Schulich School of Business
York University
Course Outline
MSBA 6400 3.00 F: Multivariate Methods for Analytics
Tuesdays, 11:30am-1:30pm
S125 SSB
Winter 2014
Instructor
Murat Kristal
S341 Seymour Schulich Building
Tel: 416-2100 ext. 44593
Email: mkristal@schulich.yorku.ca
Office Hours: By appointment
Administrative Support: Paula Gowdie Rose, S337N SSB, 416-736-5074
Course Description:
This course provides a critical overview of the issues and methods involved in
conducting empirical Operations Management (OM) research. This is a required course
for doctoral students majoring Operations Management.
Cross-listed to OMIS 7300 3.00
Course Overview
This course covers fundamental issues in conducting empirical research in business
administration (e.g., operations management, information systems, marketing,
organizational behavior, corporate strategy). The course is organized according to the
stages in the empirical research process focusing on research methodology. The course
begins with an overview of multivariate methods. The course then considers merits of
alternative research methods and designs, such as partialling and statistical control,
nominal and ordinal independent variables, interaction effects and multi-group analyses,
curvilinear and piecewise linear effects, discriminant and log-linear analysis, cluster
analyses, multivariate regression and canonical correlation. The course than integrates
measurement and model analysis through structural equations modeling.
Course Requirements
Students are required to fully participate at weekly class meetings. Each meeting will
include discussion of specific readings and reporting of any assigned homework,
including data analyses. At each class meeting, students should also turn in written
answers to the assigned homework questions.
Evaluation of Student Performance
Marking in this course conforms to the MBA program's guidelines for grading of
electives. That means that the class average must be less than or equal to A- .
Individual assignments also tend to conform to this rule. Grades out of 100 can be
translated approximately to a letter grade as follows: C- = 60-63, C= 64-66, C+= 6669, B- =70 -73, B= 74-76, B+= 77-79, A-=80-83, A= 84-86, A+ = 87-100.
The course grading scheme for Master’s level courses at Schulich uses a 9-value gradepoint system. The possible course letter grades for a course (and the corresponding
grade points awarded for each grade are:
A+
A
AB+
B
BC+
C
CF
9 grade points
8
7
6
5
4
3
2
1
0
Students are reminded that they must maintain a cumulative GPA of at least 4.2 to
remain in good standing and continue in the program, and a minimum of 4.4 to qualify
for their degree. Schulich grading guidelines mandate a section grade point average
[‘GPA’] of between 4.7 and 6.1 for core courses and a section GPA of between 5.2 and
6.2 for electives.
Your performance in the course will be evaluated on six (6) individual assignments,
each weighted at 16.67% (100%).
Assignments:
Assignments are due at the beginning of each class. I will not accept any late
assignments submitted later. These answers should be brief (usually 4-6 pages,
single spaced) and to the point.
When analyses are involved, written answers should include supporting results (e.g.
excerpts from computer output, the computer code, brief summary tables). These
written answers serve two purposes. First, this course emphasizes hands on application
MSBA 6400 F, Winter 2014
Page 2
of various empirical tools and for learning purposes, there is no substitute for using
these tools and summarizing the results in writing. Second, written answers to
homework questions provide an important means of monitoring learning throughout the
course. When these answers show that certain topics are not well understood, these
topics will be reviewed in class. There will be three major parts to the assignments:
 Answers to the questions regarding the specific methodology covered that week
in class.
 Analysis and reporting of the data.
 Evaluation of the application of that week’s method in articles published in
scholarly journals.
Readings, along with the assignments, will be posted on the CMD as the course
progresses.
Students may consult with each other regarding the concepts and principles underlying
methods used. However, written answers should represent the work of each individual
student. Also, as a general assignment for all class meetings, all students should bring
one or more discussion questions based on the assigned readings and should be
prepared to initiate a discussion of the questions (these questions need not to be
included in the written answers to the assignments). These questions will be addressed
as time permits. Written answers to homework assignments will constitute 100% of the
final grade.
For assignments that require data analyses, students must have access to statistical
software. All the analyses should be conducted in SAS. I will not accept other
statistical packages such as SPSS, or SYSTAT. I also require you to submit your
code in the appendix of each assignment.
Assigned Readings
Required Reading Materials:
Hair, J.F. Jr., Black, W.C., Babin, B.J., Anderson, R.E. 2009. Multivariate data analysis.
Upper Saddle River, NJ: Prentice Hall.
Pedhazur, E.J., & Schmelkin, L.P. 1991. Measurement, design, and analysis: An
integrated approach. Hillside, NJ: Erlbaum.
Raykov, T., & Marcoulides, G.A. 2008. An introduction to applied multivariate analysis.
New York, NY: Routledge.
Williams J.M. 1995. Style: Toward clarity and grace. Chicago. IL: University of Chicago
Press.
MSBA 6400 F, Winter 2014
Page 3
Pedhazur and Schmelkin (1991) will be the primary text for the course. The books by
Hair et al. (2009) and Raykov and Marcoulides (2008) provide practical guidelines for
multivariate analyses. Williams (1995) provide background that will be very useful for
written assignments for this class and papers for other degree requirements and for
publication. I also compiled a list of books that might be helpful for conducting analyses
in SAS.
Complementary Reading Materials:
Hatcher, L. 1994. A step-by-step approach to using SAS for factor analysis and
structural equation modeling. Cary: NC: SAS Publishing.
Littell, R.C., Milliken, G.A., Stroup W.W., Wolfinger, R.D., & Schabenberger, O. 2006.
SAS for mixed models. Cary: NC: SAS Institute.
Spector, P.E. 2001. SAS programming for researchers and social scientists. Thousand
Oaks. CA: Sage Publications.
