MGS 8110: Applied
Regression Analysis
Syllabus for Summer 2014
|
Instructor: Dr. Satish Nargundkar Office
Hours: By appointment |
E-Mail : snargundkar@gmail.com Phone: (678) 644 6838 CRN: 52307 ALC 403 MW
7:40-10:10 PM |
Text:
Applied Regression and
Correlation, by Jeremy Miles and Mark Shevlin, Sage Publicaitons, Washington, D.C. 2011. ISBN 9 780761 962304. ISBN 13: 978-0-7619-6230-4. Should be available in the University Bookstore. Also online
– click
here.
Description
The basic aim of this course
is to improve the student's understanding of the various uses of regression
analysis. Both simple and multiple regressions are discussed in detail,
including non-linear estimation. Acquainting the student with the assumptions
of the general regression model is another aim of the course. Practical
suggestions are given for checking the regression model by examining the
residuals. More than one software program may be used for regression.
Learning Outcomes/Course Objectives
Upon completion of the
course, students will be able to:
1. Develop a general framework
for decision support within organizations.
2. Perform basic data cleaning
operations..
3. Build Simple and Multiple
Regression models to understand relationships and predict the values of the dependent.
4. Use both continuous and
categorical independent variables in models.
5. Interpret results in the
context of the issue at hand.
6. Evaluate regression models
for error, goodness-of-fit.
7. Compute and interpret
interaction effects.
8. Analyze and interpret moderation
and mediation effects.
9. Check for violations of
regression assumptions.
10. Compute the sample size
required for a study given a required degree of power.
11. Discuss how causality is
related to regression analysis.
Grading:
|
Item |
Points |
|
Course Average |
Grade |
Course Average |
Grade |
|
Assignments /
Participation |
20% |
|
94-96,
97+ |
A,
A+ |
77-79 |
C+ |
|
Midterm |
30% |
|
90-93 |
A- |
73-76 |
C |
|
Team Project |
20% |
|
87-89 |
B+ |
70-72 |
C- |
|
Final Exam |
30% |
|
83-86 |
B |
60-69 |
D |
|
Total |
100% |
|
80-82 |
B- |
Less
than 60 |
F |
All assignments should be
posted to the MyRobinson website. It is also recommended
that you post them to your own website.
Final Project
Collect non-time-series data on any topic of interest to you, preferably
something related to your work (application of the techniques from this course
to your work will add value to your organization and to you). You must have at
least 5 independent variables (can be a mix of categorical and numeric) in your
final model and a dependent variable (numeric). The number of observations will
depend on the circumstances. However, the minimum sample size should conform to
the guidelines presented in Miles and Shevlin
(M&S).
1. Show descriptive statistics
of all the variables in a table in the Appendix. [get some feel for the data]
2. Show relationships of each
independent variable individually with the dependent using scatter plots in the
Appendix. [get some initial feeling about which variables are more related to
the dependent variables, and look for possible outliers or influence points]
3. Correlation analysis of
independent variables.
4. Perform regression analysis
to show overall model for predicting the value of the dependent.
5. Check for the violations of
the assumptions of (Ch. 4 of M & S) and issues (Ch
5 of M & S) with the regression model in your data. Describe what steps you
took (if any) to account for these violation and issues in your final model.
6. Interpret the results and
write a report.
7. Make an in-class
presentation (no more than 15 minutes) of all your work.
Submission:
Report
Introduction
1. Why you chose this topic
(some background, motivation).
2. What you are trying to
predict (e.g., price of houses)
3. Why you choose certain
independent variables (e.g., # of bedrooms) for your project [Logical
explanation is required.]
4. How you collected the data
(e.g., survey or from Web) [You should clearly write the source of the data.]
Analysis
5. Findings from descriptive
statistics, histograms and scatter plots [write
what variables are expected to have relationship with the dependent variable.]
6. A short discussion of the
insignificant independent variables in the full model (i.e., a model with all
the independent variables to start with)
7. Order of the dropped
independent variables in subsequent regressions with reasoning of such order
till you reached the model where all independent variables are statistically
significant.
8. Test for the violations of
the assumptions and other issues with this model. (e.g.
heteroskedasticity etc.) Discuss the steps you took
to address those.
9. Equation of the final model
(i.e., a model with only significant independent variable(s) and after
addressing the problems noted in step 8) with the estimated values of the
coefficients and the estimated standard errors.
10. Performance of the final
model. i.e., How good is your model? [Low R-squared
does not mean low grade.]
11. Findings from the final
model. i.e., interpretation of the coefficients of the independent variables
[Make sure that all the coefficients make sense. If it does not, explain
further how such odd coefficient can be justified]
Conclusion [No tables or diagrams]
12. A paragraph summarizing what
you sought out to model in the modeling, what were your experience in doing
that (including any surprises) and your conclusions.
Appendix – This should contain the tables, diagrams, and regression result
tables etc.
Statistical Package for Analysis – You may use any statistical package including but
not limited to SPSS, Stata, R, SAS, Minitab.
General Policies:
1. Students are expected to
attend each class (who knows, you may actually enjoy the class!), arrive on
time and participate in class discussions. As the instructor, I also make the
commitment to come to class prepared!
2. Turn off cell phones, pagers, stereos, TVs, etc. when in class. Treat the instructor and each other with courtesy.
Honor
Code:
It
is more honorable to get any grade with your own work than to get a
better grade by using someone else’s work as yours. While discussion with
classmates is encouraged, any work submitted must be your own (or your group’s,
for group projects). Evidence of plagiarism/cheating on an assignment/exam will
result in a failing grade for that assignment/exam, or even for the course.
Course Assessment:
Your constructive assessment
of this course plays an indispensable role in shaping education at
Accommodation of Disability
Students
who wish to request accommodation for a disability may do so by registering
with the Office of Disability Services.
Students may only be accommodated upon issuance by the Office of
Disability Services of a signed Accommodation Plan and are responsible for
providing a copy of that plan to me by the second week of classes. The Office
of Disability Services of GSU can be reached at http://www.gsu.edu/disability/.
A
copy of the Requests for Academic Accommodations form can be downloaded from
here
If
the student wishes to give the examinations at the Office of Disability Services
then he/she needs to submit the R.I.T.A form which can be downloaded from http://www.gsu.edu/images/Disability%20Services/Updated_-_RITA_form_-_2011.pdf.
GSU Academic Support
The Counseling Center: http://www.gsu.edu/counseling/
The Counseling Center offers several workshops and
seminars designed to promote student academic success. Most of these services, like study and
note-taking skills, are offered in the Life Skills Lab. These services are free
of charge and are available to all students. Please contact the Counseling
Center at 404-413-1640 to make an appointment.
The Office of Student Support Services: http://www.gsu.edu/oeo/sss.html
Student Support Services (SSS) offers one-on-one
academic counseling to students who seek to develop academically. The academic
counseling is tailored to fit the individual needs of each student and can
include such issues as test anxiety, test taking strategies, time management,
and study skills. This is for college students who demonstrate academic need
and are first generation college students, and/or low income, and/or have a
documented disability.
For more information contact SSS at
404-413-1680 or by clicking on the link above.