to use a computer language or software package(s) with matrix algebra and ordinary least squares regression capabilities.
Most of the content of the course will cover a substantial proportion of the main textbook: Damodar N. Gujarati (1995), Basic Econometrics, Third Edition, New York: McGraw-Hill, Inc. This book provides an understanding of regression analysis of the appropriate depth and scope. It does not place heavy mathematical demands on the student, but the examples are nevertheless very illustrative and clear. While the title of the book would suggest an applicability only to the field of economics, the content is actually useful in may other fields including psychology, geography, political science, environmental affairs, business, public administration, and urban studies, among others. For a more mathematically sophisticated treatment of regression analysis see: William H. Greene (1997) Econometric Analysis, Third Edition, Upper Saddle River, New Jersey: Prentice Hall Publishers.
Additional reading will be assigned from the supplementary text: Peter Kennedy (1998), A Guide to Econometrics, Fourth Edition, Cambridge, MA: The MIT Press.
The course will emphasize homework problems. There will be 12 assignments, the grades for which will be assigned on a pass/fail basis. To receive a passing grade, the student must demonstrate first that she/he has made a concerted effort on each of the problems. A grade of "fail" is also possible on extremely late assignments (one that is over five working days late).
Overall course points will be allocated as follows. Homework 25%, Test #1 20%, Test #2 20%, Final exam 35% (20% in class and 15% take home).
Projected Schedule
TU Jan 19 Introduction / Basic Matrix Operations
TH Jan 21 Matrix Operations
Homework 1. Matrix manipulation.
TU Jan 26 Gujarati: Chapter 1: The Nature of Regression Analysis
Kennedy: Chapter 1: Introduction
TH Jan 28 Gujarati: Chapter 2: Two Variable Regression Analysis
Kennedy: Chapter 2: Criteria for Estimators
Homework 2. Complete exercises 1.1, 1.2 (except e), 2.1, 2.2, 2.3, 2.4, 2.7, 2.9 (provide the squared residuals), 2.10, 2.11, and 2.12.
TU Feb 2 Gujarati: Chapter 3: Two Variable Estimation 3.1 3.4
Kennedy: Chapter 3: The Classical Linear Regression Model
TH Feb 4 Gujarati: Chapter 3: Two Variable Estimation 3.5, 3.6, 3A.1 3A.4
Kennedy: Chapter 4: Interval Estimation and Hypothesis Testing
Homework 3. Complete exercise 3.2. By hand, compute regression coefficients, their standard errors the standard error of the estimate, r2, and make a regression plot for problems 3.16 and 3.17 in Gujarati. Do problem 3.18. Also, using data distributed in class and a computer, take 4 separate random samples of size 30 from a data set of size 100, compute regression coefficients, standard errors, standard errors of the estimates, r2, and plot the regression line for each on the same coordinate axis system
TU Feb 9 Chapter 4: The Normality Assumption and Maximum Likelihood Analysis
4.1 4.5
TH Feb 11 Chapter 5: Two Variable Confidence Intervals and Hypothesis Tests
5.1 5.4
Homework 4. Complete exercises 5.1, 5.2, 5.3, 5.8, 5.9 (a and b only) and 5.19. Also, using data distributed in class and a computer, take 3 separate random samples of size 30 from a data set of size 100, compute regression coefficients, standard errors, standard errors of the estimates, r2, and confidence intervals for each coefficient.
TU Feb 16 Chapter 5: Two Variable Confidence Intervals and Hypothesis Tests
5.5 5.11
TH Feb 18 Test #1. (Covers matrices and Gujurati through Chapter 5)
Homework 5. Using data distributed in class and a computer, take 3 separate random samples of size 30 from a data set of size 100, run and interpret regressions, do individual and group predictions (with confidence intervals), and do normality checks on the residuals.
TU Feb 23 Chapter 6: Extensions of the Two-Variable Linear Regression Model
TH Feb 25 Chapter 7: Multiple Regression Analysis: The Problem of Estimation
7.1 7.4
Homework 6. Complete exercises 6.1, 6.2, 6.16, 6.17, 6.22, 7.1, 7.2, 7.18, 7.20 7.21, and 7.22.
TU Mar 2 Chapter 7: Multiple Regression Analysis: The Problem of Estimation
7.5 7.12
TH Mar 4 Chapter 8: Multiple Regression Analysis: The Problem of Inference
8.1 8.5
Homework 7. Complete exercises 8.15 (except (f)), 8.16, 8.29, 8.31, and 8.32
---- Spring Recess ----
TU Mar 16 Chapter 8: Multiple Regression Analysis: The Problem of Inference
8.6 8.12
TH Mar 18 Test #2 (covers through Chapter 8 in Gujarati)
TU Mar 23 Chapter 9: The Matrix Approach to Linear Regression Model
9.1 9.5
Homework 8. Use the matrix approach to linear regression to complete exercises 9.9 and 9.10 plus those assigned in class.
TH Mar 25 Multicollinearity
TU Mar 30 Multicollinearity
Homework 9. Complete exercises 10.5, 10.11, 10.12 and 10.26. Also, as specified in class, create a data set with two independent variables and one dependent variable. First, make one of the independent variables highly collinear with the other. Diagnose the problem, take appropriate remedial measures, and report the results. Then, keeping the first independent variable the same, reduce the level of collinearity in the second one. Again, diagnose the problem, take remedial measures and report the results. Repeat one more time.
TH Apr 1 Heteroscedasticity
TU Apr 6 Heteroscedasticity
Homework 10. Complete exercises 11.1, 11.12, 11.18, and 11.19. Using the data set distributed in class, diagnose heteroscedasticity and take appropriate remedial measures. Report all results.
TH Apr 8 Autocorrelation
TU Apr 13 Autocorrelation
Homework 11. Complete exercises 12.1, 12.2, 12.24 and 12.32. Also, diagnose and take appropriate remedial measures for the spatially autocorrelated data provided in class.
TH Apr 15 Model Specification
TU Apr 20 Dummy Independent Variables
Homework 12. Complete exercises 13.14, 13.15, 15.23, 15.31, 15.32, 16.9, 16.13
TH Apr 22 Models for Dummy Dependent Variables
TU Apr 27 Autoregressive and Distributed Lag Models
TH Apr 29 Review
Mo May 3: 4pm 6pm Final Exam