Stokes, M.E., Davis, C.S., & Koch, G.G. 2000. Categorical data analysis using the SAS
system. Cary: NC: SAS Institute.
Other reading materials will consist of chapters and journal articles, and copies of these
materials will be e-mailed. After the course begins, I will also recommend that you
purchase several of the monographs published by Sage (often referred as the “little
green books”).
Academic Honesty
Academic honesty is fundamental to the integrity of university education and degree
programs. The Schulich School will investigate and will act to enforce academic honesty
policies where apparent violations occur. Students should familiarize themselves with
York University’s policy on academic honesty. It is printed in full in your student
handbook and can also be viewed on-line on the Schulich website, clicking through as
indicated:
Schulich website ‘Programs’ -> ‘Master’s Degree’ -> ‘MBA' -> ‘Academic Honesty’
While academic dishonesty can take many forms, there are several forms of which
students should be highly aware because they are the ones that are most likely to occur
in the context of a specific course.
1) Plagiarism. Plagiarism is the presentation of information, ideas, or
analysis generated by other people as being your own. It includes
direct quotations as well a substantive paraphrases where the course of
MSBA 6400 F, Winter 2014
Page 4
that information or idea is not clearly identified to the reader. Students
should be careful to present their written work in a way that makes it
completely clear in each and every case where a quotation, a
paraphrase, or an analysis is based on the work of other people. (This
includes information from all sources, including websites.)
2) Cheating. Cheating is an attempt to gain an unfair advantage in an
evaluation. Examples of such violations include (but are not limited to)
consulting prohibited materials during an examination or copying from
another student.
3) Failure to follow limitations on collaborative work with other
students in preparing academic assignments. Each class differs in
the mix of assignments and group-versus-individual preparation that is
allowed. The instructor will make clear the extent of collaboration
among students that is acceptable among students on various pieces of
assigned work. Students should abide by those limitations and, if they
are unsure about whether a certain level or form of collaboration would
be acceptable, to clarify that question with the instructor in advance.
4) Aiding and abetting. A student is guilty of violating academic
honesty expectations if he/she acts in a way that enables another
student to engage in academic dishonesty. If a student knows (or
should reasonably expect) that an action would enable another student
to cheat or plagiarize, that student’s action constitutes an academic
honesty violation. Illustrative examples include making your exam
paper easily visible to others in the same exam or providing your own
working or finished documents for an ‘individual assignment’ to another
student (even if that other student said that he/she just wanted to ‘get
an idea of how to approach the assignment’ or ‘to check whether they
had done theirs correctly’).
5) Use of academic work in more than one course. Generally,
academic work done for every course is ‘new’ work, done for that
course only. If a student wishes to use some or all of the academic
work done for an assigned task in one course in another course, the
student must get explicit, prior permission from both instructors so that
they agree that the scope and nature of the overlapping use of that
work is such that it can fairly be counted toward both courses.
Schedule of Topics and Readings
The following list of lecture topics and readings indicate the material to be read,
reviewed and/or prepared for the various class sessions. If any changes in this schedule
MSBA 6400 F, Winter 2014
Page 5
become necessary, notifications will be posted in the course CMD, and where such
changes need to be announced between class sessions, an email will be sent to
students’ Lotus Notes email accounts, notifying them of the change.
DATE
SESSION
TOPIC (S)
January 7
1
Overview of Multivariate Methods
 Framework for classifying models (univariate vs. multivariate
models, continuous vs. categorical variables)
 Preparing data for analysis
January 14
2
Partialling and Statistical Control
 The meaning and uses of statistical control: simple, semipartial, and partial correlation
 Partitioning explained variance
 Suppressor variables
January 21
3
Nominal and Ordinal Independent Variables
 Dummy coding
 Effect coding and contrast coding
 Cost of dichotomizing continuous variables
January 28
4
Interaction Effects and Multi-Group Analyses
 Meaning of interaction
 Estimating interactions between various combinations of
categorical and continuous variables
 Higher-order interactions; scaling issues
February 4
5
Discriminant and Log-Linear Analysis
February 11
6
Cluster Analysis
 Cluster analysis compared to other multivariate methods (e.g.
principal components analysis, MANOVA, discriminant analysis)
 Grouping objects vs. variables
 Measures of similarity, hierarchical vs. non-hierarchical
clustering methods
February 18
7
NO CLASS – READING WEEK
February 25
8
Mutlivariate Regression and Canonical Correlation
 Advantages of multivariate analysis over separate univariate
analysis
 Canonical correlation as an extension of univariate regression,
discriminant analysis, MANOVA, and principal components
analysis
 Multivariate regression and multivariate hypothesis testing, set
correlation
MSBA 6400 F, Winter 2014
 Problems of OLS regression with categorical dependent
variables
 Discriminant analysis and its association with MANOVA and
canonical correlation
 Estimation and interpretation of logistic regression equations
Page 6
DATE
SESSION
TOPIC (S)
March 4
9
Other Issues in Regression Analysis
 Generalized Linear Models, violations in assumptions of
regression analysis
March 11
10
Path Analysis
 Exogenous and endogenous variables; direct and indirect
effects
 Spurious and unanalyzed associations
 Decomposition and reproduction of correlations
 Assessing model fit
March 18
11
Exploratory Factor Analysis
 Principal components vs. common factor analysis
 Extracting and rotating factors
 Construction of factor scales and scores
March 25
12
Confirmatory Factor Analysis
 Comparisons between exploratory and confirmatory factor
analysis
 Model specification, identification, and estimation
 Assessment of model adequacy and fit
April 1
13
Structural Equation Modeling
 Comparisons between path analysis and structural equation
modeling
 Alternative methods for handling measurement error
 Specification and identification of structural equation models
MSBA 6400 F, Winter 2014
Page 7